TRANSLATING EVALUATIVE DISCOURSE: THE

SEMANTICS OF THICK AND THIN CONCEPTS

SHYAM RANGANATHAN

A DISSERTATION SUBMITTED TO THE FACULTY

OF GRADUATE STUDIES
IN PARTIAL FULFILMENT OF THE REQUIREMENTS
FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY

GRADUATE PROGRAM IN PHILOSOPHY

YORK UNIVERSITY,
TORONTO, ONTARIO

SEPTEMBER 2007
TRANSLATING EVALUATIVE DISCOURSE: THE SEMANTICS OF
THICK AND THIN CONCEPTS

by Shyam Ranganathan

a dissertation submitted to the Faculty of Graduate Studies of

York University in partial fulfi llment of the requirements for
the degree of

DOCTOR OF PHILOSOPHY
©
Permission has been granted to: a) YORK UNIVERSITY
LIBRARIES to lend or sell copies of this dissertation in paper,
microform or electronic formats, and b) LIBRARY AND
ARCHIVES CANADA to reproduce, lend, distribute, or sell
copies of this dissertation anywhere in the world in microform,
paper or electronic formats and to authorize or procure the
reproduction, loan, distribution or sale of copies of this
dissertation anywhere in the world in microform, paper or
electronic formats.

The author reserves other publication rights, and neither the

dissertation nor extensive extracts from it may be printed or
otherwise reproduced without the author’s written permission.
ABSTRACT

According to the philosophical tradition, translation is successful when one has substituted words and

sentences from one language with those from another by cross-linguistic synonymy. Moreover,

according to the orthodox view, the meaning of expressions and sentences of languages are

determined by their basic or systematic role in a language. This makes translating normative and

evaluative discourse puzzling for two reasons. First, as languages are syntactically and semantically

different because of their peculiar cultural and historical influences, and as values and norms differ

across cultures, it is unlikely that languages will have synonymous evaluative and normative

expressions. If translation is only successful by cross-linguistic synonymy, it would seem that we

would not be able to translate the value theoretic claims of persons from radically different cultures.

But it is with such persons that dialogue on evaluative matters is imperative to resolve ethical and

axiological differences that could be the root of conflict. Second, as the orthodox account of meaning

renders it linguistically relative, it is unlikely that expressions across languages will be cross-

linguistically synonymous. Thus, on the orthodox account of translation, translation is indeterminate

(as W.V.O. Quine has argued) or impossible (as Jacques Derrida has argued). In this dissertation I

argue for a novel theory of meaning and translation based on innovations in the translation studies

literature and my prior work in cross-cultural research, which I call “Text-Type Semantics” or “TTS.”

TTS explains how translation is successful while affirming radical cultural and linguistic diversity.

On the basis of TTS I argue that we need what I call the “Quasi-Indexical” account of thick and thin

concepts (or QI) to translate normative and evaluative discourse. According to QI, thick and thin

concepts are best thought of as a subset of philosophical concepts defined by their roles in

philosophical texts. I argue that QI and TTS succeed where competing accounts in the moral

semantics literature (such as Non-Analytic Naturalism and Expressivism) fail.

iv
 Abstract………………………………………………………………...…………. iv

I. INTRODUCTION .......................................................................................... 1

I.1. What Translation and Ethics Have to do with Each Other .............................1
I.2. The Institutional-Textual Turn......................................................................15
I.3. Antecedents...................................................................................................20
I.4. Looking Ahead..............................................................................................36

II. TRANSLATION, EQUIVALENCE AND TEXT-TYPES ................................ 40

II.1. Introduction...................................................................................................40
II.2. The Liberated Equivalence Conception of Translation ................................43
II.2.1 Criticisms of Equivalence in Translation ....................................................43
II.2.2 Advantages of LET .....................................................................................45
II.3. Philosophical Criticisms of Equivalence I: Deconstruction .........................50
II.3.1 Heidegger’s Hopes for Translation .............................................................50
II.3.2 Deconstruction’s Response .........................................................................51
II.3.3 Equivalence against Deconstruction ...........................................................53
II.4. Philosophical Criticisms of Equivalence II: Indeterminacy of
Translation ....................................................................................................57
II.4.1 Enlightenment and Positivism.....................................................................57
II.4.2 Quine’s Thesis and Argument Against Equivalence and Determinacy ......58
II.4.3 Davidson’s Attenuation of the Indeterminacy of Translation .....................61
II.4.4 Overcoming the Threat of Quinean Indeterminacy of Translation: LET....64
II.4.5 Responding to Quinean Criticisms..............................................................68
II.4.6 Indeterminacy of Translation And Difficult Texts......................................80
II.4.7 Responding to Davidsonian Indeterminacy of Translation.........................85
II.4.8 LET and Radical “Translation”...................................................................85
II.5. Skopos Theory and LET ...............................................................................87
II.6. Functionalism................................................................................................89
II.6.1 Functional Linguistics and Translation.......................................................89

v
 II.6.2 Nida’s Dynamic Equivalence and Translation Criterion ............................92
II.6.3 Criticisms of Functionalism ........................................................................93
II.6.4 Functionalism, And Inter-Cultural Communication ...................................95
II.6.5 The Limited Legitimacy of Functionalism................................................101

III. TEXT-TYPES AND SEMANTIC DETERMINACY ................................... 103

III.1. Meaning and Translation ............................................................................103

III.1.1 Conversation and TTS...............................................................................112
III.1.2 Science, Objectivity and TTS ...................................................................115
III.1.3 Language and Meaning .............................................................................126
III.2. Why Determinate Translation Must Preserve Meaning.............................130
III.2.1 The Case of “Technical” and scientific Translation and the Question of
the Determination of Reference ......................................141
III.3. Text-Types do not Reduce to Systematic or Literal Features of a
Language.....................................................................................................148
III.4. The Objectivity of Determinate Translation ...............................................152
III.5. Text-Types in Comparison: Philosophy, Science and Literature
Broadly Conceived 155
III.6. Finding our Feet..........................................................................................161

IV. ETHICAL NATURALISM AND LINGUISTIC ACCOUNTS OF MORAL

SEMANTICS ................................................................................................ 168

IV.1. Non-Analytic Moral Naturalism .................................................................170

IV.2. Moral Twin Earth........................................................................................173
IV.3. Why the Moral Twin Earth Argument Fails ...............................................177
IV.4. Quality Science Fiction for the Analytic Philosopher ................................180
IV.5. The Explanatory Failure of Ethical Naturalism ..........................................193
IV.6. Native Linguistics, Intuition and Ideology .................................................212
IV.7. Conclusion ..................................................................................................221

V. EXPRESSIVISM, TRANSLATION AND NORMATIVE DISAGREEMENT..... 225

vi
V.1. Expressivism ...............................................................................................226
V.2. Indeterminacy, Impossibility, and Non-Equivalence of Translation ..........228
V.3. Indeterminacy, Translation, Languages and Texts .....................................230
V.4. Text-Types and Translation ........................................................................231
V.5. Classical Expressivism and the Task of Translation...................................241
V.6. Expressivism Revised .................................................................................242
V.7. Conclusion ..................................................................................................249

VI. QUASI-INDEXICAL ACCOUNT OF THICK AND THIN CONCEPTS ......... 252

VI.1. Review of TTS and the Semantic Foundation of QI...................................253

VI.2. Indexicality and Philosophy in the Literature.............................................256
VI.2.1 Kaplan’s Theory........................................................................................256
VI.2.2 Indexicality and Ethics..............................................................................257
VI.2.3 Indexicality and Epistemology..................................................................266
VI.2.4 Indexicality, Key Philosophical Terms and Translation ...........................272
VI.3. TTS and Quasi-Indexicality........................................................................273
VI.3.1 The Basic Model and its Variables ...........................................................273
VI.3.2 Thin Terms 284
VI.3.3 Thick Terms 288
VI.3.4 Thick-Thin Distinction and “Justice”........................................................289
VI.4. Criticisms and the Distinction of QI ...........................................................292

VII. TRANSCENDING CONTEXTS .............................................................. 298

VII.1. Contextualism vs. Cotextualism .................................................................299

VII.1.1 Contextualism and Semantic Indeterminacy.............................................303
VII.1.2 Semantic Determinacy and the role of Institutions in TTS .......................309
VII.1.3 How TTS Avoids the Problems of Contextualism Part 1—Meaning is
Textual ............................................................................310
VII.1.4 How TTS Avoids the Problems of Contextualism Part 2—Social
Practice that Allows for Determinacy is Institutional,
not Contextual.................................................................311
VII.1.5 Objections from Contextualism ................................................................314

vii
 VII.1.6 What we can learn from Contextualism....................................................317
VII.2. Insensitive Semantics..................................................................................318
VII.2.1 TTS and Insensitive Semantics .................................................................319
VII.2.2 TTS and Cappelen and Lepore’s Tests .....................................................324
VII.2.3 Troubles of Cappelen and Lepore’s Project ..............................................327
VII.2.4 TTS, Derrida and the Fulfilment of the Promise of Insensitive
Semantics ........................................................................340
VII.2.5 Moderate Contextualism and QI ...............................................................343
VII.3. QI, Belief and Knowledge ..........................................................................343

VIII. CONCLUSION ................................................................................... 355

VIII.1. Summary of Argument for QI.....................................................................355

VIII.2. History, Culture and Objectivity.................................................................359
VIII.3. Objectivity and Epistemic Pluralism .........................................................367
VIII.4. Relativism: Castles in the Sky ....................................................................373
VIII.5. Conclusion ..................................................................................................390
Endnotes………………… ....................................................................................392
References………………… .................................................................................425

viii
 I. Introduction

“For most people, translation is a threatening process…” Thomas Kuhn, Structure of

Scientific Revolutions

I.1. What Translation and Ethics Have to do with Each Other

My thesis is that we need what I call the “Quasi-Indexical” account of thick and thin concepts

(henceforth QI) to translate what I will call “normative discourse” or alternatively “evaluative

discourse.” By “normative discourse” or “evaluative discourse” I designate discourse that is

characterized by verbal markers such as “ought to” or nominal markers such as “good” and “ethical.”

QI construes the foundational concepts of philosophy as devices of translation. Normative/evaluative

concepts are best thought of as a subset of these concepts. The full defence of QI involves a defence

of the account of meaning and translation that underwrites QI, which I call “Text-Type Semantics”

(henceforth “TTS”). Both will be defended in light of general considerations from the philosophy of

language, semantics and translation theory, as well as the particular requirements of translating

normative discourse as informed by my own cross-cultural research.

Translation is a topic that many of the big figures in twentieth century western philosophy

wrote on (such as Rudolph Carnap, W.V.O. Quine, Martin Heidegger, Hans-Georg Gadamer and

Jacques Derrida), but the largely sceptical consequences of this avenue of investigation have lead to

translation fatigue, with the consequence that the topic is seldom discussed in the recent literature.

Much of this dissertation tracks issues in translation theory and the philosophy of language, and is a

response to the sceptical conclusions about translation that one finds both in the Analytic and

Continental traditions. This dissertation will not only show how translation is not philosophically

problematic at all, but it will present the first systematic and principled account of translating

philosophy and normative discourse in the literature. Surprisingly, nothing like this has been

1
undertaken before, even though translation is indispensable to the operation of philosophy as a

discipline concerned with canons that are culled from disparate cultural and linguistic sources.1

My prior work consists in cross-cultural research, particularly on problems in translating

Indian ethics from Sanskrit into English. It may thus surprise some to find out that I consider myself

primarily a moral philosopher. How is it that someone who considers himself an ethicist comes to

spend so much time writing on translation and the philosophy of language?

There are several reasons why we should be concerned simultaneously with ethics and

translation. First, I’ll put forward what I think is a mediocre reason to be concerned with the two

simultaneously. Some translation theorists and cultural studies critics have argued that we should

constrain translation and cross-cultural research in light of ethical considerations, because, as

translation is a mode of cross-cultural interaction, and as a translator one is in the business of re-

presenting the views of people to novel audiences, translators bear the responsibility for facilitating

intercultural communication. This is a great burden and the only way to do it responsibly is in light of

ethical considerations (cf. Chesterman 1997; Venuti 1998).

In general, people should conduct their professional life and activities in light of ethical

considerations. The notion that the translator must especially be concerned with such issues seems to

me to be superfluous. So, I take this to be a mediocre reason to run a concern for translation and

ethics together.

Here is a slightly better reason for thinking that ethics and translation should go together. We

live in a multicultural world. Diversity is an actual reality in all societies: even those in which the

ruling classes want to minimize diversity. If we are to properly educate people we must make sure

that our education reflects the cultural diversity of our populations and the world. At present,

educational systems are characterized by a chauvinism that privileges some cultural traditions over

others. This does not reflect the reality of diversity in student populations, or the reality of human

history.

2
 I much prefer this reason for taking translation seriously, but it’s not obviously an ethical

concern, for we can understand the imperative to properly represent humanity’s shared intellectual

and cultural accomplishments as a purely educational or epistemic imperative. Educations fail people

when they are one-sided. That it is possible to get a degree in “philosophy” by only studying a small

smattering of European and English-speaking thinkers in most philosophy departments in the West

(for instance) is an educational shortcoming. It leads to much nonsense, particularly in the form of

uneducated generalizations trotted out by professors (e.g., “Hume was the first philosopher to

question induction…” “Thales was the first philosopher …”—both false, but one wouldn’t know that

unless one studied other philosophical traditions). Many departments claim to provide specialized

training in the history of philosophy, but this amounts in practice to a survey of the history of

philosophy in the Western tradition. We are certainly failing on epistemic grounds by short-changing

ourselves and our students by focussing so closely on only the Western tradition. But it’s not

obviously or primarily a moral failing.

Here is a better reason for thinking that the two issues should go together. Translation is our

best way to achieve cross-cultural communication, and sadly, we do not communicate across cultures

very well. Adequate communication will certainly not solve all the world’s problems, but if we are

not very good at communicating with each other across cultures, the chances that we should

unintentionally offend each other is relatively high. Wars have broken out over lesser issues. If ethics

has some intimate connection to the way we live our lives and relate to each other, then being able to

properly understand each other, through translation, would certainly be an asset to moral deliberation.

We would then have some understanding of each other’s aims, which could variously be factored into

our practical deliberations.

I like this reason for thinking that translation is important to moral philosophy. But I think

there is an even better reason for being concerned with the two. Knowledge increases when we

communicate with others. There are very few major accomplishments in life that are solitary. We are

3
able to do what we do because of each other’s help. We are, after all, social beings quite incapable of

a healthy life without each other, and it is our innate dependency upon each other that is our strength.

If we can break cultural barriers and communicate about ethical issues, the way we do about science

and medicine, for instance, we would certainly be in a better position to have moral knowledge, it

seems to me. No one thinks that scientific or mathematical knowledge is had best by cloistering

oneself with one’s compatriots, nor does anyone think that their native language furnishes them with

all the information and concepts necessary to do a good job of science and mathematics. What people

learn about these matters from their parents and guardians is limited, and usually not correct. People

who are serious about these issues self-select, find others with common interests, and jointly

deliberate and work on increasing knowledge in their respective fields, after gaining the appropriate

institutional training. If we were to approach moral knowledge in the same light, then indeed we

would be in an enviable position.

There would also be an immense practical advantage to being professional and international

about moral knowledge in exactly the same way that scholars are professional and international about

science and mathematics. We would be able to make life better for all concerned—ethically better.

And, if we could truly coordinate our actions in light of such knowledge, then much conflict in the

world would certainly be attenuated.

Great idea. It turns out however that there is a long tradition of thinking about ethical

knowledge as a species of cultural, local knowledge. While people are often comfortable with the

notion that scientific and mathematical knowledge requires unusual institutional training, there is a

diffuse belief that ethical knowledge is not like this.

Every major tradition of philosophy has had its defenders of what we might call ethical

localism.2 In the West, the tradition goes back to Aristotle who held that ethical knowledge is

something we gain from the right upbringing. In China, Confucius appeared to hold a similar view, to

the extent that he thought that li, or the right or propriety, is deeply social and embedded in the norms

4
of filial piety. In ancient India, there was a similar tradition of Brahminical thinking that moral

knowledge is something gained from a sacred tradition, and encapsulated in treatises on ethics

(dharma stra-s) that detailed the obligations of all persons of “Aryan” society. In all three cases,

ethical knowledge is a matter of being in the right geographical location and the right social context.

No special schooling, aside from the right type of socialization, is necessary or possible in ethics.

A disturbing similarity between these various approaches to ethics is the notion that some

people simply do not count for much. Aristotle thought that some people were born for slavery

(Politics I.iii-vi, Nicomachean Ethics IV). The orthodox Brahmins that cultivated the treatise on

ethics imposed a strict hierarchy on their ethical thinking, with the result that the weight of a moral

offence was in part proportionate to the place of the victim in the hierarchical order. People on the

lowest rungs of this hierarchy were conceptualized as more expendable.3 Confucius had little respect

for women and thought of them as difficult.4 Advocates of ethical localism typically criticize these

faults of the past. But it is unclear why we should think that simply criticizing the past makes us any

better at knowing what is ethically right if such knowledge is simply a species of cultural, local

knowledge. To the extent that there has been some moral progress in human history (which has hardly

been uniform, and moral regress happens all the time) it seems that human beings have had to gain

ethical knowledge the hard way, by experimenting, making huge mistakes and later regretting them

(sexism, slavery, racism, homophobia, speciesism… the list will no doubt go on). If we were

professional and international in ethics in the way that we are about mathematical and scientific

issues, then it seems that we could avoid much of this because we would not think that what is local is

identical with what is right. Translation and moral philosophy would thus go hand in hand.

Translation would allow us to do and think about moral philosophy, and moral philosophical

problems, in a manner that was not specifically local in nature, and we might thus be able to avoid the

pitfalls of thinking that what is local is always what is correct in ethics. This, I think, is the best

reason listed so far to think that translation is vital to ethics.

5
 Here we come up against several obstacles to such aspirations.

The first set of problems concern controversies in theoretical ethics about what it is that we

are talking about when we talk about moral issues. There are seemingly interminable debates about

moral semantics, which are debates about the meaning of our moral expressions in particular

languages. There is a long sceptical tradition that suggests that ethics is really not about anything

objective at all: it is at worst a means of expressing our emotions, and at best a means of expressing

our outlook about how we would like the world to be (cf. Ayer 1946 [1936], 108-112; Stevenson

1944, 3; Blackburn 2000 [1998], 50; Gibbard 1990, 7-8). On the opposite end, some theorists argue

that ethical discourse is really about natural properties that science can study. On such an account,

ethical knowledge is a species of scientific knowledge (cf. Boyd 1988; Brink 1989, 132; Railton

1986, 204; cf. Railton 1996). Others find this incredible. Famously, John Mackie (1977) argued that if

we are to take ethical claims seriously then they must be understood as making claims about the

moral valence of objects and properties in the world, but yet physics and our best science make no

room for queer properties that somehow demand our compliance. We can add to this list the views of

the localists, who take ethics to be a type of expression of local culture. Given so much controversy,

it is not even clear that there is any objective subject matter of ethics for us to even deliberate about.

Then, there are even deeper objections, hailing broadly from what might be called the

philosophy of language, semantics and translation theory. The most powerful of such objections is

that meaning in general is a local, sociological, cultural affair. The very notion that we could

contemplate ethical claims apart from such contexts is dubious on this account.

Related to this thesis is the very troubling argument that translation is indeterminate.

Arguments for this thesis are classically found in Quine’s Word and Object (1960), though similar

arguments can be found in the work of Jacques Derrida. While some technical criticisms are lucid,

such as Scott Soames’s recent argument that Quine’s thesis depends on the dubious premise of

behaviourism, or that it relies upon equivocating between two definitions of “determination” (Soames

6
2003, 2:243-255), most criticisms of Quine (including Soames’s, I think) miss the forest for the

trees.5 There is a very simple way to recast Quine’s argument based upon his own characterization of

the thesis of the indeterminacy of translation that is not technical and thus resistant to the usual

criticisms.6 Here it is. Translation proceeds by attempting to find synonyms across languages. A

successful translation preserves meaning by reconstructing a text whose components all have the

same meaning as the components of the original. But languages are historically particular, culturally

distinct artefacts with internal relations and differences. When we are translating we treat expressions

as though they have atomic, distinct meanings. But because their meaning is a function of linguistic

factors that are culturally and linguistically specific (such as, for instance, a language’s grammar that

makes obligatory certain semantic devices, such as gender, number, and tense, and makes unavailable

grammatical devices of other languages), no expressions across languages will really be equivalent in

meaning and thus the very project of translation is doomed to have no objective standard of success.

No empirical evidence can help us decide cases of translation because we are really trying to do the

impossible, namely, find semantic equivalents among linguistic items (words and sentences) that are

in no strict sense equivalent. Translation is in some strict sense impossible, except that we press on

with it for pragmatic reasons, and our results are uncertain, or indeterminate.

Listed here is a mixed bag of theses that are widely considered and taken seriously that singly

or jointly present problems for understanding how moral knowledge could be international,

professional, objective and not simply a projection of local and parochial ways on the world. Some of

these theses concern ethical knowledge specifically. For instance, anti-realist views about ethics,

localist views that ethical knowledge is had by consulting one’s own cultural authorities, or

conflicting and problematic theses in moral semantics make ethical and normative knowledge that is

anything but a parochial projection on the world, difficult to contemplate. However, some of these

theses are a threat to all objectivity in knowledge.

7
 The most threatening to knowledge as such, I believe, are the views that meaning in the final

analysis is linguistic and that translation is indeterminate. Really, these are two sides of the same coin.

If meaning is linguistic in nature (i.e., a property of languages), then it is an empirical question as to

whether all languages instantiate the same meanings. Given the culturally peculiar histories of

languages, it is highly doubtful that languages will present us with the same semantic profiles. If

languages as such are semantically distinct, but meaning is in the ultimate analysis relative to a

language, then translation becomes puzzling. The only rational constraint that we have on successful

translation is that it preserves meaning: if a translation and an original have different meanings, it is

difficult for us to understand how they can count as the same work, or even express the same

knowledge. If meaning just is linguistic, and if translation is indeterminate, then science,

mathematics, and all knowledge that aspires to objectivity becomes threatened, for we have no reason

to believe that our scientific and mathematical knowledge is anything but a reflection or expression of

our culture as such. This is not to say that we wouldn’t have reason to take science and mathematics

seriously. They have proven themselves useful beyond what the imagination could conjure up on its

own and thus they do seem to be getting at something very important about reality and life. But, if our

best science and mathematics could not be determinately translated, then we would have reason to

doubt that our knowledge, as useful as it is, is truly objective. Despite the technological facility it

provides, it could simply be a very useful expression of our culture. The notion that we were getting

at anything independent of our mores and language would be very doubtful.

The notion that scientific knowledge is at base cultural is nothing new, and many academics

may have become rather comfortable with the idea. Thomas Kuhn argued in his Structures of

Scientific Revolutions that scientific knowledge is founded upon a paradigm, which he likened to a

linguistic practice that was formed around shared theoretical, methodological and terminological

resources. The truth of scientific claims is quite unintelligible apart from such a foundation, Kuhn

argued. Kuhn reasoned that defenders of contrary paradigms were faced with the challenge of

8
translation when they communicated with each other: each would attempt to understand the other in

terms of their own linguistic resources and the results would not always be perfect understanding or

clarity. Kuhn was thus putting forward a type of cultural relativism in science. Despite this, Kuhn

thought that his account was not problematically relativistic, but it is difficult to see why not,

particularly because he is lead to the conclusion that the very truth of scientific claims is only a matter

that makes any sense relative to specific scientific theories (Kuhn 1970, 202-206). Worse, Kuhn did

not appreciate the trouble that his account of translational indeterminacy could cause to his view of

scientific knowledge. Kuhn had gleefully deferred to Quine’s account of translation as a model for his

(Kuhn 1970, 202 fn.17). But he did not notice that Quine thought that his argument for translational

indeterminacy threatened not only cross-linguistic attempts to determine meaning, but intralinguistic

efforts as well. Having established to his satisfaction the conclusion that translation is indeterminate,

Quine reasons that if our best, empirically respectable attempts to translate lead to indeterminacy, we

have no reason to think that there is anything objective that they are tracking. And as translation

attempts to preserve meaning, it seems that the indeterminacy of translation shows that there is

nothing objective about meaning (Quine 1960, 73). Relativism thus is not even protected as a safe

harbour from the nihilistic tendencies of translational indeterminacy on Quine’s account. Quine,

oddly, remained a scientific realist despite his arguments about translational indeterminacy, but it’s

really unclear how these two positions are consistent.7

Many have rejected Quine’s thesis outright without much argument,8 but few have attempted

to demonstrate the source of Quine’s error. For instance, Hilary Putnam has rejected Quine’s thesis

outright as based on the dubious premise that there are no conceptual truths (Putnam 2001 [1992]),

but his most systematic approach to the issue of the relationship between culture and knowledge is

not too different from Kuhn’s. Putnam draws a distinction between metaphysical realism, which holds

that the world is the way it is independently of our interests, and cultural relativism, which holds, on

his account, that the world is exactly the way a culture says it is. In between there is room for what he

9
called “internal realism.” In internal realism, truth comes about through a combination of our interests

and the brute makeup of the world. The advantage to this position is that it allows us to recognize the

existence of artefacts, which can only be accounted for by reference to our interests and perspectives,

but it does not have to descend into the absurdity of cultural relativism (Putnam 1987, 1-39, 1988,

114; cf. Putnam 1982).

I find this response unsatisfactory. Perhaps a metaphysical realism that denied reality to

artefacts is a bad thing, but is there any significant distinction between internal realism and cultural

relativism? A cultural relativist does not have to hold that any old conviction in a culture is true, only

those that are sanctioned by the relevant cultural authorities, and those underwritten by its mores.

Putnam’s famous view on meaning is that it is a cultural phenomenon, underwritten by a “linguistic

division of labour,” which apportions important, deep questions about meaning and truth to relevant

experts in a linguistic community (Putnam 1975, 144). If languages are different, and surely they are,

and if cultures are different, and surely they are, then it seems that what we have in internal realism is

a rather sophisticated form of cultural relativism. Putnam is famous for changing his mind, and thus

we should probably not hold him to his past views. But his most recent writings appear to be even

more relativistic in some respects and even less sensitive to issues of objectivity and translation.9

The realist might wonder why I should be bringing this up: so what if science is culturally

relative? We still have good reason to think that it’s objective. After all, it allows us to build nuclear

bombs, send people in to outer space, construct useful technologies, and cure diseases, to name a few

coups. Where’s the room for doubt? Science may be culturally relative, but all this shows is that some

cultures are more advanced than others, and are more onto the facts of reality, while others are stuck

in some stunted stage of human development.

This response is seductive, I think. The trouble with this response is that it assumes that there

is some type of independent measure against which the epistemic advances of a culture can be judged.

If meaning, and hence knowledge, is culturally relative, there is no such measure. Sure, some cultures

10
can blow things up with bombs, but in order for this to be a sign of objective, epistemic advances, we

need some way of independently assessing the question, and those cultures that eschew such

diabolical ways could simply scoff at our claims to objectivity: rather, from their perspective, all of

our technology might be nothing more than the toys of the devil.

Donald Davidson has attempted to dismiss worries of translation as no threat to objectivity.

He affirms the indeterminacy of translation, but he argues that this fact about differing languages

should worry us no more than the fact that we can measure temperature in Celsius and Fahrenheit.

(Davidson 2001 [1977], 224-225, 1986, 313) Davidson’s response, though ingenious, doesn’t lessen

the threat of the indeterminacy of translation to objectivity. Davidson’s idea is that all languages

holistically deal with the same objective world, and thus any differences that they have among them,

though the cause for indeterminacy in translation, will be minimal and regional. But if this is so,

Western science, and traditional magical thought of an isolated culture, which both attempt to

“measure” the world, and are reflected in differing languages, would have no common arbitrator, or,

if one takes the objective world that all must live in as this arbitrator, then it seems that the world is

quite equivocal between the two approaches to measuring the world. Davidson as a person from a

culture that prizes science could believe that his way of measuring the world is better, as someone

who is a proponent of the metric system might think that the imperial system is inferior, but he

wouldn’t have a culture-independent reason for thinking so, and the members of the magical society,

like members of a society who measure the world with imperial units, would have no culture-

independent reason to switch. They may end up switching because of practical pressure from the

outside, but this would not be an epistemic consideration, just as a society of metric fans that adopts

the imperial system because of pressure from the outside world is responding to a practical pressure

and not an epistemic pressure that derives from some intrinsic superiority of the imperial system.

Here, the analogy of measurement shows itself to be off the mark. Differences in conviction are not

simply about how we carve out linear progressions, but are rather explanatory, and they often take us

11
in very different directions. The analogy between measurements assumes some type of common

reference against which changes can be calibrated. Davidson believes that he shows us what this is:

namely, it is truth. We start with assuming the truth of claims people make, relative to their

environment, and attempt to systematically understand their utterances from there. The common

world thus presents us with the independent measure to judge claims by. However, Davidson also

recognizes the holistic nature of natural languages (that, for instance, what a sentence means is a

function of many other facts about a language) and this will lead to internal differences among

languages that can lead people speaking differing languages to diverge on the truth values that they

assign to sentences, relative to the common world—differences that come to light once they have

been able to “translate” each other’s language. Thus, the thermometer ends up being relativized on

this story too. Davidson himself believes that his account reduces the threat of the indeterminacy of

translation. But even if this is so, it remains a problem that can’t be eliminated on his account.

This might be too fast. Perhaps people with differing languages, after recognizing that they

are having conflicting results in translation, can work out together the common semantic differences

between the languages and come to realize that in reality, their languages simply measure the world

with differing units. Well, if they could work this out, translation between their languages wouldn’t

be indeterminate any more. But, on Davidson’s holistic account, it is. One might think that this not

such a problem.10

It is important to note that bringing heavy metaphysics into the picture does relatively little to

soften the blow of cultural relativism and problems of translation. We might think, as some authors

appear to, that what we need is an account of meaning that appeals to metaphysically serious matters

(possible worlds, perhaps, or ontological particulars, abstract intensions, natural kinds and the like).

Such metaphysical referents and supports of language are by definition extra-linguistic and thus it

seems that we would have the right type of bridge we need to overcome cultural differences and

translate the claims of science. All we would have to do is line up expressions across languages that

12
have the same metaphysical underpinnings and, on this basis, we could bridge the cultural divide.

Translation would thus be a simple word for word, expression for expression exchange across

languages that are synonymous and guaranteed to be synonymous by virtue of their shared

metaphysical meanings. Here’s the problem with big metaphysics: if translation is a linguistic affair,

then translation is at the mercy of the actual semantic resources of a language. Our language, for

instance, might have devices for a certain type of meaning, but there is no credible reason to think

that all languages must share such big metaphysical foundations.11 Particularly, if we are enamoured

with the epistemic accomplishments of our culture (say, science) and we find this lacking in another

culture with a different language, we have very little reason to think that their language has the

semantic resources to articulate our science.12

What is far more plausible, and supported by the social scientific data, is that the semantic

profiles of natural languages are quite different. Indeed, if one actually attempts natural language

translation, as opposed to thinking about it from an armchair, what one realizes quickly is the notion

that there is some type of common semantics underlying all human languages is quite implausible.

The meaning of words, real words, develop in historical contexts over time, according to forces that

do not admit of one easy explanation (cf. Traugott and Dasher 2002; Sweetser 1990). Even if we

assume that natural selection presents itself as a basic pressure that all languages must survive, this in

no way exhaustively determines what meanings, concepts or ideas a culture must have to get along.

If there is any doubt, all we need to do is try to construct bilingual dictionaries. Very few words

across languages admit of easy translation. This is most striking in the case of languages from

disparate geographical and historical contexts, such as English and Sanskrit. Both are Indo-European

languages, and Sanskrit may be the oldest surviving such language to date, but we need to go back

several millennia before we come to a common historical antecedent for these two languages and this

shows in the efforts of scholars to match up words across such languages. The polysemy and

indeterminacy of meaning at the boundaries of languages is a fact turned up by scholarship.

13
 Why should natural languages matter? Because these are the systems that express cultural

differences. If contemporary science is to be vindicated as not simply an expression of a culture, it

should be in principle possible to translate a work of science into any language. But, however, if

meaning is relative to a language—metaphysical as it might be—it is an empirical question as to

whether languages have all the same meanings. In fact, it is highly likely that Sanskrit lacks many

necessary semantic devices to articulate contemporary scientific theory. (For instance, while Sanskrit

has a word for “atom,” namely “anu,” it has no term for the notion of an electrical “charge.”13) But if

this is so, Sanskrit’s inability to articulate contemporary scientific theory is an affront to science’s

claim to cultural-transcendence and objectivity. This is a fact: natural languages are historical and

cultural artefacts whose semantic commonalities are usually rough and contingent. To deny this is to

retreat into an arm chair world of fantasy.

My view is not that the texts of contemporary science are culturally relative, or that they are

untranslatable into Sanskrit, or any language, but the problem I am trying to highlight is this: The

contemporary philosophical scene’s difficulties with the topic of translation, borne out in some

instances in conclusions that translation is indeterminate or problematic, or in other instances by a

total avoidance of explaining how translation is possible in light of the reality of linguistic and

cultural difference, poses a threat to our ability to understand how knowledge as such can transcend

culture and language. It fosters an environment that is friendly to foggy thinking. Some philosophers

have attempted to vindicate the objectivity of ethics by arguing that it is just a species of scientific

knowledge (cf. Boyd 1988; Brink 1989, 132; Railton 1986, 204, 1996). However, if translation is

indeterminate, science is really no better off than what many philosophers take to be the plight of

ethics. Scientific and ethical knowledge would thus be no more than local, culturally contingent

knowledge, which lacks objectivity as it is a mere expression of culture.

14
 I.2. The Institutional-Textual Turn
So what is the guilty thesis? The pervasive view, buttressing most all research in the

humanities today, is that meaning is ultimately linguistic. This is the so-called “linguistic turn.” I take

this to be an extremely pervasive paradigm in philosophy in the Western tradition that still has a hold

on the way philosophers think about philosophical problems, though they may not themselves readily

advertise or admit the linguistic underpinnings of their work. Michael Dummett has famously argued

that this turn begins with Gottlob Frege, who held “the belief, first, that a philosophical account of

thought can be attained through a philosophical account of language, and, secondly, that a

comprehensive account can only be so attained” (Dummett 1994, 4). Dummett recognizes that

Edmund Husserl, a contemporary and correspondent of Frege, shared much with Frege in

philosophical interest and that at the time of their work it would have been difficult to distinguish

these two thinkers as initiating differing schools of thought (Dummett 2001, xviii). Frege is

commonly thought of as the grandfather of the Analytic tradition of philosophy, and Husserl of the

Continental tradition. Husserl, however, does not explicitly take the linguistic turn as Dummett

understands it, and thus we might have reason to think that the linguistic turn is a peculiarly

“analytic” phenomenon.

I think this characterization of Frege’s move, while historically accurate, hides the semantic

significance of the thesis. Thoughts, for Frege, are third world, Platonic entities (Frege 1988 [1918-

19]), but they also correspond to what we often consider today to be propositions. Propositions are, in

a certain theoretical orientation (such as Frege’s), the ultimate semantic entities. Thus, Frege’s move

was to look to language to elucidate the nature of meaning. This, I would recommend, is how best to

understand the linguistic turn. So understood, it is not by any means a move that peculiarly

characterizes Analytic philosophy. According to Cristina Lafont, this move can be traced back to

eighteenth century German philosophy in the work of Johann Georg Harman, Johann Gottfried von

Herder, and Wilhelm von Humboldt, and is taken up again in the work of Martin Heidegger

15
(Husserl’s famed student) and Hans-Georg Gadamer (Heidegger’s famed student) (Lafont 1999). If I

were to trace the genealogy of this movement, I think we could also find roots in Kant, who, in the

Critique of Pure Reason, made both concepts (expressed by words) and judgments (expressed by

sentences) front and centre in semantics. This approach to meaning in the Western tradition may even

stretch back to the Socratic concern for definitions as a means to solving philosophical problems. Any

account of semantics that takes the paradigm, determinate cases of meaning to have a structure

delimited by linguistic categories (such as words or sentences) is part of this paradigm in my view.

On this way of understanding the linguistic turn, it is both very prevalent and not something very

new. Philosophers who work only in the Western tradition might be led to think that there is some

type of huge gulf between the analytic and continental traditions. As someone who has an active

research project in a major non-Western tradition of philosophy, the differences between so-called

Analytic and Continental approaches to philosophical problems strike me as quite internecine.

Analytic philosophers tend to look to science for inspiration in their thought on language and

meaning. Continental philosophers tend to look to literature and arts for inspiration in their thought on

language and meaning. That’s about the difference.

An inevitability of the linguistic turn is that meaning in its most formal or determinate form is

treated as a property of languages—usually its words and sentences, sometimes more broadly its

grammar. Translation, in turn, is conceptualized as a linguistic operation or conversion. Just as the

linguistic turn makes languages out to be the substratum of meaning, it treats languages as the objects

to be translated. I call these theses that result from the linguistic turn, respectively, the linguistic

account of meaning and linguistic account of translation. These theses are really two sides of the same

coin though they don’t always take a bow together. I call this coin the “linguistic paradigm.”

In the place of the linguistic paradigm, I think we should embrace what might be called an

institutional-textual turn. I shall call the account of semantics involved in this shift Text-Type

Semantics, or TTS. TTS does not reject the meaningfulness of language, the practical utility in

16
appeals to linguistic meaning, the importance of language in human life, or the role that culture places

in shaping people’s thought, opinions and projects. Neither does it reject the notion of linguistic

competence and that there are better and worse ways of expressing oneself with language. Indeed,

TTS takes very seriously the notion that languages and cultures are historically evolving phenomena

that can be studied by the social sciences. It also takes very seriously the notion of real and radical

cultural diversity.

However, TTS treats determinate meaning as a property of texts of specific types. A text-

type, on this account, is like a genre, except that it has an institutional underpinning, and it primarily

concerns texts, which are paradigmatically bounded artefacts, with beginnings, endings and some

place for authorship. We could also use the term “discourse-type” as a synonym for “text-type,”

though the notion of a text serves rather more clearly to emphasize the extended and bounded nature

of the semantic phenomena TTS paradigmatically tracks. I take the notion of the text-type from the

translation studies literature. However, the role it plays in TTS differs greatly from its typical

conceptualization in the translation studies literature. There it is often treated as a linguistic

phenomenon. (Indeed, it has its origins in contrastive linguistics.14) I think this is a great mistake. To

conceive of text-types as linguistic phenomena is to make them incapable of showing how

translational determinacy is possible. The utility of the text-type is its ability to mediate between

differing semiotic systems, not all of which are linguistic. (Texts can be composed of icons, for

instance, pictures or symbols. The earliest texts of humanity were likely pictographic. [cf. Daniels

1996, 3.]) We continue to employ such nonlinguistic means to compose texts. Think for instance of a

pie chart, labelled with pictures, which represents data that could be determinately translated into a

table with numbers. There is also a long tradition in Indian philosophy of writing texts in the s tra

format, which is almost a nonlinguistic manner of composing philosophical texts by employing words

primarily as ambiguous symbols instead of grammatical devices.) TTS urges us to shift our thinking

about translation in a manner that results in viewing linguistics as a matter of secondary and

17
contingent importance to translation and meaning. This is necessary. Languages are ultimately

expressions of culture, which are contingent, indeterminate phenomena that shift and change

according to various factors, including, and perhaps most importantly, translation. Meaning is

something far greater than language, and this is necessary to recognize for us to avoid the descent into

cultural solipsism that has characterized Western philosophy since the outset of the linguistic turn.15

The notion of an “institution” is polysemous and ambiguous—indeed, polysemy and

ambiguity is the stuff of language according to TTS. The term “institution” is at times used for

important cultural practices. However, it also has another direction in meaning, one concerned with

strong and resilient social practices (Scott 2001, 48) that can transcend cultures and can be

international and transcultural in nature (Miller 2007). The type of institutions relevant to TTS does

not have anything essentially to do with robes, head dresses and big buildings. Modern medicine,

mathematics, physics and literature, for instance, are not identical with the cultures of their birthplace

(in many cases, there are multiple cultural birthplaces). Rather, they can be studied and practiced

anywhere in any language and the texts that underwrite these practices can be written in any

language. Their translatability is guaranteed by the formal features of the text-type that defines these

various institutions. The text-type specifies features, or text-type features, of an original text that must

be preserved in translation. Thus, the job of translation is not to reproduce a text that is like an

original in all respects, but to produce a text that is identical with respect to text-type modalities.

When translation proceeds according to such institutional considerations, and there is no resulting

controversy among the relevant experts trained in the text-type and the subject matter under

consideration, the translations are determinate, and serve as a normative point of reference for the

very functioning of the epistemic institution and for language use. Controversies in translation of texts

of a definite type, if they should arise, are always resolvable by institutional means, I shall argue.

Linguistic differences pose no obstacle to translation for linguistic meaning is indeterminate to begin

18
with. Experts add to the polysemy of a language by introducing new terms or reusing old expressions

in new ways to facilitate translations.

Because text-type institutions are social practices, it is possible for them to change over time,

but the changes are slow, and have an important effect on the text-type they preserve. For instance,

philosophy is one such text-type, supported by a specific institution, which has undergone changes

over time. Much that is considered today as science was in its inception philosophical. As research

into these matters became so specialized that they could no longer be supported by the text-type of

philosophy, they required their own vectors of meaning, translation and institutional definition. Thus

new textual institutions were born, and old avenues of inquiry that appeared live came to be closed-

off for philosophy. For instance, psychology was shed from philosophy in this fashion, and while

some philosophers have projects that overlap with psychology, their work as philosophers is typically

informed by a clear appreciation for the distinction between philosophical analyses of issues of

interest to psychology, and the clinical inquiry into the subject. When philosophy gives birth to a new

science, the text-type of philosophy is sharpened.

TTS invites us to make a paradigm shift in how we think about meaning, and this shift is

profound, given the linguistic orientation of philosophy and the wider intellectual tradition for over

the last century in the West. A chief virtue of this account is that it is naturalistic — that is, it looks to

actual translational practice, and the practice of our epistemic institutions, as providing the raw data

for semantics. It is corroborated by my work in South Asian Studies and Indology. However, the

results of adopting this naturalistic approach to translation and meaning are that we must give up the

notion that there is one story to tell about how to determine meaning, over and above a very general

account that defers to textual institutions and their text-types. Thus, Naturalism as it is usually

conceived of in the philosophical literature as a semantic thesis about the role of science in

underwriting meaning is out.16 Indeed, I shall argue that my naturalistic approach to understanding

19
meaning and translation leads us to recognize the autonomy of the semantics of philosophy and

philosophical issues, and I think it also takes us to a philosophical realism, in all areas of philosophy.

Each text-type institution that preserves, and refines, a text-type is its own fountain of

knowledge for its text-type features provide the concepts and principles to determine its special type

of knowledge. I take this to be an extremely advantageous result of TTS. It sharpens not only what

philosophy has to offer, but also what literature, the social sciences, and the natural sciences have

each to offer. Such clarity is indispensable today, in a time when various pernicious forces seek to

undermine the results of our epistemic institutions, and degrade the natural sciences to the mere effort

to interpret the evidence or provide an account of the facts “that makes sense” to us—as though

knowledge was a democratic issue that each of us had to decide for ourselves. For too long we have

humoured the idea that knowledge must be amenable and flattering to our personal outlook—a type

of toleration for half-baked intuitions that is an inevitable outcome of attempting to understand

meaning in terms of language, which is defuse and intelligible to all in a speech community. The

move to recognize the autonomy of our various epistemic institutions that TTS underwrites also leads

us to appreciate the proper place of philosophy. There are some questions that only philosophy can

answer, and we lose sight of this if we think that there is one grand story to tell about knowledge.

TTS thus protects and clarifies the contribution that our various epistemic institutions can yield, but it

also contradicts the characteristically messianic approach to knowledge and meaning that

characterizes so much Western thought, and analytic philosophy. Knowledge is as diverse and

pluralistic as determinate meaning is. Science has no place of privilege in knowledge.

I.3. Antecedents
The view that I shall defend is that normative and evaluative discourse characterized by terms

such as “good” and “ought to” is really a species of the text-type of philosophy and thus to translate

normative and evaluative discourse is to translate philosophy. Terms such as “good,” “ought to,” as

well as terms such as “real,” “knowledge,” “beauty,” “wrong,” “right,” “logic,” and “meaning” (in

20
English of course) are key philosophical terms and the conceptual content of these terms as markers

of the philosophical text-type is set out by what I shall call QI. QI is thus an account of a certain type

of concept, namely philosophical concepts, but these concepts are not reducible to features of a

language, rather they characterize a certain type of text defined by its translatability into any semiotic

system. Key philosophical terms are what I call “text-type features.” The presence of such terms or

symbols bearing the conceptual content of QI (say, in a symbolic or graphical text of philosophy) is a

textual point of reference. The translator of the philosophical text attempts to understand the various

meanings of a text in light of these concepts in order to produce a successful translation of a text as a

text of philosophy. These concepts are incarnations of the very text-type of philosophy, which I

submit is concerned with the articulation, application, debate, criticism and investigation of universal

and general theories on a varieties of topics, on the basis of considerations that are thought to be

objectively persuasive. Each philosophical concept instantiates this structure, but with a distinct

personality. Thus, in translating a text in light of these concepts, we translate a text in light of the very

textual structure of philosophy.

In a dissertation it is common to note the historical antecedents of the view one is putting

forward. One cannot make a contribution to scholarship without being cognizant and respectful of the

current and past state of scholarship. Unfortunately, with respect to the topic of normative and

evaluative discourse translation, there has been very little significant work.

Part of the problem with the scarcity of work on the topic is that the very notion of translation

is foggy against the backdrop of the linguistic paradigm. There are some cursory comments on the

challenges of translating ethical discourse in the metaethics literature (as we shall see in chapter 4),

but these are incredibly naïve and not based upon any serious or systematic thinking about translation.

But the metaethics literature cannot be faulted too greatly for the recent Western philosophical

tradition has a very problematic relationship with translation. Translation theory continues to be of

some interest outside of philosophy in the interdisciplinary field of “Translation Studies,” though the

21
empirical and applied nature of that interdisciplinary enquiry in some respects constitutes a rejection

of theoretical approaches to translation, and this can only be explained by reference to the failure of

the linguistic paradigm to explain how translation can be determinate.

The only scholar that I know of who has done significant work on the problem of translating

normative discourse, particularly in light of the challenges of cross-cultural research, is me. My

interest in this topic began when I took a “break” from philosophy as a graduate student and was

enrolled in the MA program in South Asian Studies at the University of Toronto. I became interested

in the view among many eminent Indologists that Indian philosophers were not very interested in

ethics or moral philosophy. To address this problem I began looking for literature on the topic of

translating normative discourse in this interdisciplinary field, spanning subjects such as the Religion,

Religious Ethics, Anthropology, Linguistics, Sanskritology, Philology and Philosophy. However, I

found nothing that could help me. The problem was not that scholars were not interested in studying

moral theories cross-culturally, but that the thinking on these matters was very crude and naïve. For

instance, typically, scholars of Religious Ethics (and, for that matter, even most analytic philosophers)

tend to treat any norm, prohibition or value concerning action across cultures as moral in concern.

Thus, for instance, if it is apparent that a culture has a prohibition against murder, this is often thought

to necessarily be a moral prohibition. But if we are really interested in cross-cultural scholarship and

the possibility of radical difference on values, we must recognize that people from diverse

backgrounds may have diverse perspectives on life and value, and that while a prohibition on murder

in one culture may indeed be part of their moral theory, in another culture or sub-culture it could be

part of their aesthetics (“murder is wrong because it is ugly”) and yet for others it could be part of

their soteriology (“murder is wrong because it keeps one in Earthly bondage”). These are logical

possibilities and thus a scholar interested in the study of ethics across cultures must keep an open

mind to such variability in axiological perspective when comparing cultures and their literature. But

this type of sensitivity was rarely to be found.

22
 This type of axiological naivety is really part of a larger problem with the literature that I

encountered: it invariably presumed that what counts as a moral prohibition or positive prescription,

or value, is not philosophically controversial within and across cultures. Thus, scholars studying

ethics in the cross-cultural context usually picked a conception of the ethical and categorized beliefs

and opinion across-cultural contexts on the basis of whether it accorded with their own conception of

what ethics or morality is about. But in cross-cultural research we are not interested in whether the

world agrees with us, but in the interesting and variety of ways that real people have differing views

on morality across cultures. If were not interested in such diversity, there is very little motivation to

study ethics cross-culturally.

The trouble with our ordinary way of thinking about ethics is that it assumes that it is an

autonomous area of thought, disconnected from philosophical controversies on the whole. But in

reality, the philosophical views that people take on a variety of matters are often directly related to

their substantive views in ethics, and thus one cannot study ethics cross-culturally without

understanding that what it is to learn about the moral outlook of a person or text is to understand it as

varying systematically according to various philosophical controversies. Some are metaphysical (such

as what beings, properties, actions, or character traits count as ethically relevant), and other such

controversies are epistemic (how we come to know what is morally right). Yet still there are

interesting axiological controversies that bear upon determining a person’s moral outlook (such as,

whether moral norms override other norms).

In the case of Indology, I found that scholars conformed to this general trend in scholarship of

not recognizing that ethical issues are philosophically controversial. Thus, they typically studied

Indian philosophical texts by reference to their own latent conception of morality. All scholars agreed

that there is a term in the Indian tradition that means “ethics” if any term does and this is “dharma.”

However, given the widespread provincial approach to morality, Indologists translated “dharma” as

23
“ethics” or “morality,” when it suited their moral outlook, and declared that “dharma” had an

innumerable number of meanings when it did not.

My solution to this problem was to argue that “dharma,” like “real,” or “ethical,” is a key

philosophical term and that the function of such terms is to articulate philosophical perspectives. I

also argued that “dharma” and “ethics” have the same philosophical function and are thus synonyms.

According to the view I put forward, the meaning of “dharma” was not determinable by reference to

the linguistic data alone, but rather had to do with its function in philosophical discourse. On this

recommendation, scholars were to treat all occurrences of this term in a philosophical text as

articulating a moral theory, recognize each designatum of this term as it was employed in particular

passages and sentences in a text as identifying some item that falls within the putative extension of

the concept of DHARMA/ETHICS, by the author’s lights, understand the philosophical theory of ethics

that emerges from this process particularly in light of other convictions of a philosophical variety we

have sound scholarly reason to attribute to an author, and then read and translate the text in light of

this theory. Of course, we could take in all manner of philosophically important information, such as

the sectarian allegiances of an author, and whether they saw themselves as merely an exponent of a

wider philosophical project or whether they thought of themselves as offering a philosophical

innovation. To understand a text this way was to be committed to understanding it in terms of the

philosophically relevant features of the text and thus any philosophically relevant information could

be factored in. Successful translations of Indian philosophical texts would thus reflect the

appreciation for the philosophical perspective of the author, and not one that was filtered through the

substantive philosophical convictions of the translator and scholar. All of this presumed a

background and working knowledge of Sanskrit: not enough to solve translational controversies, but

enough to start on the task of discerning the philosophical theories of an author given a sufficiently

robust understanding of Sanskrit to yield multiple prima facie readings that the approach to studying

philosophical texts that I was urging could decide between.

24
 This was the beginnings of what I now call QI (though at the time I put this theory forth, it

was not forwarded on the basis of a worked-out theory of meaning and translation). As a result of this

understanding of “dharma” I was able to provide an account of the history of Indian philosophy that

turned conventional wisdom on its head. In light of this account, not only was ethics an important

topic for Indian philosophers, but four out of the nine canonical schools of Indian philosophy take

ethics to be their foundational concern and understand their entire philosophical project as one of

moral philosophy. Moreover, the philosophical diversity that is present in the Indian tradition,

particularly on moral issues, is remarkable on this account, thus providing prima facie cross-cultural

evidence that the approach that I am urging is hitting at something objective in cross-cultural research

and is not a mere function of the scholar’s projection of his own moral values on his historical

subjects. My MA thesis, Ethics and the History of Indian Philosophy was published by the leading

Indology publisher, Motilal Banarsidass (2007a), and as a result of this ground work, and my work in

translation theory since, I have a translation of a canonical text of Indian philosophy (Patañjali’s Yoga

S tra) in press with Penguin to be published through their Black Classics Series (2007b). To my

knowledge, this is the first principled translation of this text that can be underwritten by a sound

theory of translating philosophy, which I shall argue for.

During the course of completing that thesis, I argued that there is a distinction to be drawn

between what I called “formal” terms, and “substantive” terms. This distinction corresponds, roughly,

to the distinction between thick and thin concepts that one finds in the contemporary metaethics

literature. QI, as I shall argue for it, draws a distinction between thick and thin philosophical concepts

that do the work of buttressing philosophical discourse, and thus the literature that touches upon the

thick-thin concept distinction is a type of scholarly antecedent to the account I shall put forward.

The metaphorical distinction between the “thick” and the “thin” has been invoked for decades

in analytic philosophy. Some uses do not have any explicit or intended connection to ethics, though

25
usually the “thick” designates a non-abstract matter, while the thin concerns matters of abstraction

(Aldrich 1971; Hodes 1990; Azzouni 1997).

The thick-thin distinction made a grand entrance into the ethics literature in Bernard

Williams’ Ethics and the Limits of Philosophy (1985). Williams argues for a distinction between two

types of ethical terms against the backdrop of what he calls “prescriptivist” semantics. One type of

term combines prescription with description, while the other is purely prescriptive (Williams 1985,

130). Alternatively, he describes one type of concept as marrying world-guided elements and action-

guiding elements (Williams 1985, 141) . He calls the former “thick” or “thicker” concepts.

Elsewhere, he notes that thick ethical concepts have more empirical content (Williams 1996 [1995],

25-6). Paradigm examples of thin terms include “good,” “right,” and “wrong,” while thicker terms

include “courage,” “kindness,” and “generosity.”

Williams (1985, 130) suggests that the distinction can be found in the work of R.M. Hare.

Hare in the Language of Morals (1952) draws a distinction between terms such as “good” and terms

such as “industriousness.” The former, he argues, are best thought of as primarily being evaluative in

nature, and secondarily as descriptive. Hare’s astute observation is that terms such as “good” do, by

way of usage, acquire descriptive associations, but these he argues are secondary for they are not

invariably connected with all usages, while the evaluative is fundamental. In contrast, terms such as

“industrious” have an evaluative meaning, but this is secondary to its descriptive meaning (Hare

1952, 121).

C.L. Stevenson may have been the first philosopher in the so-called “non-cognitivist”

tradition in the West to recognize that ethical judgments can combine descriptive and evaluative

components. Stevenson also, and before Hare, distinguishes between evaluative terms that have

descriptive content on occasions of usage only, and those that have it more conventionally. He also,

to my knowledge, was the first author in ethics to use the term “thin” to characterize the former, more

abstract manner of ethical meaning (Stevenson 1944, 94). From what I can tell, the thick-thin

26
distinction as something explicitly labelled with the metaphorical access of thickness begins with

Stevenson.

While Williams defers to Hare as the origin of the distinction, the distinction between thicker

and thinner terms as it plays out in Hare and Stevenson’s early writings is both similar to and distinct

from Williams’. Williams is loath to admit that thinner moral vocabulary have descriptive content,

while both Hare and Stevenson recognize that usage provides it with such content. However, their

account is also similar to the extent that they hold that some moral vocabulary is best explained as a

combination of descriptive and nondescriptive components. For this reason, we might call the

distinction identified by Williams the Williams-Hare-Stevenson account, or the two-components

account of ethical meaning. On this account, thick concepts are a combination of two distinct

components (a non-cognitive meaning and a descriptive meaning), while thin concepts are composed

of just an evaluative component.

The two-components account is likely the most common account of the distinction (cf.,

Gibbard 2003a, 304), and it is certainly the conception against which alternative accounts are

provided. It is a favourite manner for accounting for the meaning of pejorative terms that are, on the

usual account, thick in nature (cf. Dummett 1973, 454), but it is quite an implausible account on this

score. For instance, according to Stevenson, “Nigger” means “Negro, Bah!”— (Stevenson 1944, 82).

I am sure most African Americans will find this a very truncated account of the significance of this

slur. The two meanings account is also a favourite of naturalists, who often argue that empirical

criteria of usage and evaluative criteria cannot be easily pulled apart. Philippa Foot made similar

observations about “rude.” On her account, it is a value term, but it has descriptive conditions of its

acceptable application from which the evaluative component can be derived (Foot 1978 [1958], 104;

cf. Anscombe 1958).

A related but critical account of the thick-thin distinction is the anti-disentanglement thesis,

or the anti-splitting thesis. This thesis is parasitic on the two-components thesis for it recognizes a

27
conceptual distinction between evaluative and descriptive meaning (that is, we can imagine and talk

about the distinction), but on this account there is no actual semantic distinction in actual cases of

moral terminology usage. To attempt to split the meaning of moral terminology into two, on this

account, is to destroy the meaning (Putnam 2002, 38; cf. McDowell 1981, 145; Murdoch 1970).

There is a strong Wittgensteinien underpinning to this account, which holds that moral meaning is

embeded within a linguistic practice.

A varient of the anti-disentanglement account is proposed by Jonathan Dancy, who rejects the

distinction between the evaluative and descriptive at the conceptual level. On his account,

“[e]valuation and description are indeed distinct as speech acts, but that distinctness does nothing to

establish a distinction between evaluative and descriptive properties, meanings, or anything else

much” (Dancy 1995, 272). Thick concepts, on Dancy’s account, provide a range of meanings that

anchor their usage and provide a rough picture of what one encounters in various contexts, without

exhaustively describing the referents (Dancy 1995, 271).

Another critical stance on the distinction holds that there is no real principled way, even on

linguistic grounds, to divide moral terminology into thick and thin categories—normative language

appears on a number of dimensions on this account (Scheffler 1987, 417-8). Call this the anti-thick-

thin distinction view.

Simon Blackburn also holds a critical view of the thick-thin distinction. Blackburn takes the

anti-disentanglement picture of thickness as a departure. On his account, thick concepts, if there were

any, would present us with a perfect amalgam of descriptive and evaluative components such that

they could not be split without destroying the meaning, However, on his account, what we find in

putative cases of thick concepts is nothing semantically settled. Rather the descriptive content of such

terms vary with usage and cannot thus be said to be semantically based (Blackburn 1992).

Gopal Sreenivasan (2001, 14) traces the distinction we find in Williams to the anthropologist,

Clifford Geertz’s famous distinction between thick and thin descriptions. Geertz actually did not

28
invent this distinction. Geertz (1973, 6) finds it in Gilbert Ryle’s Collected Papers (1971†).

According to Ryle, thin descriptions of events do not make reference to the intentions of actors and

only include descriptive content, while the thicker descriptions do include an attitudinal and

intentional component to their description (Ryle 1971b [1968]; cf. Ryle 1971a [1968]). The Ryle-

Geertz distinction actually inverts the relationship between descriptivity and thinness in comparison

to Williams and Stevenson’s account: on the Ryle-Geertz account, it is the thin descriptions that are

purely descriptive, while the thicker descriptions make reference to intentions, including attitudes. As

an elucidation of Williams’ distinction, the Ryle-Geertz distinction seems to miss the mark. However,

Ryle was Williams’ teacher, and it may be that the metaphor of the thick and thin made its way via

Ryle to Williams17—though, as we’ve seen, it can be found much earlier as it appears in Stevenson’s

early and important work.

The most original linguistic account of the distinction in the literature to date is provided by

Christine Tappolet. On her account, thick concepts are best explained by reference to the two-

components of the Williams-Hare-Stevenson model. The term “good,” however, is not best explained

as simply being the evaluative half of a thick concept. Rather, it is better understood as being

determined by affective concepts, just as the concept red is determined by red objects. Tappolet’s

suggestion is that we should reject the thick-thin distinction as Williams’ paints it, and return to the

more traditional distinction between general and specific concepts. “Good” would gain its meaning

by virtue of more specific value concepts, even if they were not thick in Williams’ sense (Tappolet

2004)

An honourable mention in this discussion goes to Stephan Burton, who argues that thick

concepts are best understood as a combination of evaluation, and a description that is criterial in

nature, thus providing a type of reason or justification for its application (Burton 1992). This

approach to thickness is quite distinct.

29
 Very well. Which of these accounts of the thick-thin distinction is correct? I do not think that

this is a sensible question. The accounts that we have been examining are all linguistic accounts of the

thick-thin concept distinction. They aim to elucidate the meaning of a class of terms in specific

languages. The notion is that there is some real fact of the matter as to which of these competing

accounts is the right account. But I think there is no good reason to take any one account to be more

correct than the other. This is not to say that such accounts are completely wrong or that they play no

role in scholarship. They can and do function as observations of different aspects of the polysemous

nature of key philosophical term usage in some languages (English for sure). They can thus function

as tenuous evidence that some terms function in a corpus as key philosophical terms of a normative or

evaluative nature when scholars move to the cross-cultural context, but the considerations that

involve identifying such terms will vary according to the cultural setting one is studying. My Ethics

and the History of Indian Philosophy sifted through large amounts of textual material that was

particular to the Indian setting to make a case for “dharma”’s functioning as a key philosophical term.

What clinches an account of whether a term has this function is whether it solves an outstanding

problem of scholarship, I think. A culture’s philosophical history will likely be the last thing that is

truly appreciated about it. The details that are relevant to determining such matters will not always be

the same and shift according to cultural contexts. The only invariant feature of such inquiry is the

text-type of philosophy.

The trouble with placing all of our bets on linguistic meaning is that language is our typical

means of social interaction. By virtue of this social usage expressions come to have meanings, but the

meaning they have is indeterminate, acquiring new dimensions with historical usage. It can be studied

empirically by linguists and philologists, but no such accounts are determinate for language use is

always subject to innovation and influence from external forces, such as translation. Many

philosophers think that they can specify the meaning of expressions in a language and that such

specifications get at some deep fact about a language. (I am reminded by these philosophers of the

30
old-fashioned grade school teacher who has very strong views about whether sentences in English can

begin with conjunctions or whether contracted words are meaningful. Or, worse, I am reminded of

unfortunate conservative criticisms of African American speech patterns, whose racist and ideological

basis is rather plain.18) I think we should be sceptical of the notion that there is such a thing as

determinative linguistic meaning, and so I shall argue.

Normally, this view on language would be seen as a declaration of post-modern uncertainty.

However, this view of language is only threatening if we operate within the linguistic paradigm. TTS,

in contrast, invites us to consider the role that our epistemic institutions play in grounding knowledge:

institutions that transcend cultural contexts because they are defined by textual concerns whose

essence is their translatability. Our best knowledge is determined by the semantics of such texts, not

language use at large. It is because there is a distinction between our institutions and language use at

large that there can be anything like culture-independent objectivity in knowledge. It is because of

such text-types that meaning, which is born locally, can comprise a text whose determinate meaning

consists in its translatability. Meaning is objective by virtue of its sublation within a text of a definite

type.

The account of the thick-thin concept distinction that I shall defend is thus not an account of

the meaning of terms in a language. Rather it is a formalization of the semantics of the philosophical

text-type---the text-type that underwrites the very institution of philosophy and the translation of

philosophical texts.

The semantics of the philosophical text-type is something that all competent philosophers

know and employ in their work. It is the standard by which our work is judged and in turn it is the

foundational norms that we employ in evaluating our students’ work.

As an undergraduate, I learned that there was a distinction between explication and criticism.

To explicate the text of a philosopher is to treat certain important terms as the focal point of

philosophical theory articulation. The terms in question, like “right,” “real,” “ethical” or “good”

31
concern differing philosophical matters. Moreover, I learned that to really explicate and grasp what a

philosopher is saying, one must not only understand how these terms relate to theories they might

articulate, but also one must understand the reasons and considerations that philosophers bring to bear

in support of such term usage. The reasons, in turn, are not simply subjective descriptions of likes and

dislikes. Rather, the reasons in question aspire to be objectively persuasive. Philosophers do not

always have theories. They may sometimes not have a theory, but employ a term like “good” or

“justice” in a dialectical search for a theory that satisfies certain considerations that they take to be

objective requirements on an adequate theory of the good or justice. Sometimes, they even make use

of such devices for purely deconstructive and critical projects, aimed at showing that on the basis of

the objective considerations, no philosophical theories are objectively correct. Understanding that

such considerations and reasons in favour of candidate theories are meant to be objectively persuasive

is the grounds on which criticism of philosophical texts is possible.

This in a nutshell is a very truncated summary of the philosophical text-type. It leaves out

many nuances that come to light in the translation of philosophy. For instance, philosophical texts can

sometimes articulate multiple perspectives, and they may even take the form of dialogues. A proper

appreciation and translation of such texts must keep track of the various theories and considerations

relative to the author of the whole text, or the characters in the texts, such as Plato’s dialogues. Either

way, the various moves and turns in philosophical texts are understood as a function of these various

variables, such as reasons, theories, and criteria according to which theories are selected.

I do not think that this is controversial at all. I would be surprised if any trained philosopher

teaching philosophy in an accredited university did not judge their student’s work by such lights. For

instance, if a student writes an essay on John Stuart Mill but attributes to him the theories and reasons

that we find articulated in the works of Jeremy Bentham by means of the term “good,” trained

philosophers would think that the student had failed to properly grasp the philosophical import of

Mill’s texts.

32
 An extremely important and distinguishing feature of the philosophical text is that the

theories and reasons that are found in such texts are articulated with key philosophical terms. A term

like “ethics,” just like “biology,” sets out an area of enquiry. But “ethics” unlike “biology” is the

manner in which philosophers articulate their competing views on ethics. “Biology” in contrast has no

text-type-theoretic role in the texts of science to articulate biological theories. Rather, “biology”

describes a type of scientific enquiry and the controversies that rage in biology have nothing primarily

to do with the proper usage of “biology.” Scientists from competing paradigms in science do not

typically feel the pressure to describe their opponent’s views in a manner that renders them

“unbiological.” Rather, they will usually attempt to argue that the opposing view is not the right

account of biological fact. In philosophy, in contrast, there is a great pressure to attempt to deprive

one’s opponent of the very right of even using a philosophical term. A Utilitarian, for instance, who

disagrees with a Kantian, is likely to say of some conclusion of Kant’s (say his view that we must

never tell a lie, come what may) that it is unethical—as likely as she is to say that Kant provides the

wrong account of ethics. This is no accident of language for it is a philosophical pattern that can be

observed across cultures. What we have at stake in the philosophical texts is the operation of key

philosophical terms as Quasi-Indexicals. That is, their operation defers to various contents that shift

according to perspective (such as reasons, and theories). Yet, such term usage also aspires to

transcend the perspective from which it is offered. I call the anatomical feature of the formalization of

QI that specifies this aspect of key philosophical terms the objectivity clause. The objectivity clause

concerns the reasons and considerations that a philosopher brings to bear in their philosophical texts

that are explicitly meant to transcend perspective and be objectively persuasive. This tension between

the perspectival aspect of philosophy and its claims to objectivity has often baffled relativists, but it is

a bafflement that comes about without appreciating its role in the very semantics of philosophy. It is

natural that relativists have typically not appreciated the semantics of the text-type of philosophy for

33
they have not thought about philosophical semantics in terms of what we need to understand in order

to translate philosophy. QI, in contrast, is developed with exactly this challenge in mind.

Key philosophical terms, as set out by QI, come in two varieties. Some are thin, others are

thick. Given the linguistic nature of so many of our texts, the expression of the basic thick and thin

model comes in two forms: one nominal and one verbal. There are many other differences too. Some

theories are normative/evaluative, which is to say that they are standards by which the world is

judged. In contemporary parlance, they have a certain direction of fit, namely a world-to-theory

direction of fit. In contrast, other theories are not normative, and they take the world as a minimum

standard of adequacy. Such theories have a theory-to-world direction of fit. The theories, moreover,

are selected for very different reasons. Ethical theories are different from aesthetic theories, which are

different from metaphysical theories, over and above differences of direction of fit. I call the criteria

that distinguish such theories the axiological differentia of a theory. Not all philosophical terms

articulate theories with clear axiological differentia. “Good” and “right,” for instance, do not. They

may be used without contradiction to articulate any type of normative or evaluative theory. All

philosophical theories carve the world into two: some aspects of the world fall in the theory, others do

not. The specific valence of the philosophical terms depends upon whether they are meant to refer to

items that fall within the theories they articulate, or whether they refer to items that fall out of such

theories. “Unreal” and “evil” refer to items that fall outside of the theories they articulate, though the

theories they articulate have different directions of fit.

Because we are linguistic creatures, we have evolved two distinct forms of QI: one nominal,

and the other verbal. I think that from a textual perspective there is no deep semantic difference

between these two forms. When authors think that the distinction is philosophically important, we

employ the relevant form in translation to the extent that it is demanded by the theories they

articulate. But from a canonical point of view, the so-called controversy over whether the normative

or evaluative is more fundamental is quite mistaken.

34
 Here is the nominal form of a typical thin term, such as “real”:

ordinary symbol s, refers to an item x that falls (within/outside) universal and general theory t that
has a (theory to world/world to theory) direction of fit, which is selected for (fill in axiological
differentia), in accordance with the relevant considerations.

The second form is calibrated for terms with a verbal function such as “ought to”:

ordinary symbol s, indicates that an argument a (satisfies/fails to satisfy) conditions specified by a

universal and general theory t that has a (theory to world/world to theory) direction of fit, which is
selected for (fill in axiological differentia), in accordance with the relevant considerations.

The basic thin model may also be altered to accommodate thick terms. Thick terms add to

the basic thin model what I call an extra-theoretic constraint:

ordinary symbol s, indicates that an argument a, that conforms to an extra-theoretic constraint,

(satisfies/fails to satisfy) conditions specified by a universal and general theory t that has a (theory to
world/world to theory) direction of fit, in accordance with the relevant considerations.

This is the verbal form. The nominal form takes on a referent:

ordinary symbol s, refers to an item x that conforms to an extra-theoretic constraint, which falls
(within/outside) a universal and general theory t that has a (theory to world/world-to-theory direction
of fit, which is selected for its (fill in axiological differentia), in accordance with the relevant
considerations.

Thick terms on this account are not comprised of a distinct type of meaning. Rather, just like

thin terms, they can be used in theory articulation. Their thickness simply consists in an extra

criterion that constrains usage over and above the basic thin model.

The question of whether a term is determinately thick or thin cannot be assessed, on this

account, by mere linguistics. Rather, one must consider the question in light of the work that a term

does in a philosophical text. However, if philosophy is particularly vibrant in a linguistic community,

the varied uses that terms take on by virtue of functioning as key philosophical terms will contribute

35
to the polysemy of a language, and this will in turn give rise to multiple significances associated with

philosophically interesting words in a community. Indeed, the history of moral semantics as a

linguistic enterprise from G.E. Moore, Emotivism, through to contemporary Expressivism and Non-

Analytic Naturalism all seize upon some aspect of the after effects of “good”’s philosophical

functioning, in English at least. This dissemination of philosophical meaning that in turn spawns

multiple linguistic significances characterizes “dharma,” I argued in Ethics and the History of Indian

Philosophy. The varied uses of such a term within a culture will thus seem extremely inexplicable to

the linguist, or give rise to interminable debates among linguistic philosophers, who do not keep such

text-type-theoretic considerations in view.

I.4. Looking Ahead

I split the task of making a case for TTS into two, slightly artificial parts. In Chapter 2 I

address the question of translational determinacy: how is it possible? The chapter addresses the

concerns of philosophers and translation theorists. On the basis of the practical challenges in

translation and the best insights from the translation studies literature, I make a case for the

indispensability of text-types to translation. Text-types, it is argued, overcome the problems of

translation as they are traditionally conceived for they provide us with a principled manner to decide

what features of an original text are to be preserved in translation. Most importantly, they allow us to

keep the indispensable, orthodox constraint on translation, namely that translations be equivalent to

their originals, without commitments to the implausible thesis that originals and translations must be

alike in every respect, or that languages are the objects to be translated. In the absence of the notion

that translation aims at equivalence that text-types help us make sense of, there is no determinate

constraint on translation. What real translators and translation theorists know, but is scarcely

recognized in the philosophical literature, is that translation is not a word for word, expression for

expression exchange across languages based upon some linguistic criteria of synonymy. Rather, it is a

36
selective process of textual reconstruction in light of the many modalities of an original text and the

filtering criteria of a text-type.

Chapter 3 deals with the semantic consequences of taking text-types seriously. The text-

type-theoretic approach to translation, coupled with the orthodox thesis that translation aims at

equivalence, leads to sceptical views about the determinacy of linguistic meaning for in each case that

translation is successful it seems that what we are doing is preserving semantic modalities of a text

relative to a text-type. I argue that in light of the unassailable success of text-type-theoretic

approaches to translation, the traditional linguistic approach to semantics is under-motivated and

should be abandoned. Failure to abandon the traditional approach to meaning in light of a sound

account of translation leads to all manner of problems in our ability to think clearly about how

determinate translation works. Importantly, the failure to abandon the linguistic approach to

semantics in light of a sound account of translation leads to the conclusion that we could perfectly

translate a text according to every reasonable expectation of an area of inquiry, but the text would

have a different meaning than the original, and thus it, and any propositions abstracted from it, could

not be attributed to the author of the original. This would have an unacceptable consequence for our

ability to conduct research internationally. All knowledge would be incurably culture-and language-

bound, even though we might have fulfilled every reasonable institutional expectation in translation

that underwrites the institution relevant claims of knowledge. The linguistic turn is shown to be

culturally solipsistic. In its place, we should embrace TTS, which I set out. This chapter also engages

in a comparison of the text-types of philosophy, science and literature. QI as I am forwarding it is

simply the formalization of the text-type of philosophy in the form of its most obvious incarnations,

namely philosophical concepts.

Chapter 4 is a central chapter in this dissertation. It deals with linguistic accounts of moral

semantics, particularly those that are alleged to be naturalistic. Linguistic approaches to moral

semantics, including naturalistic accounts that hold that science can lay bare the meaning of moral

37
concepts, as well as criticisms of such views in the literature (such as Mark Timmons and Terry

Horgan’s famous “Moral Twin Earth” argument), are shown to all be quite incapable of explaining

normative discourse translation. QI, in contrast, is shown to do the job where other accounts fail.

Chapter 5 addresses the question of whether Expressivism can provide a semantic basis for

the translation of normative discourse. I show that Expressivism fails, and the reason it fails is

because it is dismissive of the philosophical reasons that authors have for their normative claims as

irrelevant to normative semantics. QI, in contrast, defers to the actual philosophical convictions of

authors and can thus accurately translate normative discourse. This chapter not only shows that

Expressivism is not up to the task of explaining normative and evaluative disagreement across

languages and cultures, but it also proves that one cannot accurately translate normative or evaluative

discourse without treating it as an aspect or version of philosophical discourse. Normative and

evaluative concepts must thus be understood as sub-varieties of philosophical concepts. QI is the only

account in the running that is based upon a sound theory of translation and explains how evaluative

and normative concepts are a species of philosophical concepts that are themselves devices of textual

translation.

Chapter 6 sets out QI in detail. QI has some features in common with indexical accounts of

philosophical concepts in the literature, provided by such authors as Jamie Dreier, David Phillips,

Keith DeRose and Stewart Cohen. However, there are very important differences as well. QI, first of

all, generalizes to all philosophical concepts, and provides a clear account of the thick-thin

distinction. Secondly, and perhaps most importantly, QI is not an account of terms or concepts of a

language, but rather of the text-type feature of philosophy. Thus it, unlike its competitors, is

indispensable to the translation of philosophy.

Chapter 7 continues the discussion on QI. Specifically, the quasi-indexical approach that it

urges might superficially make it appear as though it is a species of what nowadays is called

“Contextualist Semantics” according to which meaning is context-sensitive. TTS, on which QI is

38
based, is shown to be strongly opposed to Contextualist Semantics. TTS, in contrast, is a deep

criticism of the notion that meaning is context-bound. QI is also defended against recent criticisms of

indexical accounts of philosophical terms from Insensitive Semantics, or Semantic Minimalism. I

argue that TTS is the only truly context-insensitive semantics, while so-called Semantic Minimalism

is merely an elaborate form of Contextualism.

Chapter 8 concludes this dissertation. I briefly summarize the various arguments in support

of QI, defend it against the charges that it does not protect the culture- and language- neutrality of

translation, and that it is too pluralistic in its conception of epistemic objectivity. I also argue that it

shows that relativism is mistaken. I stress that my argument for QI does not depend upon it being an

effective argument against relativism. Rather, the considerations in favour of QI have to do with

translation, for the thesis I am defending is that we need QI, and its account of thick and thin

concepts, in order to translate normative discourse. That it can also provide a foundation for a

response to the relativist is an important reason for being interested in QI, TTS and translation.

39
 II. Translation, Equivalence and Text-Types

“1. A translation must give the words of the original. 2. A translation must give the
ideas of the original. 3. A translation should read like an original work. 4. A
translation should read like a translation. 5. A translation should reflect the style of
the original. 6 A translation should possess the style of the translation. 7. A translation
should read as a contemporary of the original. 8. A translation should read as a
contemporary of the translation. 9. A translation may add to or omit from the original.
10. A translation may never add to or omit from the original. 11. A translation of verse
should be in prose. 12. A translation of verse should he in verse.” (on the
contradictory advice to translators in the literature —Savory 1968, 54)

“This phenomenon, whereby a theorist makes global observations on translation in

general, but actually means only one, often narrow area of it, still characterizes
translation studies today—to the detriment of a general theory of translation.” (Snell-
Hornby 1988, 14)

Even Quine’s definition of indeterminacy in translation feigns to be upset about the

same text leading to different translations “which stand to each other in no plausible
sort of equivalence relation however loose” (1960:27, italics ours). Who told
Quine that translation was expected to depend on equivalence? (Pym 2004, 58)

…when it comes to LSP texts [i.e. specialized texts], translation scholars seem to have
little difficulty in accepting the notion of complete logical equivalence… whereas
equivalence has otherwise become something of a four-letter word.…(Madsen 1997,
287)

II.1. Introduction
In this and the next chapter, I shall discuss the question of translation—what it is what

philosophers and translation theorists have had to say about it. In this chapter, I shall present an

account of translation based upon the best insights from the translation studies literature while

criticizing the irrational and unfounded convictions in this literature (based, largely, upon the

diffusion of the linguistic turn in the humanities). The next chapter continues with the issue of

translation, but brings in to discussion the other side of the coin: meaning. The topic of meaning is

never far from the topic of translation and the general omission of the term “meaning” from the

discussion in this chapter is contrived. Translation distinguishes itself, among semiotic processes of

40
conversion, in being concerned with semantic content preservation. If we take away the notion that

translations are semantically equivalent to their originals, we have very little left supporting the

notion of translation as distinct from adaptation, interpretation or what is sometimes known as

paraphrase. However, I shall leave the word “meaning” largely out of the conversation in this chapter,

and reintroduce the topic in the next chapter. Once we have a clear account of what determinate

translation is about, we will be able to have a clear account of determinate meaning.

In this chapter, I shall defend the traditional conception of determining success in translation.

According to the orthodox notion of translation, translation is successful when the items translated are

equivalent with the originals. However, I shall argue that this feat of equivalence must be relativized

to institutional standards. Translations will thus be equivalent to originals with respect to not all, but

some features of the original. The type of feature that an original and translation share defines what I

will call a “text-type.” I take the notion of a text-type from the translation studies literature.22 A text-

type is like a genre except that it has widespread institutional support. In arguing that translation can

be unequivocally successful when it preserves features specified by a text-type in translation, I am

arguing that translation is really only determinate when we translate texts of a definite type.

Language, as such, cannot be translated determinately.

I will call the defining features that must be preserved in translation text-type features. On

the view I will forward, not all texts of a certain type are translations of each other, though those that

bear the text-type features in the same way can stand to each other as translations. For example, if

rhyming on a certain theme chosen by an author is a text-type feature, a translation of a text of this

genre will also rhyme on the same theme. The two texts may differ in many respects, including (very

likely) their literal meaning, however, so long as the text-type feature is borne in the same way—

which in this case amounts to the same theme being rhymed—the derivative text is a translation. Such

a feature would not be borne in the same way if, for instance, the derivative text rhymed on a different

theme.

41
 It follows from the view that I set forth—which I will call the Liberated Equivalence

Conception of Translation (or LET)—that different types of texts must be translated according to

different strategies because their text-type features differ. Hence, scientific texts ought to be translated

differently than poems, for what are essential in the two differ widely. I shall call a theory of how to

translate texts of a certain text-type a “text-type theory of translation,” or a “text-type theory” for

short. Moreover, on the view I shall present, the institutionalization of such text-type theories plays an

important role in resolving controversy in translation.

In this chapter I shall argue for LET by first examining what I shall call pre-reflective and

naïve conceptions of equivalence in translation. LET succeeds exactly where these restricted accounts

of equivalence in translation fail: in being both realistic in aim and custom-tailored to specific text-

types. I will then examine philosophical criticisms of equivalence issuing from the recent Continental

and Analytic traditions. Investigating the threats from Deconstruction and Quine’s indeterminacy of

translation thesis will confirm a theme: critics of equivalence in translation have an unrealistic

conception of equivalence in mind, which LET disavows. The positive case for the practical

implementation of LET will largely be made in responding to Quine’s indeterminacy of translation

thesis. Next, I shall distinguish LET from an influential school in translation studies: Skopos Theory.

After this, I shall examine a very influential school of translation theory that, when elevated to the

status of a universal theory of translation, puts forward a monistic conception of equivalence in

translation. The school may be called Functionalism, and I shall argue that in its universal

manifestation it leads to patent absurdities that are remedied by LET. This discussion shall place us in

a position to search for an account of the text-type features of philosophy in the next chapter.

Comparing philosophy to other text-types in the context of translation promises to yield insights into

the peculiar nature of philosophical discourse that translation ought to preserve, and which a

semantics of value concepts ought to respect.

42
 II.2. The Liberated Equivalence Conception of Translation

II.2.1 Criticisms of Equivalence in Translation

What is translation? Pretheoretically, we may be inclined to answer that translation is a

process whereby the “meaning” of one text (called the source text or ST) written in a language

(usually called a source language or SL) is transferred or reduplicated in another text (called a target

text, or TT) in a distinct language (called a target language or TL). A text is a translation, on this

account, owing to its derivative status with respect to an original. Moreover, a translation, on this

account, mirrors, reflects or re-presents the meaning of an original text. Let us call this the Pre-critical

Equivalence Conception of Translation or PET (cf. Catford 1965, 20-1). (As we shall see, this is only

one version of the Equivalence Theory.) On this view, to read a translation is thus to be given access

to the ST without being able to comprehend the SL.

One objection to PET is that it renders translation a phenomenon that only occurs between

distinct languages. However, translation is a wide process, involving the conversion of many

meaningful systems. For instance, translation might proceed intralingually (via a process of

rewording, which involves conversions of signs into other signs within the same language),

interlingually (across languages), and intersemiotically (across symbolic systems, such as in the case

of translation from a full language into a narrow system of icons).23 There may be instances where

translation within a language is an illuminating and legitimate exercise (say when one translates a

technical work into ordinary language). Moreover, the Continental philosopher John Sallis argues that

it is even reasonable for us to countenance the translation of poetry into pure (non-lyricized) music,

which if possible would constitute a case of intersemiotic translation (Sallis 2002, 120-2). In

restricting translation to an interlinguistic affair, PET cannot account for these prima facie processes

of translation. The defender of PET might insist that this is not a failure. Rather, translation is

essentially a linguistic affair, they might argue. The critic however will rightly point out that the value

43
of translation is its ability to overcome limitations of mediums of communication. Thus there is

nothing conceptually problematic about the idea of translating across symbolic systems. If a theory of

translation should fail to meet this challenge, it can reasonably be thought to fail as a theory of

translation for it would not be able to show how content comes apart from its representation.

These objections to PET are fair. Translation is supposed to facilitate the presentation of

meaningful content in a medium or manner that it was not originally presented in. There is thus

something counterproductive about PET’s implication that translation cannot proceed

intralinguistically or intersemiotically. An acceptable account of translation must allow for such

textual conversions. However, this objection, while important, is not prominent in the literature

against equivalence conceptions of translation.

According to another criticism, what makes PET objectionable is that it allegedly implies that

a necessary condition of a text being a translation is that it reproduces literal significances of the ST

(cf. Holmes 1988 [1978], 100-1). Contingent factors such as the terminological resources of

languages, the rhetorical and figurative repertoire of linguistic communities (their shared metaphors,

similes, imagery and cultural references) and obligatory features of grammars24 that are not shared by

two languages make the expectation that a translation be the literal equivalent of the original wholly

impractical in many cases.25 If I were to translate the sentence “He was sweating like a pig” into the

language of a community of pig herders according to a theory that equated meaning with the literal

meaning of the sentence, the result would likely be a misunderstanding of the purport of the original,

for pigs in reality do not sweat much at all, if at all (hence they require mud baths to cool off), but the

phrase “sweat like a pig” in English is meant to denote copious perspiration. Critics of equivalence

appeal to such examples to cast doubt on the idea that translation can be the literal reproduction of the

meaning of a ST in a TT.26

Mary Snell-Hornby in her introductory book entitled Translation Studies: an Integrated

Approach provides a different argument against equivalence in translation:

44
 …it presupposes a degree of symmetry between languages which makes the postulated equivalence
possible. Nowhere is the fallacy in such thinking better illustrated than in the term equivalence itself.
From… selections of definitions one has the impression that the German term Äquivalenz and the
English term equivalence are identical and themselves exemplify the linguistic relationship they set
out to denote…. In fact the opposite is true: on closer investigation subtle but crucial differences
emerge between the two terms, so that they should rather be considered as warning examples of the
treacherous illusion of equivalence that typifies interlingual relationships (Snell-Hornby 1988, 16-7).

Snell-Hornby clearly throws the baby out with the bath water here: she fails to recognize that

‘equivalence’ itself is not a monolithic concept that specifies in every instance some coincidence of

properties, nor does it require that two items be equal in every respect in order for them to be

equivalent in some respect. (For instance, to say that two meals—one Chinese and the other

Mexican—are equivalent in nutritional value is not to say that they are identical in every respect, or

that they even look and taste like each other). As it always makes sense to ask in what way two items

(such as a ST and TT) are equivalent, an equivalence conception of translation does not require that

there be a terminological or rhetorical symmetry between languages in every respect.27

II.2.2 Advantages of LET

We may thus distinguish between two theoretical versions of the equivalence conception of

translation: a naïve version and a liberated version. The Naïve Equivalence Conception of Translation

(henceforth “NET”) requires that translation consist in a complete re-rendering of every feature of the

ST in the TT. The Liberated Equivalence conception of Translation (henceforth “LET”) attempts to

recreate in a TT the text-type feature of the ST in the same way.

LET has several advantages. First, it is not forced to turn the anthropological reality that

languages and linguistic communities differ with respect to rhetorical or terminological/conceptual

resources into a universal obstacle to translation. For a ST and TT may stand to each other as original

to translation despite these differences if they achieve some type of satisfactory equivalence,

45
understood in terms of text-type features and the relevant particulars of a text. (Perhaps the

equivalence will be achieved by sacrificing a rhetorical dimension while maintaining symmetry of

terminology or conceptual instantiation, while other times translation is effected by retaining the

rhetorical character while adapting to native terminological resources.)

Secondly, LET provides a broad (albeit general) normative criterion against which a

translation can be appraised. It explains how, for instance, the Good News Bible has some claim

(however weak) to being called a translation of earlier Bibles in Greek or Latin or of the King James

Version, but how it is implausible that Marx’s Capital is a translation of any version of the Bible.

Third, it provides us criteria for assessing the adequacy and quality of a translation: if an

intended TT departs from a ST in some crucial way that does not bear the text-type feature in the

same way, we have grounds for saying that the TT is a bad (or no) translation. Fourth, LET builds on

a long tradition of conceptualising translation along the line of equivalence.

As evidence of the defuse presence of the equivalence conception of translation in the history

of Western thought on translation (and a latent pluralism about what equivalence consists in), we can

look to John Dryden who crystallizes the competing approaches to translation before him (stretching

back, arguably, to St. Jerome, and Cicero). Dryden surmises that there are three ways to translate a

text: metaphrase (also called “literal translation” where the TT mirrors the ST both syntactically and

terminologically), paraphrase (where the TT is not expected to be equivalent to the ST with respect to

syntax and terminology but in terms of sense), and imitation (were the TT is produced with the aim of

duplicating some aesthetic features of the ST). Dryden favours paraphrase, in most instances, over the

other forms of translation (Dryden 1992 [1680]). However, all three manners of translation aim at

some type of equivalence between a TT and its ST. Most authors who have followed Dryden have

continued to affirm the role of equivalence in securing a translation, but like Dryden have privileged

one or two such methods as translation while deriding the contraries. This is particularly true of

philosophers. Schopenhauer appeared to prefer metaphrase to alternatives, suggesting that languages

46
can and ought to be enriched in order to effect translation, while eschewing both paraphrase and

imitation even as means of translating poetry (Schopenhauer 1992 [1800]). Nietzsche

characteristically rejected the truth-theoretic approaches to translation (both metaphrase and

paraphrase) and held that translation can only proceed where imitation is possible, and failing that

“translations” are at best generalizations of the original (Nietzsche 1992 [1882]).28 Schleiermacher

recognized that translators have two choices: to bring the audience to the style and mode of the ST, or

rework the TT so that it is in conformity to the target audience’s capacities. In all matters of

translation, Schleiermacher advocates the former and hence affirms metaphrase (or something close

to it) as the means by which translation ought to proceed (Schleiermacher 1992 [1938]). Much later,

Quine, following through to dissolution the project of Leibniz and the Logical Positivists (of

constructing an ideal language into which natural languages can be translated), distinguishes

“translation” from paraphrase. Paraphrase, according to Quine, does not aim at synonymy and thus

does not make claims of equivalence of meaning, though it can be employed to substitute expressions

with more lucid elements (Quine 1960, 161, 250). This formulation threatens translation for Quine,

for he can make little sense of the idea of synonymy. But it speaks to the prevalence of some form of

an equivalence conception of translation in philosophical reflections on translation. LET thus is heir

to this tradition, but surpasses it in its flexible employment of the concept of equivalence.

Fourth, even those who criticize equivalence in translation theory, such as James Holmes,

subscribe to some liberated conception of equivalence, though they are often loath to admit it

(Holmes 1988, 37, 43 fn.10). Theorists such as Holmes and Snell-Hornby who are critical of

equivalence are often also pluralists about translation strategies, and hence some type of liberated

conception of equivalence is implied in their pluralism. Given the pluralistic leanings of much recent

work in translation studies, it is not surprising that liberated conceptions of equivalence have been

entertained. For instance, W. Koller in his article “Equivalence in Translation Theory” recognizes

five types of equivalence: (1) denotative equivalence, (2) connotative equivalence, (3) text normative

47
equivalence (Koller elucidates this by reference to the idea of stylistic equivalence, but it seems close

to my idea of a text-type feature), (4) pragmatic equivalence, and (5) formal equivalence (or artistic-

aesthetic equivalence) (Koller 1989 [1979], 100-1). Koller appears to champion LET when he writes

that the norm of equivalence in translation “does not say anything about the kind of relation” between

a ST and TT, and that “this must be additionally defined” (Koller 1989 [1979], 100). LET however is

not restricted to countenancing only five conceptions of equivalence in translation and counsels not to

decide in advance the types of equivalence at play in translation. (Koller himself does not seem to

place any great weight on his list of equivalence types being exhaustive).

Fifth, equivalence as a criterion of translation is indispensable for the very purpose of a

translation is to take the place of the original (if it didn’t, we would never allow our students to write

essays based upon translations as though they were writing on the work of the original author). If a

translation cannot be said to be equivalent to the original in some critically relevant way, we have no
29
justification for relying upon a translation even as a first step introduction to a text. However,

equivalence conceived in a monistic fashion cannot do justice to the varying types of translation tasks

that face translators. Hence, some type of liberated conception of equivalence that changes according

to text-type is necessary.

Sixth, LET does not tie translation to any particular text-type or medium, nor does it demand

that STs and TTs be composed in distinct languages. Hence it can smoothly account for how

intersemiotic translation and intralinguistic translation are translations on the same footing with any

other interlinguistic translation. Orthography to sign language, jargonese to non-technical talk, poetry

to music, may all be explained as translational conversions so long as the text-type theory that is

employed specifies a type of equivalence that can be achieved across the respective mediums.

Finally, LET is an instructive platitude, I shall argue, for most theories of translation err

precisely when they attempt to restrict equivalence in translation to a definite set of criteria for all

translation. When theories of translation take a stand on how all STs and TTs must be equivalent (if

48
they specify, for instance, that all translations must be equivalent in “meaning,” conceived of in some

traditional, truth-theoretic manner, or that all translations must conform to the cultural values of the

target culture), they stop being useful general theories of translation. At best they thus become useful

recommendations for how translation in a certain region ought to proceed (say, scientific, poetic, or

Biblical translation), and at worst they become improper generalizations from particular cases.

Criteria for equivalence in translation, I shall argue, are regional, and to suggest otherwise is to be

guilty of a certain type of provincialism.

One might argue against LET by claiming that the word “translation” is in fact ambiguous.

Consider, for instance, a position that is consistent with LET. In some circumstances, it may be

appropriate to call an imitation authored in a TL the translation of some ST. For instance, to translate

a comedy, or a poem that makes cultural references or relies upon rhetorical devices not shared by the

SL and TL, one may substitute a work in the TL that has certain aesthetic similarities with the ST.

According to LET, we might be justified in calling this substitution a “translation” if the TT shares a

text-type feature with the ST in the same way. Perhaps this qualifies as a “translation” in some sense,

but it seems to depart from a lay concept of translation, dominated by the idea of literality. Let us call

this the ambiguity of “translation” objection to LET. The supporter of LET, I believe, can respond by

conceding, for the sake of argument, that the term “translation” might be ambiguous, and that perhaps

there is a distinct concept of ‘translation’ for each justifiable set of criteria employed in evaluating

whether two items stand to each other as ST and TT. ‘Translation’ might thus be a cluster concept or

a family resemblance concept (in Wittgenstein’s sense). However, this should not affect the

plausibility of LET as an account of how we ought to use the term “translation.” Moreover, if LET is

the correct account of translation, it shows us something interesting: that all things that we are

inclined to call “translations” share something—namely a liberated conception of equivalence

according to which a ST and TT stand to each other in the relation of original and translation. This is

significant. If such a commonality can be found, it suggests that it is not the concept of ‘translation’

49
that is all that ambiguous, but that it has a broad range of application. What seems like ambiguity

from one perspective is thus really an appreciation of the liberated conception of equivalence at work

in the concept of translation.

In the following two sections I shall address philosophical objections to LET from the

Continental and Analytic traditions. In both cases I hope to show that the criticisms against

conceiving of translation in terms of equivalence rely upon unrealistic expectations of what

translation could aim at. LET avoids such unreasonable expectations.

II.3. Philosophical Criticisms of Equivalence I: Deconstruction

II.3.1 Heidegger’s Hopes for Translation

Much of the Deconstructionist tradition can be understood as a response to and attempted

resolution of the internal tension in Martin Heidegger’s thought on meaning and translation.

Heidegger in a relatively early stage of his thought was of the view that translators in the Western

tradition on the whole look back on Anaximander (the oldest of the pre-Socratics) through the

ontological lenses of post-Socratic philosophy (Platonism and Aristotelianism) but also through the

related metaphysical framework that presumes that the structure of the world is settled, discrete and

transparent. Heidegger rejects such readings, because they forced Anaximander’s thought into the

framework of metaphysical thinking and at once criticizes it for not respecting all its distinctions.

Heidegger’s solution is not to appeal to a social science such as philology or historiography. Rather,

he holds that one must attempt to access the fragment by becoming accustomed to the

prephilosophical, poetics of the Greeks (Homer’s Iliad, for instance). We must thus render ourselves

(and thus in a sense “translate ourselves”) into the language of the original in order to translate the
30
fragment (Heidegger 1984 [1946]).

50
 II.3.2 Deconstruction’s Response

The tension in Heidegger’s approach was not lost on Deconstructionists—particularly his

student Derrida: the very expectation that there is a kernel or original meaning of a text that is

unalterable but can be accessed through some form of investigation presupposes a metaphysical

framework—particularly one of presence. This is the idea that the significance of an item is an

irreducible property of it (whether the property be the meaning of a text, or the text itself) that can be

retrieved or summoned up, if we only entered into the proper investigation (Derrida and McDonald

1985, 114-117).

Jacque Derrida’s response to this early Heideggerianism is to reject the idea of a kernel of

meaning and replace it with the idea of writing (i.e. an inscription) that forms the foundation for

“meaning” by its connectivity and deferral to other inscriptions. Writing is characterized by what

Derrida calls differance (conjuring up the idea of both particularity and the activity of deferring). (It is

thus not unnatural to appreciate Derrida as some type of semantic connectivist, if not holist.) This

view not surprisingly poses some difficulties for translation, if translation is conceived of as the

retention, conservation and transportation of meaning from the ST to the TT. The original text, on the

Derridian account, is itself fraught with semantic multiplicity and play that is a function of the

complex, non-foundational, deferring function of writing, which always sends the reader to another

inscription for the sought after clarity. Moreover, the movement from one system of writing to

another involved in translation always involves a process of loss from the ST with its unique
31
inscriptional relations. Thus, Derrida in a round table discussion on translation states:

Translation can do everything except mark this linguistic difference inscribed in the language, this
difference of language systems inscribed in a single tongue. At best, it can get everything across
except this: the fact that there are, in one linguistic system, perhaps several languages or tongues [i.e.
systems of meaning]. Sometimes—I would even say always—several tongues. (Derrida and
McDonald 1985, 100)

51
Moreover, translation is involved in a fundamental paradox, according to Derrida, which he calls the

“double-bind”: translation at once involves the mandate of transferring meaning, but by virtue of

transgressing systems of writing must also respect its impossibility (Derrida and McDonald 1985,

102). The reality of loss in translation is brought home by several examples provided by Derrida, the

most salient being the translation of the Greek “pharmakon” in the Platonic dialogues. In his Plato’s

Pharmacy, Derrida notes that the term has several meanings, including ‘poison’ and ‘medicine,’ and

that Plato often relies upon the ambiguous nature of the term “pharmakon” to advance the plot.

Derrida calls this simultaneous use of the term in an ambiguous fashion that takes on one meaning in

the present context, while also referring to itself as an ambiguous symbol, and to multiple meanings

from the perspective of the various characters in a plot, anagrammatic writing—a feature of writing

that cannot survive translation, if one chooses any one of the component significations to render it
32
(Derrida 1981 [1967], 98).

Michel Foucault also appears to undercut the equivalence conception of translation in a

parallel fashion. Foucault quotes Jorge Luis Borges in his epigraph to Language, Counter-memory,

Practice as saying, “The fact is that every writer creates his own precursors. His work modifies our

conception of the past, as it will modify the future” (quoted in Foucault and Bouchard 1977, 5).

Foucault thus echoes the Derridian notion that there is no kernel of discrete meaning in an original

text aside from an open-ended play of writing, but adds to this the observation that the meanings of

historical texts are determined by many subsequent, institutional factors as well. The factors in

question are brought to the fore in Foucault’s unique idea of an “Author-Function”, which replaces

the traditional idea of authorship. Whereas the traditional notion of authorship regards the author as

the source of the meaning of a text, the author function replaces this picture with an analysis of the

institutions and factors at play in interpreting, controlling and distributing texts that are attributed to

an author (Foucault 1984).

52
 Deconstruction is not itself a theory of translation, and it appears to have no great use for the

concept apart from treating it as a locus for critical inquiry into scholarship.33 While it thus might not

offer a positive account of translation, it seems to problematize equivalence conceptions of

translation. To summarize, Deconstruction offers two criticisms of equivalence conceptions of

translation: (a) STs lack a discrete set of meanings that subsequent translations can be judged against,

and (b) the process of translation is a process of differentiation, not a process of semantic retention.

Edwin Gentzler, writing on Deconstruction and translation theory, suggests that one of its

implications is that the ST has no clear privilege in the process of translation, for a “translation” can
34
function as a factor that revises our conception of the “original” (Gentzler 2001, 145, 150). If this is

so, then Deconstruction appears to problematize an equivalence conception of translation in an

additional manner: (c) the TT is not to be judged in light of the ST, but the ST is constituted (and

hence judged) by the TT. Can an equivalence conception of translation defend itself against these

objections?

II.3.3 Equivalence against Deconstruction

By abandoning the view that equivalence in translation must be secured by making the TT

instantiate all characteristics of the ST, LET can embrace many of the insights of Deconstruction

(particularly (a) and (b)) without compromising the idea that a translation is justified in light of its

equivalence with some crucial feature of the ST. What features are relevant to judging a translation as

equivalent to the TT are not specified by LET. Rather it holds that different text-types call for

different criteria and thus that it is unreasonable a priori to specify how translation is to proceed in

every case.

Deconstruction’s putative dethroning of the importance of the ST and the promotion of the

importance of the TT would cause problems for LET as it is formulated, for according to LET a

translation is to be judged (at least in part) in accordance to its fidelity to an original text. The insight

53
of Deconstruction according to some is that TTs (among other texts) constitute STs as much as STs

constitute TTs, and hence there is no factual basis for setting up the ST as the standard against which

a TT is to be judged. I take this last criticism to be an error that results from a conflation between two

distinct notions: ‘translation’ and ‘interpretation.’ The two concepts are often treated

interchangeably—particularly in the Hermeneutic tradition (Heidegger, for instance, was reputed to

have held that every translation is itself an interpretation, cf. Lilly 1991 [1956], vii; Gadamer 1996

[1960], 384, 387)—but there are important differences between the two. Translations and

interpretations are easy to confuse, for both function as proxies for original texts. However, an

interpretation is an explanation: it seeks to explicitly shed light on a distinct corpus. It is an

intermediary. A translation, in contrast, takes on the literary identity of the original: it is treated as

though it is the original—and moreover, there seems to be nothing wrong with this, if the translation

is good. This is borne out in the normal inclination to treat the author of the original as the author of

the translation.35 The distinction is made clearer by considering a secondary work in philosophy, say

the Routledge Philosophy Guidebook to Plato and the Republic (Pappas 2003), and the Republic, by

Plato. In a philosophy class on Plato, both interpretations and translations may be consulted, and the

former may be presented as the authoritative introduction and account of the thought of Plato, but no

professor will (or ought to) allow students to treat the Routledge Philosophy Guidebook to Plato and

the Republic as though it were the work of Plato. The distinction between interpretations and

translations is further obscured by the notion that we need to interpret a text and come up with a

unique account of what it means in order to produce a translation. However, if text-types play any

determinative role in translation, we need no totalizing account of the content of a text, for translation

is always only a process of selecting some features of a text out for reproduction in a TT. Determinate

translation is thus an extremely self-conscious process that involves a thorough analysis of the various

features of a ST and a self-conscious decision to preserve some features in a translation.

54
Interpretations, in contrast, frequently do not involve this degree of self-critical awareness: they may

be totalizing explanations of the content of a text.

What does this tell us about the Deconstructionist idea that “original” texts are constituted by

later texts, including translations, and thus that it makes no sense to judge a translation against an

original? It tells us that Deconstructionists may be correct that texts come to be seen in new light due

to complex social factors, including the authority given to certain interpretations (certainly this is

correct about isolated cases, if not as a generalization). However, a translation is not in itself

explicitly an interpretation. Thus, it still makes sense to judge a translation in light of “originals”—all

the while acknowledging that what we take to be the importance of an original text is something that

is determined by many factors that are contemporary to us.36

Finally, it is worth noting that though Deconstruction’s criticism of translation appears to

radically threaten the enterprise of translation by banishing from translation the very idea that there is

a stable or original kernel of meaning that a translation must approximate, the criticism is relatively

benign for it leaves intact the idea that translations are to be judged in part by their ability to

approximate the semantic play of the ST and moreover that this data itself (difference) is determinate.

Indeed, it only is on the basis of the norm of equivalence and the determinacy of difference that the

very idea that translation involves a loss of the original and a differentiation can make sense. As a

concrete example of this type of tacit reliance upon a principle of equivalence in translation and the

determinacy of difference we need look no further than Derrida who repeatedly takes us back to the

original Greek of Plato as a standard against which translations are judged (recall he criticizes

translations of “pharmakon” such as “remedy” instead of “drug” or “medicine” for destroying the

anagrammatic element of the term that allows it to refer at once to beneficent substances and poisons)

(Derrida 1981 [1967], 98).

Derrida appears to have eventually recognized the relative harmlessness of his critique of

translation, for it leaves untouched the possibility of the inessential being left behind while retaining

55
what is relevant to the concern of the translator—a view that resonates with LET. In his article “What

is a ‘Relevant’ Translation,” Derrida has a final go at the notion of translation. He criticizes

translation by drawing an intriguing analogy to the forced conversion of the Jewish merchant Shylock

in Shakespeare’s The Merchant of Venice, who is first pressured to accept a different form of payment

than the one that was contracted for, next to forgive the debt owed to him (even though forgiveness is

supererogatory, and not an obligation) and then forced to convert to Christianity while having his

entire wealth confiscated. This process is precisely the process of translation, according to Derrida,

where the foreigner’s wealth is appropriated by a domestic culture, while the foreigner’s integrity and

values are not respected. The process appears acceptable from the perspective of Christianity, with its

emphasis on spiritual rebirth and conversion (the translation is by analogy the rebirth of a ST) but will

always be an affront to the ST culture (Derrida 2004 [2001]).

While the essay is intriguing, displaying Derrida’s unique prowess at recognizing analogies

(and a sophisticated analysis of The Merchant of Venice that paints Shakespeare as the confessor of

Christian anti-Semitism, rather than an anti-Semite himself), it is a relatively weak criticism of

translation, from the perspective of the philosophy of language, for the process of translation is not

rendered essentially incoherent on this account. Rather, if correct, it shows that translation is morally

repugnant though possible. The vices of translation have to do with the putative coercive power of the

target culture. However, we ought to question whether it is always the target culture that sets the

unilateral standards and terms of translation—indeed, even if this happens in practice, why ought we

to think that this is the ideal that translation ought to strive for? Translation might equally be seen as a

process of accommodation of the foreign. Translation involves loss, as Derrida notes, but loss itself

may be part of a type of economy (something Derrida is fond of emphasizing) and economic

transactions need not always be exploitive to those who are a party to them. A better analogy for the

ideal of translation is the successful integration of immigrants in a multicultural society; neither side

loses anything essential; both are enriched. It would be wrong to assume that this is what always

56
happens. But it would also be wrong to assume that it could not happen, and that we shouldn’t be

striving for this ideal, in society and in translation.

II.4. Philosophical Criticisms of Equivalence II: Indeterminacy of

Translation

II.4.1 Enlightenment and Positivism

The Analytic tradition of philosophy, like its Continental counterpart, has also offered a

criticism of equivalence conceptions of translation in Quine’s conception of the indeterminacy of

translation. If successful it is more damaging than the criticisms of Deconstruction for it not only has

equivalence conceptions of translation in its sights, but also the very idea that we could always

determine which of competing translations is the better or worse though the translations themselves

are logically incompatible.

Quine’s arguments for indeterminacy in translation are best understood as the culmination of

an enlightenment project of constructing an ideal or universal language into which natural languages

could be translated. The idea has roots in the thought of such philosophers as Descartes (Descartes

1970 [1629], 6) and Leibniz (Leibniz 1966 [1666], 11).The latter conceived such a language as an

“alphabet of human thought” that would be readily intelligible to all human beings regardless of their

spoken language. According to Rudolph Carnap, in so conceiving of an ideal language in this manner,

Leibniz envisioned a Begriffsschrift (‘Concept-writing’) in Frege’s sense (Frege 1967 [1879]).

Carnap, the famous member of the Vienna Circle, was inspired not only by Leibniz but by the

international language movements of his day, particularly Esperanto (Carnap 1963, 66-71). Carnap

thus strove, within the wider project of Logical Positivism (which aimed at, among other things, the

promotion of science and the pruning of metaphysics and other vestiges of pre-scientific thinking off

of human thought), to develop an ideal, canonical language serviceable to the sciences, into which all

empirically meaningful sentences of natural language could be translated (thereby functioning as a

57
formal test for the meaningfulness of the natural language claim) with the scaffolding of modern

formal logic.

Carnap operated with various empiricist conceptions of meaning to sharpen language through

translation. The first theory he espoused was an explicitly empiricist theory of meaning, according to

which the meaning of nonlogical terms are phenomenal (sense) data (Carnap 1967 [1928]). Carnap

found that this was unworkable and in the 1930s reconceived meaning in terms of descriptions of

definite space-time points (a position known as physicalism) (Carnap 1937, 1954 [1936]).

One of the explicit aims of Carnap’s Logical Syntax of Language was to specify the general

syntactic structure of any language, whatsoever. Among the basic principles of all such languages are

thus formation rules that among other things specify the permissible notation, sentences and

transformation rules of the language. Translation and interpretation, understood in terms of a strict

logical and empirical equivalence of sentences across languages, are made possible on Carnap’s

account by reference to such basic principles of syntax (Carnap 1937, 222-233).

II.4.2 Quine’s Thesis and Argument Against Equivalence and

Determinacy

Against Carnap’s early views,37 Quine argues in Word in Object that translation is a

problematic project according to a thoroughgoing empiricism because the totality of all possible

empirical evidence cannot always determine whether a given translation—to the exclusion of other

contrary translations—is right or wrong, even though the translations in question are (logically)

incompatible (Quine 1960, 27). In other words, on Quine’s account, we may be left with a set of

equally acceptable but mutually incompatible translations, none of which we are inclined to think is

objectively correct. This is so because, on Quine’s view, there is nothing for a translation to be right

or wrong about (Quine 1960, 73). This is Quine’s indeterminacy of translation thesis.38 The

indeterminacy of translation on his view is not to be confused with “the platitude that uniqueness of

58
translation is absurd.” Rather it is the more “radical” thesis that there can be competing empirically

justified approaches to translation (what he calls “translation manuals”) that rule each other’s

translations out of order and perhaps even that the resulting translations will be divergent in truth-

value (Quine 1960, 73-4).

While Quine’s thesis is meant to apply to all translation, he brings it into relief by examining

the problem of radical translation, which is the field linguist’s task of compiling a manual to translate

a hitherto isolated people’s language into English. The only data that the linguist can go on “are the

forces that he sees impinging on the native’s surfaces and the observable behaviour, vocal and

otherwise, of the native” (Quine 1960, 28). With such constraints the closest thing to the meaning of

observed linguistic behaviour is what Quine calls stimulus meaning, which is the ordered pair of

stimuli that respectively causes assent and dissent to a sentence (Quine 1960, 31). Varying types of

sentences can be discerned for taxonomical purposes, on the basis of what Quine calls analytic

hypotheses. These pair up observation sentences in the SL and TL, and what Quine calls stimulus

analytic and stimulus contradictory sentences (sentences whose assent and dissent are tied

exclusively to a certain stimulus). Included amongst analytic hypotheses are some insights into

intrasubjective synonymy of expressions that appear to have the same reference but different senses

(in Frege’s sense); however, nothing like firm rules of translation can be offered for such cases, for

the information is at best idiolectical (Quine 1960, 68). The problem with the analytic hypotheses is

that the data they are founded upon—stimulus meaning—“must woefully under-determine the

analytical hypotheses” (Quine 1960, 72). The result of this underdetermination of rule by data is that

many mutually exclusive manuals of translation between a given SL and TL are possible, and

moreover some of these manuals may be logically incompatible yet both empirically adequate in light

of the totality of all possible empirical evidence. The absence of any possible empirical evidence to

decide this type of controversy is Quine’s indeterminacy thesis.39

59
 Quine’s global criticism of translation is also a criticism of equivalence conceptions of

translation, in so far as he understands translation as governed by the ideal of finding synonymous

sentences (in the sense of “same meaning,” cf. Quine 1960, 27, 61).40 One problem with synonymy,

on Quine’s account, is that it is implausible to find two sentences in different languages with the same

stimulus meaning. This must be so even with seeming paradigm cases of stimulus synonymy, like

Quine’s hypothetical case of “Gavagai,” which his linguist wishes to translate as “(lo a) Rabbit.”41

The more general problem with translation, and its search for synonymous sentences, is that it raises a

“conflict of parts [i.e. sentences] seen without the wholes [i.e. a given language]” (Quine 1960, 78-9).

Translation is thus doomed to fail because the items it attempts to find equivalences for are not in any

reasonable sense equivalent—each playing a unique part in its native language. (The Derridian

criticism of translation by appeal to writing thus resonates with Quine’s observations.) Quine’s

solution is to abandon translation, and to replace it with “paraphrase.” According to Quine, with

respect to paraphrases “one may in a non-technical spirit speak fairly enough of synonymy, if the

claim is recognized as a vague one and a matter of degree. But in the pattest of paraphrasing one

courts confusion and obscurity by imagining some absolute synonymy as goal” (Quine 1960, 161).

It is important to note here that while Quine’s indeterminacy of translation thesis is based

upon a criticism of the presence of synonymous sentences across languages, it is distinct in his view

from the terminological problem of synonymy. On Quine’s account, languages might differ in

terminological resources and thus lack terminological synonymy but yet be determinately

translatable. For instance, a language might have a name—”ak-!ha!”—for a certain time of the day,

roughly between 4:00pm and 6:00pm, for which English has no terminological equivalent. This poses

a problem for literal translation, but not for determinate translation for Quine, for we can without

controversy spell out in English what “ak-!ha!” is though we lack a single word for the concept. A

translation that failed to represent ak-!ha! appropriately could be thus criticized as missing the mark..

Of course, if there were two contradictory ways of rendering “ak-!ha!” in English and no way to

60
choose between them, this problem of terminological synonymy would result in an indeterminacy of

translation. The issues are thus not truly distinct.42 Difficulties in finding uncontroversial

terminological equivalents across languages will give rise to indeterminacy at the sentential level. As

it is a matter of fact that words across languages do not admit of easy, uncontroversial equivalents, it

seems that Quine’s problem runs deep.

II.4.3 Davidson’s Attenuation of the Indeterminacy of Translation

Quine’s pupil, Donald Davidson, may be seen as offering a solution to the scepticism

engendered by Quine’s approach. Specifically, with a small innovation to Quine’s machinery,

Davidson saves some notion of equivalence that is of use to translation, though he seems to reject the

idea of sentential synonymy (cf. Davidson 2001 [1967]). The innovation is Alfred Tarski’s

disquotational schema, in which a quoted sentence in a first order, object language that is said to be

true is paired up in a biconditional with an unquoted sentence in a second order, metalanguage

(Davidson 1996 [1974], 461). The schema is variously called “Tarski’s schema” or a “disquotational

schema.”

Tarski introduces his schema to shed light on the notion of truth: any adequate account of

truth on his view must account for the fact that,

“p” is true if and only if p (Tarski 1944, 343).

According to Davidson, we can develop a theory of meaning for a SL by recursive application of this

schema to the SL sentences, each of which is treated as the quoted object language sentence while the

TL sentences comprise the metalanguage on the right side of the biconditional (Davidson 2000b, 70-

2, 1996 [1974], 468). Thus, the sentence “Gavagai!” can be interpretable or translatable as “There

goes a rabbit!” if “Gavagai!” is true if and only if there goes a rabbit! While Tarski thus begins with

the idea of meaning as fundamental and moves towards an account of truth, Davidson regards his

project as taking truth as the primitive notion, and reversing the procedure to arrive at translation

61
(Davidson 2001 [1973], 134). In so doing, Davidson does away with Quine’s reliance on stimulus

meaning to anchor interpretation and translation.43

The application of the disquotational schema to the problem of translation assumes that one

can judge whether the truth conditions of a SL sentence (treated as the object language sentence that

is quoted) are equivalent to those of the metalanguage sentence on the other side of the biconditional.

Davidson makes use of a principle that Quine employs (in the context of determining the logical

connectives of sentences) to solve the problem of determining the truth of SL utterances across the

board: what Quine and Davidson called the principle of charity (Davidson 2001 [1973], 136, fn.16;

2001 [1967], 27) and what Davidson later called the principle of rational accommodation (Davidson

2000a, 23-4, 1996, 66-7). Earlier formulations of the principle counselled us to choose interpretations

and translations that “maximize agreement” between the informant and the interpreter (Davidson

2001 [1968-9], 101) or “to read some of [one’s] own standards of truth into the pattern of sentences

held true by the speaker” (Davidson 1986, 316). In formulating the principle of charity thus,

Davidson was merely taking on Quine’s formulation (who derived it from N.L. Wilson, 1959, 532).

However, in liberating the principle of charity from the narrow task of determining the logic of a SL

in Quine’s work,44 Davidson appears to open the floodgates to a type of cultural imperialism that has

translators take all STs to be in keeping with our worldview. Davidson later retreats from this

formulation, and instead suggests that the point of charity is to render the total behaviour (linguistic

and other) of a speaker or author intelligible (Davidson 2000a, 23-4, 1996, 66-7).45 So understood,

charity counsels us to attempt to widen the scope of agreement in our interpretations, thus converting

what would otherwise be a difference between incommensurable conceptual schemes into substantive

differences on how to employ concepts common to both the interpreter and interpreted. Thus, our

initial hypothesis of wide spread agreement gives way to interpretations that render our interlocutors

intelligible, though perhaps wrong about certain matters by our lights (Davidson 2001 [1974]).

62
 In employing the principle of charity/rational accommodation across the board, and in forcing

interpretation to fit in the mold of Tarski’s schema, Davidson believes that his theory significantly

reduces the indeterminacy of translation (Davidson 2001 [1979], 229-230, 1996 [1974], 463)—

something which translation theorists should be interested in. Charity thus narrows the range of

possible disagreement, including disagreement between translations, and Davidson’s employment of

Tarski’s schema to generate translations does away with Quine’s extremely narrow sense of meaning

as sensory stimulation. This widens the range of equivalence available to the Davidsonian translator.

Moreover, Davidson’s account dispenses with the possibility of conceptual relativism of SLs and

TLs, which Quine’s analysis appeared to raise (cf. Quine 1968). While the indeterminacy of

translation is not eradicated on Davidson’s account, Davidson holds that Quine and his indeterminacy

of translation as such “should be viewed as neither mysterious nor threatening. It is no more

mysterious than the fact that temperature can be measured in Centigrade or Fahrenheit (or any linear

transformation of those numbers)” (Davidson 1986, 313).

It is unclear how Davidson’s account of Quine’s (and his) indeterminacy of translation thesis

consoles those who are interested in translation, for what we translate are not whole systems (such as

languages) but texts, and as such the problem remains of how to reconcile the divergent implications

of incompatible translations in our language. No confusion arises from switching between differing

systems of measurement for we are aware that the measurement is made relative to a system of

measurement, and we are also able to independently measure their divergence. By contrast,

competing translations on Davidson’s account do not have any independent measure by which we can

uncontroversially correlate the divergence by which we measure the world with our sentences, for

differing speakers, on this account, will assign differing truth values to the translation of their

interlocutor’s sentences. There is thus no language independent perspective from which we could

track such changes relative to the common world.

63
 While Davidson offers a theory of meaning in place of Quine’s idea of a translation manual,

and thus reduces the threat of indeterminacy in significant ways, he does not offer much solace to the

translators of texts. We are still left with the problem that there may be in many cases no obviously

right translation.

II.4.4 Overcoming the Threat of Quinean Indeterminacy of

Translation: LET

The way to overcome the threat of Quine’s Indeterminacy thesis, I believe, is to address his

conception of equivalence. For if equivalence in translation can be secured, there would be no

indeterminacy in translation. For Quine, equivalence of sentences is an equivalence of meaning, and

the only scientific way that he can make sense of this is in terms of stimulus synonymy—though he

grants that stimulus synonymy falls short of our ordinary notion of synonymy (Quine 1960, 37).

Because Quine believes no two sentences across different languages will share the same stimulus

meaning, actual translations can never be grounded in scientifically respectable data for equivalence.

This constitutes the ontological basis for the indeterminacy of translation.46 Translation on this

picture is worse than guesswork, for at least in the case of guesswork there is some plausible ideal of

being correct that one aspires towards. In translation, there is no ideal of being correct that one can

scientifically aspire towards. However, a supporter of LET will find Quine’s restriction of

equivalence to such a narrow and idealized phenomenon of stimulus meaning perplexing. Such a

criterion for equivalence might appear reasonable when translating texts (including utterances) of a

purely observational nature (bereft of weighty theoretical considerations), but it seems to fail us in

every other sphere of translation where what we wish the ST and TT to be alike in respect to

properties that cannot be specified in terms of observational support.

If equivalence is liberated from a naïve conception and employed liberally as it is in LET, it

seems that there is no universal danger of indeterminacy in translation, for what determines

64
translation is not an ideal but unachievable conception of equivalence, but criteria of equivalence

relative to text-types that, unlike stimulus synonymy, can be achieved. Regional, text-type relative

criteria for equivalence may not determine unique translations, but the resulting translations need not

be indeterminate (that is, without standards of right and wrong to measure against). What will speak

to the determinate nature of the non-unique translations is their intertranslatability according to the

same principles that produced them.

Consider, for instance, the case of the text-type of novels. One of the text-type features of a

novel is timing: scenarios and plots must be divulged to readers at an uninterrupted pace.

Parenthetical editorial notes and footnotes designed to provide background information to the reader

can kill the delivery, just as an explanation of the humour of a joke can render it not funny. Further,

consider the scenario raised by the translation theorist Hans J. Vermeer:

Someone had bequeathed a certain sum to two nephews. The will [written in French] had been folded
when the ink was still wet, so that a number of small ink-blots had appeared in the text. In one place,
the text could read either as deux ‘two’ or d’eux ‘of them’. The lawsuit [that ensued] was about
whether the sentence in question read a chacun deux cent mille francs “to each, two hundred
thousand francs”, or a chacun d’eux cent mille francs “to each of them, one hundred thousand
francs.” (Vermeer 1989, 186)

If this scenario occurred as a minor incident in a novel written in French, the (interlingual) translator

would have to find a way to retell the scenario in a distinct target language without interrupting the

tempo of delivery and without the aid of the original ambiguity of the French text—that is, if the

translator were to respect one of the critically relevant features of a novel, namely uninterrupted

tempo of the plot. Vermeer suggests that the translator could do this by “perhaps introducing an

ambiguity concerning the presence or absence of a crucial comma, so that 2000,00 francs might be

interpreted either as 2000 or as 200000 francs” (Vermeer 1989, 186). Those with Quinean

sensibilities will note that such a “translation” departs widely from the original in truth-value, and

they would thus not even count it as a translation. Moreover, the translation criteria that Vermeer calls

65
upon could just as easily sanction the substitution of the original ambiguity in French with “1000,00

francs” or some other ambiguous number that would serve the purpose. The same principle guiding

translation here would not only produce differing translations, but translations that are not even

logically equivalent—a scenario analogous to the supposed contradiction that arises from divergent

translation manuals on the Quinean account. However, from the perspective of LET, any resulting

translation that preserved the numerical ambiguity of the bequeathed sum of money would be

intertranslatable for what is crucial in the context of such novel-translation is not the literal content of

such minor details but their function within the overall story. A translation that failed to preserve this

functional equivalence would be false, and those that preserved it would be equally and determinately

acceptable.

What then of non-artistic and theoretical texts of the sciences? How does LET propose to

solve the problem of translating such texts? LET could overcome indeterminacy of translation of

such texts with a double-pronged approach:

(1) texts are subsumed under the heading of text-types with achievable criteria of

equivalence, thereby furnishing both STs and TTs with standards to be measured

against (and hence, something to be right or wrong about), and

(2) texts are moulded as per text-types.

Where the totality of empirical evidence is Quine’s standard against which translations are

judged, LET assumes that the criteria against which translations are judged are institutional, discipline

relative text-type features that can be duplicated across translations in the same way. The purpose of

text-moulding is to ensure that the translations truly meet the criteria of equivalence with a source text

by also achieving the same relationship of equivalence with alternative translations, which, left to

their own devices, might permit contrary and contradictory readings and implications—even among

experts. This may be necessary, for instance, if differing translations call upon differing terms in the

TL, that approximate the extension of a term in the SL and thus function well as translational

66
equivalents, but their unchecked interpretation in the TL will lead to problems and potential conflicts

as their ordinary extensions are not identical in the TL, nor with the SL term. It may also be necessary

if the only plausible translation going leads to readings and implications that are foreign to the ST.

Text-moulding is achieved by such framing devices as marginalia, glosses, and footnotes, which

caution readers against problematic inferences and interpretations of target text language usage,

which follow naturally from colloquial usage of the target language, but are absent in the technical

language of the TT. Such framing comments serve the purpose of altering the functioning of the

target text language to the behaviour of the ST. Text-moulding will thus not only make use of novel

employments of domestic terminology, but neologisms and other innovations. Text-moulding is a

long established strategy to render translations determinate.

Text-moulding will appear objectionable to those who believe that a TT must be readily

intelligible to the target audience or that translation must always take a target language as it is, and

not attempt to enrich it or alter its behaviour. Quine tacitly assumes both. However, LET has no use

for the idea, found so objectionable by Derrida—and Quine!—that there is some original inviolable

core of a language that is forever unchanging (Quine 1990 [1951]). Especially in the context of

technical, scientific and theoretical texts, where STs are replete with stipulative uses of ordinary

terminology and neologisms (Pinchuck 1977, 166, 177; Sager 2001), it seems perfectly appropriate

that TTs also employ domestic terms exotically and novel terms domestically. Where Quine sees

indeterminacy, LET thus sees an opportunity to mold TTs into equivalents of STs. This textual

approach to translation can license such liberties with language because it does not conceive of

translation as an aim to match up words or expressions across languages in the abstract. Rather,

translation is a textual process, and thus our standard of equivalence is not based upon some a priori

conception of the fixed limits of the possibilities of expressions in a language, but rather what

contribution they do and can make to texts of a certain type. The type in question will be a latent

feature of the text, instantiated not in isolated words and sentences, but in an overall synergy of its

67
constituents. But in recognizing this synergy, we recognize a phenomenon that is not reducible to its

components. It is rather a feature of organization. We succeed in translation when we make a text that

mirrors just the right features of the ST. The fact that none of the constituents of the ST and the TT

are full equivalents when considered apart from texts and in the abstract matters not. Indeed, the fact

that there are contradictory ways of translating words and sentences considered in the abstract across

languages also matters not, for our task of translation is not to understand the contribution of words

and sentences to texts in the abstract, but rather in their relationship with other words and sentences in

a text, particularly in light of the resulting text-type patterns that emerge from such combinations. We

can take this stand on translation because we recognize it as a textual process. Quine and the heirs of

the linguistic paradigm cannot for they conceive of translation as fundamentally a linguistic process

of matching up language on the bases of an unachievable synonymy. Quine and others who have

despaired over translation give up on translation just when the real task of translation begins. They

think that the task of translation is a matter of determining the translations of words and sentences in

the abstract. This, according to the view I am urging, is simply the preparatory stage for real

translation. Real, or the paradigm cases of translation occurs when we translate whole texts, given a

suitable understanding of the multiple significances of their constituents, gauged in the abstract, and

an understanding of their synergistic results when organized in a text. The synergy that results is the

instantiation of the text-type. And, I might add, there may be multiple such synergies in every text,

instantiating multiple text-types. But this is no obstacle to translation, for translation is a selective

process of retaining some and not all features of a text in translation.

II.4.5 Responding to Quinean Criticisms

Objection 1. The Quinean might object to my argument, with its appeal to “text-moulding,”

that it obfuscates the real issue: — “How do you know you have got it right when you claim that a ST

and a TT are equivalent in the same way? Under Quine’s account, stimuli provide the independent

68
standard against which the ‘meaning‘ of languages (and thus texts) can be assessed. But the argument

that you are putting forward does not make use of any such idealized extra-linguistic benchmark to

measure meaning and thus determine translation. Your proposal seems like a whole lot of hand

waving, and no substance.”

Response 1. The Quinian argument for the indeterminacy of translation shows, if anything,

that the so-called objectivity of stimuli does nothing to anchor objectivity in translation, and thus I do

not appeal to it because I have learned from Quine’s misadventures in translation. Moreover,

translation manuals as Quine identifies them are not the means of determining translation on his

account or mine, and as these are ideally based upon stimuli, we have little reason to feel any lacking

for not appealing to them in our account. Rather, what Quine calls translation manuals are really the

raw resources that must be in place in order to produce real translations. They are the result of

linguistic and philological research, not translation. That they show that there are conflicting

“translations” of words and sentences across languages in the abstract does nothing to worry us.

Indeed, we should welcome such insights as indispensable to understanding the various possible

contributions any device can make to a text. Moreover, objectivity on the Quinian account is seriously

compromised by the individualistic nature of linguistic and translational research by his lights.

Translation manuals and any other individualistic, non-cooperative approach to translation and

interpretation are bound to yield problems of indeterminacy because the norm governing their

development is one of solitary work (the job of the field linguist in radical translation and that of the

interpreter in radical interpretation is conceptualized as the work of a single individual, and not guilds

or unified associations of linguists and interpreters) though the problem that radical translation gives

rise to (an alleged indeterminacy between competing translations and interpretations) occurs in a

social context of there being more than one prima facie, acceptable “translation” or “interpretation”

that are incompatible.47 Like all areas were we find objectivity, translation too must be understood as

requiring the cooperation of various experts. Some experts will produce lexicons and gauge the

69
multiple ways semiotic resources can be matched up across semiotic systems, like languages, and

some experts will understand the way that specific text-types play themselves out in translation.

Translation is difficult because it requires us to combine knowledge from these disparate sources. But

with such knowledge, we can objectively identify when the synergistic combination of semiotic

devices yields texts of certain types because such diagnoses appeal to institutional criteria that apply

consistently across the board in assessing texts and translating them. Just as objectivity comes about

by consistently applying the same length in measuring distance, objectivity in translation comes about

by the consistent application of text-type criteria in translation. As individuals can make mistakes, we

require an expert, institutional backdrop to provide an independent check against the work of

translators, but the decisions of the wider institution can be objective when they appeal to underling

textual norms that define text-type institutions themselves.

Given the importance of the institutional support to text-types it may seem as though

translation has nothing objective to respond to. But this is a misunderstanding. Text-types help

approach texts in a manner that tracks their features, given a concurrent appreciation of the semiotic

devices that comprise a text. For Quine, the fact that we come to multiple and at times conflicting

accounts of the significance of such pieces implies that translation is indeterminate. However, this is

only because he does not recognize the role that text-types play in choosing between competing

significances in translation. The institutional context is thus simply a professional backdrop to

properly apply a text-type in translation and to solve controversies should they arise with further

research and, if necessary, new translation conventions that will facilitate conceptual innovations to a

target community. The institutional backdrop to translation is thus judged in light of their

responsiveness to the actual features of a ST in light of the discipline-relative text-type. Our goal is

ultimately to understand the features of a ST in such a manner that we can recreate a TT with the

same features, informed by a text-type. Thus, even though the job of the translator occurs within a

certain context, the goal of translation is to realize how aspects of a text can be realized with distinct

70
semiotic devices, thus demonstrating how such features transcend the contexts of their native

linguistic and cultural communities.

Whereas there may be several competing translations or interpretations on Quine’s and

Davidson’s respective views that are forever equally acceptable but incompatible, the same is not the

case with text-type theories as I conceive them. In other words, on the account I am proposing, while

there may be alternative translations, they will not necessarily be incompatible: if they are a result of

one and the same text-type theory. If the text-type has been satisfactorily applied in translation as per

institutional standards, then they will be compatible and mutually translatable as per the same text-

type considerations. If they are a result of competing text-type theories, then the translations

themselves will not be in competition as they track different features to be preserved in translation.

Thus the Upanisads might be translatable both as poetry and as philosophy. If experts translate a text

according to one and the same text-type but find that their translations conflict in some institutionally

relevant way, then they have occasion to rethink their approach. It may be that one or both has failed

to recognize some relevant detail in the ST, or that the ST is fundamentally and irresolvably

contradictory, in which case the right translations should mirror this tension in the right way on text-

type-theoretic grounds, thus rendering the resulting translations alternative but mutually consistent

translations.

Objection 2. The Quinean might object that LET simply socializes the Quinian approach to

translation by rebranding translation manuals as text-types. Alternative text-types do not exclude the

logical possibility of alternative translation conventions, but merely stand behind the idea of

institutionalizing some over others. Your institutionally sanctioned text-type theories are like Quine’s

translation manuals. If this is so, doesn’t it seem that the problems at the individual level simply

reappear at the social level? We should expect that it does, for this type of reappearance of problems

at the individual level at the social level can be observed in other so-called solutions to problems in

the philosophy of language, such as rule-following paradoxes.

71
 Response 2. While there is a sense in which an institutionally sanctioned text-type is like a

Quinean translation manual, there is a major difference. Text-types, according to LET, do not set out

with the goal of conserving every feature of a ST in a TT. Rather, they set out with the goal of

translating a text read as a text of a certain type. Thus, translation on this account is a self-consciously

selective process while in the Quinean account it is not. If a text is translated according to contrary

text-types, the resulting translations do not run afoul of any norms in translation according to LET. In

other words, the resulting translations are not indeterminate, but merely acceptable alternatives

intelligible relative to distinct text-types. It is important to remember that, for Quine, the

indeterminacy of translation is not simply the possibility of alternative translations. Rather, it is the

conundrum of equally acceptable but incompatible translations (Quine 1960, 73). Quine characterizes

the incompatibility of translation in terms of logical incompatibility (i.e., contradiction). However,

when we liberate equivalence in translation and understand it relative to text-types, standards for the

compatibility of translations are relativized to the text-types. It thus makes no sense to complain of an

aesthetic or poetic translation of a certain ST because it fails to maintain the philosophical

significance of the ST, any more than it makes sense to criticize the quality of a portrait for failing to

be a topographical representation of the same person.

While text-type theories thus determine translation in their respective fields, there may be

controversies of interpretation that have a direct effect on how texts are translated. Scholars in their

respective disciplines ultimately resolve these controversies by further research, argument, and by the

production of influential translations. We recognize those who produce translations under the

guidance of specialized disciplines as translators, though no translator can ever claim full credit for

their work, as it is parasitic upon the research and consensus of broader disciplines that resolve

controversies around text-types. None of this happens in a vacuum, however, sealed off from wider,

interdisciplinary and philosophical debates in academia, as no area of study—and no text—is truly an

island. Philosophy as the traditional mother discipline in the West and the final resort of theoretical

72
perplexity can contribute something (though probably less than most philosophers of language may

think) to text-type theories of translation. Of course, the relatively young discipline of translation

studies will have much to offer as well. However, from the normative perspective of our institutions

of text-types, there are no such things as unspecialized translators at large who are competent to

translate. Good (i.e., determinate) translations are produced by those who have a special expertise,

fostered institutionally, in certain text-types. 48

Paradoxes of rule-following that render any interpretation one cares to give of past practice as

consistent with it do not affect the account I am providing here for the account is institutional. A

characteristic of institutions of the type I am highlighting is that they are comprised largely of

authorities who are not experts but practical authorities, charged with the day to day functioning of

the institution. (An expert authority is someone we listen to because they know better, such as our

mechanic. A practical authority is someone we listen to because they have power over us regardless

of whether they are right or wrong, such as a police officer, judge, professor or teaching assistant.)

Thus, most philosophers, for instance, are not really experts on the text-type of philosophy or the

whole institution of philosophy as it spans continents. Indeed, most philosophers know very little

about the history of philosophy and are usually concerned with narrow regions of its history. They

tend to confuse philosophy done in certain sub-traditions with philosophy as such in part because they

do not explicitly understand the underlying textual norm that underwrite its determinate translation—

norms that abstract from the particular style of doing philosophy and sets out philosophy as such—

though they will tend to implicitly employ the text-type norms of philosophy in teaching students

about philosophy. (It also doesn’t help that they tend to be very uncurious about philosophy from

alien traditions.) As a result, they can be incorrect about what philosophical discourse is like and we

can judge this error in light of past practice, defined by the dominant decisions of past practical

authorities in philosophy. After some time, wide spread error in the manner in which authorities

conduct themselves in an institution can change the very character of the underlying norms, but in the

73
short term we can understand their activity as a departure from past practice. It is because past

practice is defined by those who have power to alter the face of the very institution (and in the case of

text-type institutions, this pertains to decisions about what types of texts are to be studied and prized,

and how to appraise and judge students) and not simply those who are expert, that there can be

something like an objective past of an institution that sets down norms for the continuation of the

institution. This is exactly the way it is with the institution of law. Not all interpretations of past

practice are consistent with the history of judicial decisions in a country, because past practice is

defined not by experts, but by judges who have an authority to institute their decisions, though such

decisions may be wrong in light of past judicial decisions. After some time, the preponderance of

such decisions in a certain direction changes the course of the institution, but in the short term we can

understand it as a deviation from the past. In the case of text-type institutions, the underlying norms

just are the text-type and innovations in text-types happen only after long, widespread deviations from

past practice.49

Objection 3. It seems that LET attempts to solve the problem of indeterminacy of

translation by ignoring what Quine takes to be the problem: namely that there is no empirical data that

settles problematic controversies in translation. The claim that seeming indeterminacy is settled by

the quasi-legal nature of institutions that preside over text-types is an admission, of sorts, that there is

nothing objective at stake in the controversy.

Response 3. If by “objective” one means that there is a fact of the matter that individuals

can be wrong about and that experts can and do converge upon, then LET paints a picture of

translation that is objective. Moreover, controversies of translation, on LET’s, account, are resolved

on the basis of the facts of the individual text considered apart from a text-type, the relevant text-type

features, past precedence in translation and institutional decisions that set down conventions in

translation to deal with novel controversies. All of these are issues that individuals can be mistaken

about and matters that experts in the relevant institutions converge upon. If by “objective” one means

74
“thoroughly divorced from the decisions of persons,” then LET does not paint a picture of translation

that is objective for translation on its account is contingent upon the decisions of translational experts

in text-type institutions to introduce, for instance, new conventions in translation to create TTs that

are equivalent to STs. However, it is difficult to understand how such a criterion of objectivity can be

relevant to any theory of translation, for translation is a textual process, and texts are our artefacts.

Indeed, even scientific knowledge, empirical as it is, will depend upon the decisions of experts, to the

extent that expert decisions contribute to the institutional knowledge of science with respect to such

issues as scientific methodology and procedure. Decision-independent objectivity may be elusive in

general.

Objection 4. Your putative solution to the indeterminacy of translation problem is weak on

details. How would you deal with putative cases of indeterminacy?

Response 4. To clarify, indeterminacy is different from the case of alternative translations. If

two translations, A or B, are merely alternatives and not indeterminate, then both A or B may be

correct. The disjunction is thus inclusive. If A and B cannot both be correct, though we have equal

reason to think both are correct, or as correct, then the choice between them is indeterminate. The

disjunction in the case of indeterminacy is exclusive. LET, as has been noted, may likely give rise to

cases where there are alternative translations, either because they are produced by the same

institutionally sanctioned text-type theory, or because alternative institutions provide divergent

translations of the same text according to different text-types. The former will arise quite frequently,

and ought not to be a worry for LET any more than the analogue was a worry for Quine, for the

disjunction here is inclusive. It might also arise when, for instance, classicists or experts in literature

produce one type of translation of, say, a classical text in Chinese, with the aim of retaining its poetic

and artistic merits, while another group of scholars, say philosophers, are interested in translating the

text for its philosophical importance. The translations will certainly diverge. But there is in reality no

conflict between the two translations for they both preserve different text-type features of the ST—the

75
disjunction here too is inclusive. The conflict would only arise if we forgot their text-type-theoretic

relevance and compared them as though they were translations of the same sort. But this would be a

mistake in reading the translations, not a conflict between the translations themselves. There is the

seeming possibility that groups of scholars from differing disciplines might haggle over a text to

make their reading the default approach, and such an institutional crisis would be an unfortunate

affair that arises from an ignorance of the role of text-types. It would be a pseudo disagreement on par

with the disagreement over whether a soccer ball is spherical or inflatable.

It is thus difficult for me to think about even a plausible case of indeterminacy of translation.

(That Quine could not or did not provide a putative example of the indeterminacy of translation is a

point rarely noted.50). As this may be a failing in imagination, I will consider Mark Lance and John

O'Leary-Hawthorne’s fictional and putative example of indeterminacy of translation. The conditions

of this putative case of indeterminacy are as follows:

(1) Translators who hail from a community that speaks TL1, are inclined to translate an SL
word a as b, while translators who hail from a community that speaks TL2 will translate a as
c

(2) Translators from TL1 and TL2 will not be able to translate b and c by each other, though
both can translate them as a.

(3) These translations would result in divergent truth values of observational sentences,
depending upon which translation they appear in (or as Lance and O’Leary-Hawthorne put it,
“different observational sentences” of the SL “come out true under the respective
translations” ) (Lance and O'Leary-Hawthorne 1997, 44-47).

To flesh out the example, Lance and O’Leary-Hawthorne ask us to consider three linguistic

communities—corresponding to three languages that miraculously have the same ostensible

vocabulary and syntax, which I shall call the SL, TL1 and TL2. In Salem, where the SL is spoken, the

word “witch” is used to refer to a group of women who dabbled in demonology, natural medicine,

76
pseudo black magic and midwifery. TL1 and TL2 both have a word “witch” but the extensions differs

from that of the Salem case. In TL1, “witch” refers to practitioners of genuine magic (not all of them

women), while in TL2 “witch” refers to practitioners of “feminist science.” According to Lance and

O’Leary-Hawthorne, the cultures of TL1 and TL2 evolved their meanings of the word “witch” beyond

a meaning that is wholly commensurate with what is found in Salem, as their respective populations

of witches refined their practices. Given that these meanings of “witch” in TL1 and TL2 differ today,

we can call their respective words “witch”1 and “witch”2. I will call “witch” in the SL from Salem,

“witch”0. This scenario putatively results in indeterminacy because, “witch”1 and “witch”2 cannot be

intertranslatable—though according to Lance and O’Leary-Hawthorne, “witch”0 from SL is plausibly

translated as “witch”1 or “witch”2—and because the observation sentence “there are witches in

Salem” will have a different truth value, depending upon whether it is “witch”1 or “witch”2 that is

used in translation. In the case of using “witch”1, it would be false, for the witches of Salem practice

pseudo magic, not genuine magic as per TL1’s account of “witch”1, while if it is “witch”2, the

sentence would be true, for the Salem conception of a witch and the TL2 conception share an

emphasis on feminist science.

In order to get any Quinean case of indeterminacy of translation off the ground, one requires

a few assumptions, and one major omission. The major omission is the role of text-type theories in

translation. But this in turn is related to an erroneous assumption, that translation is an effort to

establish correspondences between languages, not texts. On the assumption that the business of the

translator is to translate languages, and not texts, one is lead to another erroneous assumption: that the

translator is constrained by some putative ordinary meaning of a word in a language. The error here

however is in the view that there is something like the correct meaning of a word, abstracted from its

textual environs.

The cotextual sensitivity of sentences is an issue that Lance and O’Leary-Hawthorne do not

address, and the putative problem of indeterminacy that they raise might not even arise if the text-type

77
of concern were, say, a poem, where the sentence “there are witches in Salem” serves a function that

is unaffected by alternative substitutions of “witch” to “witch”1 and “witch”2 in TT1 or TT2. Lance

and O’Leary-Hawthorne however do give some indication of the type of text-type that is relevant to

their thought experiment. For them, the scenario presents a problem for anthropologists from the

cultures of TL1 and TL2 studying Salem. Thus, the relevant text-type appears to be the ethnography.

The ethnography, above most other text-types, is the most immune to the type of problem that

Lance and O’Leary-Hawthorne envision because key native concepts are most always preserved in

them, introduced with framing comments, or what I call “text-moulding.” It is a strategy that

anthropologists have employed for some time: nothing new here (cf. Messick 1993). The ethnography

thus most closely resembles Schleiermacher’s ideal form of translation, where the TT reader is made

to do all the work of approaching the ST’s concepts. If the concept ‘witch’ as found in Salem is of

ethnographic importance for anthropologists studying Salem, then when they compose their

ethnographies they will include the appropriate framing comments (footnotes, glossaries, or other

marginalia) that explain how the word “witch” in the ethnography functions, in contrast to its

functioning in other texts that it is domestically associated with. Ethnography, like so many

important text-types, thus involves a process of educating the reader, not simply pandering to their

preconceptions. The textually moulded rendition of “witch”1 and “witch”2 in TL1 and TL2 thus

become intertranslatable, and the apparent conflict that Lance and O’Leary-Hawthorne envision

disappears.

Lance and O'Leary-Hawthorne’s approach to solving putative problems in translation is to

understand translation as an effort to “form a single community where previously there were two”

(Lance and O'Leary-Hawthorne 1997, 20). I think there is something admirable about this approach,

in so far as it recognizes that success in translation may require change in practices. However, the

error that this approach makes is to presume that translation involves the whole community. This is

virtually the same error that Quine makes, in assuming that translation concerns a word for word,

78
sentence for sentence, matching of languages. It is obvious to anyone who spends any time

translating, that it is not languages or communities that are translated, but texts. The problem of

indeterminacy as Quine conceives it, and Lance and O'Leary-Hawthorne provide an example for,

only arises when one hypostasizes a SL and TL as the texts to translate. In this unfortunate move, the

linguistic philosopher deprives herself of a tool to effect translation: language. Really, we should

recognize that the objects to be translated are texts, and regard language as a mouldable medium to

achieve these ends.

Objection 5. There is something suspicious about using the Quinean scenario as a foil to

argue for LET, for it seems that Quine in his concern with translation is actually interested in

something very different from you and, admittedly, traditional translators. Traditional translation is a

three party phenomenon, where a translator mediates between a ST author and a TT audience.

Quine’s thought experiment concerns a two party scenario: the linguist, concerned with a translation

for their own sake only, and a native informant. Perhaps “translation” is not the best label for this and

that is why Davidson wisely replaced the idea with interpretation, which does have a two party

structure. Given that the concerns of Quine and Davidson are in some sense to interpret people, and

languages, and not texts, why should we agree with you that Quine and Davidson are incorrect about

translation as you understand it? Is it not more reasonable to conclude that Quine and Davidson are

simply not talking about the issues that you are interested in and thus cannot be used as foils to

articulate your view?

Response 5. I believe my response to the problem of indeterminacy posed by the Quine-

Davidson tradition shows how it is that their view is not divorced from the concern of ordinary

translators. For if their views pertained to a purely autonomous domain of inquiry, then LET should

not have anything to say about how a problem that they identify can be resolved. But LET does have

things to teach us about how we can dream up cases of indeterminacy of “translation” a la Quine and

how we can resolve them. We dream them up by omitting text-types from our analysis and

79
hypostasizing languages into texts. We solve them by re-understanding texts as the objects of

translation, and languages as the defeasible rules of syntax, vocabulary and other building blocks that

texts can be constituted by.

While Davidson does move to the idea of “interpretation” in his articles, it is clear that it is a

successor concept to Quine’s “translation” and moreover Davidson himself sees himself as defending

the Quinean conception. Thus, even if one corrects the error of Quine’s taxonomy, there is enough

connection between the Davidsonian project and the Quinean project for Davidson’s views to be

relevant to our concerns. For, the solution proposed here solves a problem that they identify, namely

how we know when we’ve got a translation right.

In short, LET shows that Quine and Davidson, and others who have written on the difficulties

of translation, are missing a piece of the puzzle, namely text-types.

II.4.6 Indeterminacy of Translation And Difficult Texts

One might argue that LET does not preclude alternative translations that exemplify the

indeterminacy in translation. If there are some text-types that specify criteria of equivalence that

cannot be achieved without compromising some relevant text-type features, such text-types would

yield translations that are indeterminate for no resulting translation would be obviously successful

though they may each be different, seizing upon different features of the ST.

For a picture of what an untranslatable text-type might look like, consider W.D. Hart’s

observation that one cannot preserve self-reference, reference and truth-value simultaneously in a

translation. Hence, if the sentence “The first word of this sentence has three letters” were translated

metaphrastically into French, it would be false, for the first word of the sentence would have two

letters (Hart 1970). If a text-type demanded that all three of these features be inviolable, we might

conclude that the translations of such texts are indeterminate, for the resulting translations could not

be assessed as correct nor completely wrong, though we have no way to decide between them.

80
 The best candidates for indeterminately untranslatable texts a la Hart’s example are

metalinguistic texts, where language itself becomes the object of discussion. These may be difficult to

translate, but their translation does not appear to be indeterminate, for what we do is simply subsume

such texts under a text-type that is not inherently untranslatable. Something (probably self-reference)

will be compromised in the translation, but in accordance with a principled decision that can make

appeal to the text-type in question. Moreover, what I have called text moulding becomes

indispensable. While self-reference or what is often called reflexivity might seem to be lost, it can be

regained by the parenthetical qualifications in texts to correct our understanding of reference to

preserve all the relevant features. Thus, for instance, we might have text-type reasons to translate

Hart’s sentence from French into English as “The (le) premier mot de cette phrase a trois lettres” or

perhaps “Le premier mot de cette phrase (the) a trois lettres”.

When we subsume a text with much metalinguistic content under a text-type that

compromises self-reference, what we thereby decide in effect is that it was more important for the ST

author to say something true than for her to be able to say it reflexively.

In some cases, such decisions to sacrifice reference or self-reference in order to retain the

putative truth of claims being made is licensed by the very text-type employed in translation.

Consider G.E.M. Anscombe’s translation of Wittgenstein’s Philosophical Investigations (1958) (my

underlining):

I.381 Wie erkenne ich, daß diese Farbe Rot ist?-Eine Antwort wäre: “Ich habe Deutsch gelernt.”

I.381 How do I know that this colour is red?-It would be an answer to say: “I have learnt English.”

Here Anscombe makes the appropriate adjustment in the reference of the sentences from German to

English in order to preserve the putative truth of the claim Wittgenstein is making in light of his

philosophical argument, which generally at this point centres around elucidating the philosophical

concept of MEANING in terms of UNDERSTANDING, and UNDERSTANDING in terms of knowing one’s

81
way around a linguistic practice. Later, she makes a slightly different adjustment but in light of the

same considerations (my underlining):

I.134 Betrachten wir den Satz: “Es verhält sich so und so”-wie kann ich sagen, dies sei die allgemeine
Form des Satzes?-Es ist vor allem selbst ein Satz, ein deutscher Satz, denn es hat Subjekt und
Prädikat. Wie aber wird dieser Satz angewendet-in unsrer alltäglichen Sprache nämlich? Denn nur
daher habe ich ihn ja genommen.

I.134 Let us examine the proposition: “This is how things are.” How can I say that this is the general
form of propositions?-It is first and foremost itself a proposition, an English sentence, for it has a
subject and a predicate. But how is this sentence applied-that is, in our everyday language? For I got
it from there and nowhere else.

This adjustment not only changes the reference but also comes very close to an adjustment in

self-reference as the sentence being quoted and the one that comments upon it are in the same

language and hence, the possible cultural peculiarity of the sentence being quoted plays no part in the

point being made. The point Wittgenstein is making here is about understanding the meaning of any

sentence, not particularly German or English sentences. This contrasts with other cases where

Wittgenstein does wish to bring self-reflexive attention to the cultural and linguistic peculiarities of

his perspective, as a speaker of German. Anscombe handles such cases by means of text-moulding:

538. Es ist ein verwandter Fall (obwohl es vielleicht nicht so scheinen möchte) wenn wir uns z.B.

darüber wundern, daß im Französischen das prädikative Adjektiv mit dem Substantiv im
Geschlecht übereinstimmt, und wenn wir uns dies so erklären: Sie meinen “der Mensch ist ein
guter”.

538. There is a related case (though perhaps it will not seem so) when, for example, we (Germans)
are surprised that in French the predicative adjective agrees with the substantive in gender, and when
we explain it to ourselves by saying: they mean: “the man is a good one.”

Here, the fact that the “we” being referred to are German speakers is important, as is, to some

degree, that this is Wittgenstein’s self-reflexive comment as someone who understands himself as a

member of the linguistic community of German speakers, and thus Anscombe preserves both the

82
rigidified reference along with self-reference by the device of text-moulding, by parenthetically

qualifying the reference of “we” with “German”. These translations are not the result of haphazard

decisions on Anscombe’s part. Rather, she is making the necessary adjustments so that Wittgenstein’s

philosophical point about the connection between meaning, understanding and a linguistic practice is

preserved through translation. She is translating Wittgenstein according to the text-type of

philosophy, not, say, the text-type of meta-linguistic texts. In every case she is preserving some

feature of the ST in the TT, even when she switches reference and or adjusts self-reference in order to

maintain the putative truth of the claim of the ST, but these decisions are not based upon atomic

appreciations of each sentence or word of Wittgenstein’s text, but in light of the philosophical theory

that Wittgenstein articulates with philosophical concepts such as MEANING and UNDERSTANDING.51 If

Anscombe were treating Wittgenstein’s text as a text about a language, and not as a philosophical

text, the adjustments in reference that she would have to make would be quite different.

No text is untranslatable. And not only is no text untranslatable, indeterminacy is simply

apparent and a function of our inability to find the right text-type (or combination of such types in the

case of anthologies) to mediate translation. The main reason that we should be surprised at the

prospect of an untranslatable text is that text-types are very slim: they do not specify the provenance

of texts, or what content they must save. Rather, they specify very general and abstract rules

according to which we can track features of a ST to be preserved in a TT. If it were anything more

robust, the very text-type would be a text itself, and any translation of a text under its guidance would

be a mere articulation of the very type. But by virtue of being a “type” the text-type abstracts from

much content. Given its selective approach to translation, it would be very peculiar if we would not

be able to construct a TT that matches a ST given the text-type-theoretic constraints. Perhaps we

could only produce a translation at the cost of text-moulding, but this typically means that what we

are translating constitutes a type of conceptual innovation or expansion of the mental horizon of the

83
target audience. As a point of translation is to learn, we shouldn’t be shy of such mind expanding side

effects of translation.

I take it that in many respects, the argument that I have presented here is consistent with the

view presented by the translation theorist Anthony Pym in his recent and important book The Moving

Text, where Pym argues that “the concept of translational equivalence deserves rather more respect

than it is currently accorded” (Pym 2004, 51).52 Pym concludes that “translation is not a mapping of

one function onto another” but rather “it is a productive function in itself.” Where I part ways with

Pym is over his view that “[t]ranslational equivalence is thus ultimately determined by what

translators actually do or have done in the past, and not by abstract comparisons between falsely

discrete languages or cultures” (Pym 2004, 62). The problem with Pym’s characterization of

translation is the notion that what translators do or have done sets ultimate standards of equivalence:

this would imply that translators could do no wrong. If this were so, there would be nothing objective

about translation. But there is. In each case of real translation, where translation involves texts, not

words or sentences treated in the abstract, we have not only the text and its many significances that

present us with data, but also the text-type to help us choose in a principled fashion what features of a

ST are to be preserved in a TT. We justify translations by our openness and self-conscious reflection

on the conditions of accurate translation. But justification is not the same as accuracy or objectivity,

but merely an assurance of the accuracy of translation. Equivalence is thus responsive to the features

of the ST in light of the text-type. In these cases, translators make TTs that are equivalent with STs,

but for good and objective reasons that can be consistently applied across many cases of translation

relative to the same type. Translation is not a projection of the whims or interpretations of a translator

on a target audience, but rather the principled construction of an equivalent text in a TT on the basis

of text-type-theoretic considerations.

84
 II.4.7 Responding to Davidsonian Indeterminacy of Translation

The preceding is an argument against the indeterminacy of translation thesis presented by

Quine. What of Davidson’s view that despite the wide-ranging application of the principle of charity

there remains a residual indeterminacy? A similar argument can be made against Davidson’s view. If

criteria for equivalence are regional and serviceable for specific text-types, then we should generally

have recourse to some principle (and device of text creation) that can narrow the question of

equivalence in translation. LET thus follows Davidson’s innovation of applying principles or criterial

considerations in narrowing the range of possible translations. But whereas Davidson regards the

principle of charity as the major criterion of translational equivalence, LET is open to many more that

are sensitive to the varying text-types. Differing but equally acceptable translations would be

isomorphic instances of the same text, differing only in features deemed to be inessential by the

relevant text-type theory.

II.4.8 LET and Radical “Translation”

As Davidson paints the picture, we have relatively little to go on when we’re radically

interpreting or translating someone. For Quine, we have even less. In the impoverished state that

Quine leaves the radical translator, he implies that life would be easier if the translator had recourse to

the knowledge about the intentions of the speaker, but these are not forthcoming given his austere

conception of admissible evidence.53 Davidson’s innovation is to reject the notion that empirical

evidence is the ultimate arbitrator of translation and interpretation and to supplement the project of

interpretation and translation with a criterial objective, by broadening the application first of his

principle of charity, and then later by widening it in his principle of rational accommodation.54 LET

as I’ve presented it furthers this move in the direction of criteria, and I believe it has the effect of

augmenting the project of radical interpretation and translation. In the context of “radical

85
interpretation” what we do is look for behavioural or contextual cues that suggest that some text-type

or another is appropriately applied to a text (including utterances) in a certain context, and then we

use the text-type theory to translate the text we are presented with interlingually or intralingually

(depending upon our linguistic kinship with the interlocutor). This is one way that we satisfy

ourselves that we understood what our interlocutors have said. What the radical translator/interpreter

does, on LET’s account, is try out different text-types in the radical context. The goal is to abstract

from discourse texts (which are often extempore), and apply text-type theories to translate them. The

decision is justified, on LET’s account, because it is our institutional standards of text-types that we

appeal to in deciding how to translate a text. One might even understand this approach as an

extension of Davidson’s principle of rational accommodation.

If new text-types are warranted because our old one results in our discarding the bulk of the

texts we are trying to translate as semantic noise, we can institutionally revise our text-type theories

or create new theories to track the novel texts we are confronting. In general, however, the results of

applying text-types to texts abstracted from discourse that we are attempting to radically translate will

not provide us with results that we will happy with in the long run. We will revise our translations in

light of continued research and likely these initial steps will contribute to the first lexicons of the new

language we are encountering. These first baby-steps at translation are really the ground work of

linguistic anthropologists and philologists in the early days of studying a culture. The goal in these

initial days must be to set some understanding of the complexities of the semiotic devices of a culture.

We can be satisfied that we have come to this point when we have acquired sufficient grasp of the

language to communicate in conversation with those we are interpreting, and the revisions that we

turn up with are minor. Lexicons and the like derived from this effort will set the ground for later

translations that will be satisfactory for these later translations will be based upon a fulsome

appreciation of the complexities of the source language in light of text-types, thus allowing scholars

to recognize text-type patterns in texts that would have been overlooked in the groping days of radical

86
translation. The ground work set for real, determinate translation of texts cannot be accomplished by

one person alone any more than one person alone can provide the observations to confirm any

empirical theory. What we require for determinate translation is often a very detailed understanding

of potentially conflicting “translations” of words and sentences that will allow us to understand the

full range of contributions they can make to texts. Text-types are made possible by these multiple

possibilities.

II.5. Skopos Theory and LET

The previous two sections attempted to address the philosophical criticism that translation

conceived in terms of equivalence is paradoxical or problematic. Against the argument that

translation is indeterminate, I put forward an account of translation that makes use of equivalence

relative to text-types, which are defined, criticized, improved and employed in the context of

institutional categorizations of texts. The argument thus makes use of the idea of institutional

expertise to resolve controversies in translation. There is an important theory of translation in the

literature that also makes use of the idea of expertise to solve problems of translation. It is the

influential, German Skopos Theory of translation.

According to Hans J. Vermeer, one of the Skopos School’s major proponents, translation is to

be distinguished from trans-coding, which attempts to transfer a code from a ST to a TT. Translation,

in contrast, unlike trans-coding, is not “retrospectively oriented towards the source text” but rather is

oriented towards the target culture (1989, 175). What normatively guides translation is thus not

always a search for fidelity with the ST, but rather a skopos that is a result of the negotiation between

the translator—who is in all matters related to translation an expert—and the agent who commissions

the translation. “Skopos,” which literally means ‘action,’ is here understood as including both the

purpose of a translation and the translation process itself.55 The commissioning of a translation, on

this account, assigns such a purpose both to the ST and the process of translation (subject to the

approval of the translator who is the final expert on “what’s what” where translation is concerned),

87
and the norms that govern a translation according to Vermeer are those set out by the commission of

the translation (1989, 174-184).

Vermeer thus provides an alternative, target side account of the norms of translation. Unlike

equivalence accounts of norms of translation, Vermeer’s Skopos Theory makes the norms governing

translation something that is up for negotiation between the agent commissioning a translation and the

translator. Equivalence between a ST and a TT according to any criterion are in the offing in judging

a translation, but only if such an objective is negotiated at the time of the commission of translation,

according to Vermeer.

From the perspective of LET or any equivalence conception of translation, the Skopos Theory

appears suspicious: it appears as a potential apology for the autocratic behaviour of translators,

instead of a normative account of translation. The problem seems to be that it is unclear what the

expertise of the translator according to Skopos Theory consists in. For LET, as I’ve argued for it, the

translator is an expert by virtue of her study and research on a text-type. The translator is thus not at

liberty to translate any ST into any TT she wishes. Rather, she has institutional standards of

translation to prima facie respect (subject to the critical reflection that text-type theories ought to be

continuously subject to).

Finally, as I’ve argued in the previous section, there will often be institutional pressure for us

to come to a consensus upon how a text is to be translated, and hence there will often not be

normative room for us to negotiate with a commissioner of a translation the shape of a translation.

However, there may be certain texts that we tolerate multiple text-type theories for. For instance, I

suspect that some types of poetry may be texts for which the relevant experts (literary scholars on

poetry) tolerate multiple text-type theories for reasons specific to the discipline. And hence, it may be

that some poetry can be translated into music, as Sallis (2002, 120-2) envisions, reproducing some

feature of the poem that is thought of as essential (as per relevant text-type) in a work of music.

88
 Where LET clearly departs from Vermeer’s Skopos Theory is in the latter’s latent

commercialization of translation. Vermeer’s theory assumes that the translator has a service that can

be commissioned (i.e., bought) by someone. On this understanding, the translator’s duty is to the

contract that is negotiated. According to LET, the translator’s duties are to faithfully employ a text-

type theory in translating, a theory that is deemed applicable to texts by scholars who are expert in the

relevant texts. Success according to LET is measured in translation in terms of increase of knowledge,

not financial profit. However, LET in one respect surpasses the flexibility afforded to translators by

Skopos Theory, for what Vermeer calls trans-coding may count as translation with respect to certain

text-types.

II.6. Functionalism

II.6.1 Functional Linguistics and Translation

In two previous sections I argued that LET survives philosophical criticisms of equivalence in

translation narrowly conceived. In the previous section on Skopos Theory I attempted to distinguish

LET from an influential theory that also recognizes the expertise of a translator as central to the

problem of determining translation. In the current section I will move to a dominant trend in

translation studies in the English language: Functionalism. Like LET, Functionalism recognizes the

central importance of equivalence in translation. Unlike LET, Functionalism rests upon a monistic

criterion of equivalence in translation. Recognizing the shortcomings of Functionalism will not only

bolster the case for LET, but it will also help us to conceptualize a text-type theory of philosophy.

J.R. Firth, an early and influential Functionalist, accounts for meaning thus:

What do the words ‘mean’? They mean what they do. When used at their best they are both affecting
and effective. A Martian visitor would best understand this ‘meaning’ by watching what happened
before, during, and after the words were spoken, by noticing the part played by the words in what
was going on. The people, the relevant furniture, bottles and glasses, the ‘set’, the specific behaviour
of the companions, and the words are all component terms in what may be called the context of

89
 situation. Meaning is best regarded in this way as a complex of relations of various kinds between the
56
component terms of a context of situation. (Firth 1964 [1930], 110)

According to the eminent translation theorist, Susan Bassnett, Firth’s conception of meaning implies

that:

In translating… it is the function that will be taken up and not the words themselves, and the
translation process involves a decision to replace and substitute the linguistic elements in the TL.
And since [a] phrase [in English] is, as Firth points out, directly linked to English social behavioural
patterns, the translator putting the phrase into French or German has to contend with the problem of
the non-existence of a similar convention in either TL culture. (Bassnett 2002, 27)

Bassnett whole-heartedly endorses Firth’s view, and provides the example of the problem of

translating the term “butter” into Italian from English to bolster the point:

When translating “butter” into Italian there is a straightforward word for-word substitution:
butter-burro. Both “butter” and “burro” describe the product made from milk and marketed as a
creamy-coloured slab of edible grease for human consumption. And yet within their separate
cultural contexts “butter” and “burro” cannot be considered as signifying the same. In Italy, “burro,”
normally light coloured and unsalted, is used primarily for cooking, and carries no associations of
high status, whilst in Britain “butter,” most often bright yellow and salted, is used for spreading on
bread and less frequently in cooking. Because of the high status of butter, the phrase “bread and
butter” is the accepted usage even where the product used is actually margarine. So there is a
distinction both between the objects signified by “butter” and “burro” and between the function and
value of those objects in their cultural context. (Bassnett 2002, 26)

The conclusion that Bassnett draws from these considerations is shocking. She writes:

The emphasis always in translation is on the reader or listener, and the translator must tackle the SL text
in such a way that the TL version will correspond to the SL version. The nature of that
correspondence may vary considerably… but the principle remains constant. Hence Albrecht
Neubert’s view that Shakespeare’s Sonnet “Shall I compare thee to a summer’s day?” cannot be
semantically [i.e. literally] translated into a language where summers are unpleasant is perfectly
proper, just as the concept of God the Father cannot be translated into a language where the deity
is female. To attempt to impose the value system of the SL culture onto the TL culture is dangerous
ground, and the translator should not be tempted by the school that pretends to determine the

90
 original intentions of an author on the basis of a self-contained text. The translator cannot be the
author of the SL text, but as the author of the TL text has a clear moral responsibility to the TL readers.
(Bassnett 2002, 30, my emphasis)

Let us bracket Bassnett’s nonsequitorial suggestion that the critic of Functionalist translation

is somehow especially concerned with divining authorial intention, more than the Functionalist. The

idea that translators always have a binding obligation to TL readers over the producers of the ST will

come as a shock to many sensitive theorists concerned with cultural representation, post-colonial

discourse and scholarly fields where the translator and scholar are conduits by which TL readers are

acquainted with the views of SL authors. Bassnett’s view that our obligation is only to the TL readers

in all circumstances has no obvious ethical foundation in any viable ethical theory to date and

promises to do nothing but further insulate TL readers in their own ethnocentrism. (Indeed, on such a

view, it seems that there is no point to translation at all!) Anthropology, the history of philosophy, the

study of religion and other important endeavours are destroyed on such an account of translation (cf.

Napper 1995; Tsering 1995; Mullin 1995, 221; Dorjee 1995).57 Bassnett’s view is not only pernicious

but irresponsible. But most importantly, it is undermotivated from the perspective of translation

theory, once we have text-types in view. The text-type provides us a way to bypass questions of

authorial intention in our first attempt to come to terms with how to translate a text. When we bring

text-types to bear in translation, we are not asking what an author intended, but rather what text-type

does a text display. Having made such decisions we can subsequently ask what an author might have

intended, and pursue the question through historical research. But this can only be accomplished once

we have some textual evidence for translating a text according to a certain text-type.58 And it is these

considerations, not how the text will be received, that are the primary considerations in translation.

Failing this, there is no objectivity or accuracy in translation.

91
 II.6.2 Nida’s Dynamic Equivalence and Translation Criterion

While Bassnett is a very influential author in translation studies, her view that translation

should always indulge TL values might seem mitigated if she were the only scholar putting her

weight behind it. However, Bassnett is neither the first nor the most influential of functionalists in

translation studies. Even Skopos Theory presents a form of functionalism that is divorced from any

explicit concern shared by Bassnett and propounders of LET—that STs and TTs must be equivalent

in some important fashion. For Skopos Theory, a TT is judged as adequate when it functions in a

certain way, determined by the negotiations of the translator and the commissioner of the translation

(Vermeer 1989, 174-184). The influence of functionalism in translation studies—particularly among

authors who write primarily in English—however appears to be a result of the far-reaching influence

of Eugene Nida.

Nida’s groundbreaking Toward a Science of Translating, with Special Reference to

Principles and Procedures Involved in Bible Translating (1964) is one of the most cited works in

translation studies over the last forty years. This influence may be a function of the breath and

scholarship of the work, which draws freely from linguistics, semantics and analytic philosophy. It is

by no means purely derivative, though much that follows in translation studies does appear to be

derivative of Nida’s magnum opus. In it, Nida distinguishes between two types of equivalence that

translation may seek to secure: formal and dynamic. Formal equivalence, according to Nida, “focuses

attention on the message itself, in both form and content”: “from this formal orientation, one is

concerned that the message in the receptor language should match as closely as possible the different

elements in the source language.” In contrast, dynamic equivalence “aims at complete naturalness of

expression and tries to relate the receptor to modes of behaviour relevant within the context of his

own culture; it does not insist that he understand the cultural patterns of the source-language context

in order to comprehend the message” (Nida 1964, 159).

92
 So far, so good, according to LET. Nida’s distinction between formal and dynamic

equivalence does not imply that any one approach to translation is superior to the other. Rather, Nida

explicitly claims that both approaches to translation have their valid application and that there are a

number of “intervening grades” between these two extremes (Nida 1964, 160). However, the dynamic

conception of equivalence is well suited to a functionalist outlook on meaning, leaving behind details

about the culture and values of the source culture in exchange for an emphasis of equality of effect on

readers. Despite his LET-like tendencies, Nida endorses a functionalist account of meaning (like

Firth’s) (Nida 1964, 37) and more importantly he proposes a functionalist criterion for evaluating

translations. He writes:

But all translating, whether of poetry or prose, must be concerned also with the response of the
receptor; hence the ultimate purpose of the translation, in terms of its impact upon its intended
audience, is a fundamental factor in any evaluation of translations. This reason underlies Leonard
Forster’s definition… of a good translation as “one which fulfills the same purpose in the new
59
language as the original did in the language in which it was written.” (Nida 1964, 162)

II.6.3 Criticisms of Functionalism

Functionalism cannot be the universal approach to translation, for at least two reasons.

First, functional translation theory goes against the goals of various text-types. For instance,

neither scientific nor philosophical texts are composed with the overriding concern for quick and easy

digestibility of their content for any reader. This is because it is more important to be right, in science

and philosophy, than to be accessible. Functionalism would however aim to improve easy uptake of

the TT at the expense of foreign ideas and values of alien scientists and philosophers. Similarly, the

anthropologist concerned with ethnography and translating claims of the informant into the language

of the ethnography must also find the Functionalist proposal to substitute cultural references of the TL

for those of the SL objectionable and counterproductive. The anthropologist rightly treats the textual

material that ethnographies are culled from as intrinsic expressions of the worldviews and values of

93
informants and their cultures. To fail to attempt to preserve these features in translation would be

tantamount to disregarding the anthropological text-type features of such texts.

A second reason that Functionalism cannot be regarded as a universal conception of

translation is that, in its Firthian moments, it makes use of a suspect conception of meaning.60

Meaning, according to Firth, is what a word does. In the Analytic tradition, this is sometimes known

as the causal theory of meaning61. The projection of the “causal” nomenclature back on to the

Functionalist tradition is telling. For instance, on the causal theory of meaning, meaning is a natural

and psychological process that occurs with linguistic stimulus of a certain sort. If translation were

built around this conception of meaning, then translation would attempt to replicate a psychological

reaction to a ST through the construction of a similarly efficacious TT, requiring no work on the part

of the TT reader, aside from the will to be the audience of the TT. Moreover, on such a view,

translation theory would be aided by disciplines such as psycholinguistics and the social sciences in

general. Functionalism in translation does idealize the passivity of the reader of the TT and also has a

tradition of conceiving of translation as a type of social science (particularly in the cases of Catford

and Nida). Such a conception of meaning leads to absurdities when applied across the board. Some

time ago, the causal theory of meaning was put forward by C. K. Ogden and I.A. Richard’s The

Meaning of Meaning (1923) and Bertrand Russell in his Analysis of Mind (1921). Wittgenstein in his

unpublished notes brings to the fore the trouble with this conception of meaning by analogy. He

writes: “‘If I wanted to eat an apple, and someone punched me in the stomach, taking away my

appetite, then it was this punch that I originally wanted” (quoted in Monk 1990, 291). We could easily

construct such a Wittgensteinian example that was not an analogy: if a certain literary work when

assigned to students (say a Shakespearean play) has the effect of making them decide that literature is a

waste of time, then it seems that the meaning of the literary text is that ‘literature is a waste of time’.

Wittgenstein’s well-taken point is that the problem for the causal theory of meaning is that effects of

causes are not unique. It may be that a certain word or text has a certain effect on its readers, but the

94
effect is not unique to the word or text nor is it even necessary, as subjective impressions from reading

experiences vary widely. Other words and works and even other events can cause the same

psychological effects on readers. Meaning is certainly not always effects language usage has on

readers and it is certainly not coextensive with subjective effects of readings. Perhaps, in some cases,

meaning will also amount to such an aesthetic or personal impact, but meaning as such cannot be

reduced to such effects.

II.6.4 Functionalism, And Inter-Cultural Communication

Functionalism is often a suspect translational strategy even in the apparently paradigm cases

of its importance. Eugene Nida, for instance, seems to think that a functionalist, dynamic approach to

translating is well suited to biblical translation. Nida presents the example of Psalm 1.1., which has

sometimes been translated with formal equivalence in mind. So translated, it reads: “walketh not in

the counsel of the ungodly, nor standeth in the way of sinners, nor sitteth in the seat of the scornful.”

However, he notes that people often read “council” into “counsel” and misunderstand that “standing

in the way of someone” in Hebrew meant ‘associating with sinners’ while “sitting in the seat of the

scornful” meant ‘joining in and making fun of God’ (Nida 1997, 189). If such passages are to be

understood, without distortion, it seems that literal attempts at reproducing the original must be

abandoned.

In response to Nida’s examples, we might question along with Peter Newmark the wisdom of

Nida’s, and functionalism’s, tacit principle that translations must be instantly intelligible. Moreover:

The translation theorist has to raise the question, in considering Nida’s dynamic equivalence, not
only of the nature (education, class, occupation, age, etc.) of the readers, but of what is to be
expected of them. Are they to be handed everything on a plate? Are they to make any effort? Are
they ever expected to look a word up in a dictionary or an encyclopaedia? I have no wish to question
the appropriateness of the Good News Bible translation, and obviously the translation of any
performatives (public notices etc.) must also be instantly intelligible. However, I am writing against

95
 the increasing assumption that all translating is (nothing but) communicating, where the less effort
expected of the reader, the better. (Newmark 1989, 133-4)

Newmark here touches upon an issue of vital importance in translation: how much effort must the

target audience be expected to expend? No good answer to this question would prescribe the same

level for all types of texts. Literary texts, it may be argued, would require a higher degree of passivity

than scientific or philosophical texts in order to facilitate their aesthetic objectives. But religious texts

often sit in the peculiar crossroads between literature and philosophy, and to the extent that

philosophical ideas of morality and metaphysics are essential to religious doctrine, it would seem that

religious texts should demand of their readers some effort in uptake. Hence, for this reason,

functionalist translation strategies (i.e., Nida’s dynamic equivalence as a translation ideal) as a blanket

policy in the context of religious texts—including the Bible—seem to be inappropriate.

There is an additional reason why functionalist translation strategies in the context of

religious texts is inappropriate, a reason closer to Newmark’s concerns: religious texts have the

intrinsic characteristic of expressing culturally specific values, and it seems that one cannot become

acquainted with the history of a religious tradition without being exposed to expressions of such

mores. Nida’s functionalism deprives us of this acquaintance.

Against the back drop of Christianity’s proselytizing zeal and historical intolerance of

cultural and philosophical differences, it seems that Nida’s approach to translating the Bible has the

effect of making Christianity less alien, and thus, less intrusive on target culture sensibilities. It thus

can be seen as a praiseworthy departure from more traditional Christian approaches to cultural

outreach (i.e., threats of eternal damnation and coerced conversion by the state in domestic and

colonial contexts). However, Nida’s approach to translating the Bible also has the effect of concealing

the alien nature of the Bible both to cultures that are the target of missionary intentions and modern

societies. There is thus something disingenuous about propagating dynamic translations in missionary

contexts.

96
 A related danger for functionalist translation strategies in the context of religious texts is that

the translator becomes the apologist for the source culture, or worse, its white-washer. Consider the

problem that R. A. Megrab faces in translating Hadith. “Hadith” refers to both a corpus that contains

sayings and describes the actions of the Prophet Muhammad, and to individual verses of the corpus.

Megrab raises the example of one Hadith, which literally translates, “No man shall be taken to task

for beating his wife.” Megrab recognizes that this claim in English would meet with widespread

disapproval in the target culture, but that it was quite acceptable in the source culture. And thus, a

translation must be produced that preserves the acceptability of the Hadith. Megrab notes other

Hadiths that recommend against striking one’s wife on the face and exhort the kind treatment of

women. Thus he reasons that “beating one’s wife” must be understood as communicating not

aggression but the effort to guide and change one’s wife (as the Hadith makes it clear that it is the

Husband who is in charge of the moral guidance of the wife). Thus, the Hadith ought to be translated

as “No man should be taken to task for acting with the intention of reforming his wife” (Megrab

1997, 234).

There are several problems with Megrab’s functionalism. First, it presumes that Hadith

cannot be riddled with contradiction. However, religious texts are most always riddled with

contradictions because they are typically collations of separate smaller texts, and while the smaller

texts might exemplify a tolerable degree of consistency, the anthology does not. (The Bible’s tension

between the Old Testament “an eye for an eye” and the New Testament “turn the other cheek” is just

one example of such a contradiction). Thus, it may very well be that one Hadith exhorts the kind

treatment of woman, while the other sanctions beating them. But, perhaps the Hadith are to be treated

as a unified text because they have the same author. If so, we still cannot decide to translate

Mohammad’s sayings simply to preserve their palatability. The moral sensibilities of Mohammed’s

day and place differ from those of ours, and moreover, from Anglophonic culture on the whole,

steeped as it is in an English tradition of liberalism. Thus, Megrab’s effort to make the Hadith

97
acceptable to readers of English is in short an attempt to project the moral standards of Anglophonic

cultures on to Mohammed. But this is absurd: there is no a priori reason to expect that Mohammed

was trying to appeal to anyone’s moral sensibilities except his own. The question of what

Mohammed’s intentions were cannot be settled without first deciding what type of text the Hadith is.

If the point of Hadith’s is to give moral guidance, then it is best treated as an instance of normative

discourse and translated according to the appropriate text-type, which I shall argue is the text-type of

philosophy. However, if this is how we are to treat the text, then the issue for us to determine is not

what our moral sensibilities are, but rather what the moral theory in the Hadith are, and this may very

well offend our standards. To determine this requires a systematic study of the text, and such a study

may lead us to conclude that the moral theory forbids all violence against wives, in which case

“beating his wife” may best be treated as an idiom that is functionally equivalent to “coercing his

wife”, or some variant. But this could not be decided on the basis of worries about the palatability of

the translation, or on the mere comparison of one Hadith against others. Rather, it would have to be

concluded on the basis of a systematic study of the moral theory expressed in the Hadith themselves.

Susan Bassnett’s claim that a disavowal of functionalist strategies amounts to the hubris of

attempting to discern the intentions of the ST author is a red herring (Bassnett 2002, 30). What is at

stake is the problem of determining the text-type. This consideration is cotextual and institutional

(…is this a text of science? of philosophy? a narrative? a comedy? what text-type feature does it

display most prominently?…). “Cotext” is a term used by some translation theorists that is to be

distinguished with what is now discussed under the heading of “context.” “Context” has come to

mean something very specific in the philosophical, linguistic and anthropological literature that is a

departure from its traditional, etymological meaning of “with text.” According to the Oxford English

Dictionary, “context” in the 16th century CE meant, “The whole structure of a connected passage

regarded in its bearing upon any of the parts which constitute it; the parts which immediately precede

or follow any particular passage or ‘text’ and determine its meaning” (Oxford University 2006). This

98
is very much what “cotext” now designates. “Cotext” concerns textual matters, of structure,

composition, and type. The reason that textuality is so important for translation is that translation

must be sensitive to issues of textual structure, order and composition, particularly in light of a text-

type. “Cotext” refers to such factors. “Context”, in the philosophical, linguistic and anthropological

literature, in contrast, has come to be used as a synonym of what is ordinarily called a

“circumstance.”62 The idea of text as TTS construes it—structured documents with beginnings, ends,

and authors, and types—is completely foreign to context as it is discussed in the recent literature.

Here, what is thought to be salient under the heading of “context” are factors such as time, place,

speakers and interlocutors themselves, their interests and attitudes (including such controversial

factors as referential intentions), and most importantly, language.

The distinction between context and cotext is crucial. Contextual considerations pertain to

pragmatic considerations. For instance, J.L. Austin’s How to Do Things with Words provides an

excellent account of pragmatic considerations that bear upon language use (2000 [1962]). He sets out

a picture of pragmatic force that explains, for instance, how the recitation of vows can marry people

in certain contexts. Texts, in contrast, are inert, and their significance does not depend upon their

reader, but rather their institutional text-type, that specifies how a text is to be translated, but also read

and understood. If this were not so, anyone could pick up a text of higher mathematics and understand

it perfectly, even if they know nothing significant about mathematics. The clearest difference between

contextual and cotextual considerations is that the felicitous recitation of a text of a marriage

ceremony can marry people, but the text itself, does nothing. Indeed, the translation of a text outlining

a marriage ceremony does not marry anyone, the reading of such texts in foreign language does not

necessarily marry anyone, nor does the placement of such a text on my desk. There can be textual

events that are like speech acts: for instance, the signing of laws by the head of state, or the signing of

wills. These events are no different than speech acts, except they use the written word as their

medium. Translation, in contrast, must treat a text as an inert unit so as to extract the appropriate

99
features into a TT. The concern of translation thus is to understand how features of a text transcend

context by virtue of their preservation in a TT. This is very different from the focus of context and

pragmatics, where the interest is in understanding the significance of, say, a speech act, or a signing

of a law, and how this has a certain practical power in light of felicitous contextual conditions.

When we are not able to come to interpretive decisions on these more impersonal standards

alone, we can refer to authorial intentions, but this is only possible once we have evidence for treating

a text as a species of a certain type. In the absence of understanding that a text has a type, speculation

about authorial intentions is as indeterminate as the text we are trying to understand. Text-types thus

help us disambiguate texts, as well as authorial intention for texts, unlike language, have authors.

Text-types vary, and some may demand a certain degree of functional translation. And we may know

this quite apart from questions of authorial intentions.

Literary texts might appear to be the most obvious candidates for functionalist (dynamic)

translation. Many literary works, such as poems and novels, lend themselves to a functionalist

conception of translation, because aesthetic effect is central to their purpose. But even here, there are

many text-types that fall under the general heading of “literature” that strive to produce an effect on

the audience that cannot be fully translated dynamically. Principle among these are texts that have a

moral or prudential objective. For to substitute some alien value for those of the source culture’s

moral or prudential advise can scarcely be called translation of a moral or prudential text (for what is

critically relevant about such text-types is lost in this move). The Princeton philosopher of language

and African-American literary and cultural studies theorist, Kwame Anthony Appiah, offers an

example of such texts that comprise the literature of a people: proverbs from a dialect of the Twi-

langauge, spoken in a region of Ghana. The problem that functionalism attempts to overcome are

presented by the proverbs: how does one translate texts with culturally specific references that are not

familiar to the target culture? In contrast to the functionalist, Appiah recommends a strategy of “thick

translation” (using “thick” in the manner of Ryle and Geertz, and not Williams—see chapter 1). Thick

100
translation, on Appiah’s account, is a type of academic translation increasingly in vogue “that seeks

with its annotations and its accompanying glosses to locate the text in a rich cultural and linguistic

context” (Appiah 2004 [1993], 399). This is a stark departure from dynamic translation, and the

functionalism of authors such as Bassnett. No “imposition” of source text values upon the target

culture occurs via thick translation. But rightfully, the target audience does the work of trying to

understand the alien.

The cases of religious and literary translation suggest the following: that Functionalism

cannot be successfully applied to all cases of translation, and moreover it is not appropriate in cases

where, ceteris paribus, aesthetic effect is not a critically relevant feature of the ST. The underlying

problem for the notion of functional translation is that it is not a textual notion, but rather pertains to

how we are to treat sentences sub species aeternitatis, based upon intuitions and reflections on how a

sentence might be received. Such reflections can hardly be said to form the objective or principled

basis for translation.

II.6.5 The Limited Legitimacy of Functionalism

Do these considerations spell the end of functionalist strategies in translation altogether? If

so, this conclusion would be at odds with suggestions in the present chapter that functionalist

considerations can be critical to translating certain literary works such as novels, where literal

meaning ought to be sacrificed for functional equivalence to preserve a text-type feature through

translation.

The arguments presented in the present section do not provide grounds for removing

functionalist translation strategies from the tools of the translator, but they do imply that they have

been overused by translators and over-recommended by translation theorists. This is in part a result of

the extension of a translation strategy that works in certain provinces in translation to the project on

the whole (or, if one pleases, to translators straying from their areas of expertise to comment on what

101
translation in all cases must be like). The problem underlying the over-application of such techniques

in translation is the failure to properly address the question of how much work we ought to expect a

reader to put into a reading. I think many translators and translation theorists undervalue how much

work readers ought to put into reading translations in part because of the commercialization of

translation, which aims like all good business towards increased user-friendliness. Translation is too

important however to let the forces of the market decide its course.

I do think that functional strategies are legitimately employed in many cases. I have

suggested that in the context of some literary works, like novels, the relevant text-type features

demand that functional compromises in minor details of a translation be sought in order to preserve

the essential features of the ST in the same way. I suspect the same is true for poetry. Like all matters

in translation, it is the text-type that helps us decide such matters. And thus, the decision to be

concerned with the impact a translation has on readers is ultimately licensed by an objective

assessment of the features of a ST in light of a text-type. A concern for fidelity and accuracy in

translation thus underwrite such functional approaches.

102
III. Text-Types and Semantic Determinacy

III.1. Meaning and Translation

In the previous chapter I argued that theoretical problems with translation that have troubled

philosophers, both in the analytic and the continental traditions, a stem from a fundamental

misunderstanding of what translation is about. Philosophers such as Quine and Derrida have

conceptualized translation as a type of word for word, sentence for sentence, exchange across
63
languages based upon some latent conception of semantic equivalence. However, the reality is that

languages have their own natural histories and the notion that we can get anything like cross-

linguistic synonymy uncontroversially off the ground to underwrite translation is doubtful. The notion

that translation can be a word for word, sentence for sentence exchange across languages is dashed

for no greater a reason than that languages are characterized by distinct grammars, that make some

types of meaning rather obligatory (for instance, gender, number, and tense) and render others quite

difficult to realize when the native grammars do not recognize such syntactic categories. The

asymmetry of the semantics of words, and grammar as such, across languages has led to a chorus of

Quine, Derrida and translation theorists in the translations studies literature opining that equivalence

in translation is a unrealizable. But if TTs cannot be equivalent to their STs, then it is difficult to

understand how it is that there can be anything like a successful translation.

A fundamental mistake of such sceptical conclusions about translation is it assumes that there

is some type of absolute, sub species aeternitatis criterion that determines equivalence in translation.

This sub species aeternitatis criterion of equivalence in translation is assumed to be the criterion of

semantic equivalence of linguistic expressions across languages, or the notion of cross-linguistic

synonymy. However, criteria of equivalence in translation are not linguistic but textual and define

different types of texts. Equivalence in translation is possible not because semantic equivalence

103
across languages at the linguistic level (abstracted from all texts) is a fact, but because translation is a

selective process that aims to preserve some features of a text in translation. Thus, the fundamental

mistake of recent negative conclusions about translation stems from the assumption that translation

aims at treating languages as the objects to be translated. But translators have never translated

languages: this is a seemingly unlimitable task. Rather, translators have traditionally translated texts

of definite types.

Thus, Quine, Derrida, et al., start off on the wrong foot in translation (or, rather, they throw in

the towel before translation really begins). They think it is about language. In reality, it is about texts

of definite types. Texts of course need not be comprised of language as the linguist understands it. All

manner of semiotic systems can comprise texts, and we can translate texts across semiotic systems (or

more properly, we can create TTs that are semantically equivalent to their ST though they are each

comprise of distinct semiotic resources) because in translation we do not attempt to preserve

everything, just some text-type-theoretic concerns. That philosophers should think about translation in

terms of language is understandable, given the shared linguistic underpinnings of the recent

philosophical tradition that makes meaning in its most determinate form a feature of, or elucidated by,

language. That great philosophers who have thought seriously about translation arrive at sceptical

conclusions about translation has not generally been construed as evidence that we should rethink the

linguistic orientation of recent philosophy. I for one am for the notion that thinking should be

responsive to the evidence. The inevitable conclusion that translation is indeterminate against the

backdrop of the linguistic paradigm is the best possible evidence that the linguistic paradigm should

be abandoned. Translation can be nothing but indeterminate against the back drop of the linguistic

paradigm for the notion of a text-type is not a category of language or grammar. It pertains to the

organized semiotic whole that makes up a particular aspect of a text. The shared history of humanity

is a history of translation and if the linguistic paradigm cannot explain how it is that we have

104
succeeded in this basic activity of intercultural communication, it shows itself to be disconnected

from reality.64

In the last chapter I made a case for LET and argued that translation aims to preserve some

features of a ST in translation, and that translation is determinate when it preserves the right features

of a ST in a TT relative to text-type-theoretic considerations. If there is a residual controversy over

translation I argued that scholarly activity can always sort out such problems, either through further

research, or translational devices such as text-moulding. The reason that expert activity can solve

problems of translation is two-part. First, since a text-type institution is individuated by its text-type,

scholars who are expert in a text-type and the institutional knowledge underwritten by such text-types

are uniquely placed to deal with controversies of translation. Secondly, given that translation is a

textual endeavour, and not a linguistic endeavour, it is always possible to create a TT that preserves

the features of a ST that are important relative to a text-type for we simply introduce novel

expressions when a target resource is insufficient or use old devices in new ways. Text moulding, an

orthodox approach to solving problems of translation, is always possible particularly because

translations need not be readily intelligible to their readers. Indeed, specialized texts of various

disciplines are not readily intelligible to readers in a source language community and there is no good

reason why we should think that all texts in translation, such as, for instance, a tract on theoretical

physics or mathematics, should be readily intelligible to anyone without the proper institutional

training. What is important is that our translations be in principle accessible by the target community,

and the existence of experts trained in the relevant text-type and the TT are all we need in order to

ensure that novel ideas are accessible to people who left to their own devices would not be able to

understand such a text. But text moulding is always possible in translation because translation is not

an effort to match up linguistic devices across languages given some prior account of their meaning.

Rather, translation aims at recreating a TT that mirrors the ST in specific features. LET has no

commitments to the inviolable nature of linguistic devices that it must respect in translation. We use

105
language or any semiotic system, to construct texts as we please, and so long as they conform to

institutional expectations, the project is successful.

I have deliberately made use of the notion of a textual “feature” in the last chapter and in my

discussion of LET, and I have avoided talking about meaning. But what indeed are these “features” of

texts that are to be preserved in translation? They are, I submit, nothing but semantic features of a ST.

In other words, the features that a text-type tracks in translation are semantic features of a text and it

is these features that a TT must embody. Thus, STs and TTs on this account are semantically

equivalent to each other. What makes this account different from the linguistic account of translation

is that the meaning that is preserved in translation, on this account, is not explainable by reference to

the semantic profiles of the constituent devices in a text sub species aeternitatis, but rather by

reference to text-type-theoretic concerns that organize and highlight aspects of the polysemy of

semiotic devices in a ST: it is the combination of these two factors that constitute a text of a definite

type, and it is this synergy that we recreate in translation, albeit with distinct semiotic devices. But the

principle remains the same. In the TT we employ distinct semiotic devices, all polysemous in the

abstract, and we select them for incorporation in a text based upon their ability to mirror the semantic

features of the ST in light of the shared text-type. Viewed apart from texts, semiotic devices each

have a one-to-many relationship with other semiotic devices, within and across semiotic systems such

as languages. But when in actual texts, they are constrained in such a manner by their fellow devices

that they come to have a narrow range of significance, and this narrow range can be semantically

equivalent to corresponding translation units in a ST when we produce translations according to text-

type. It may still be possible that the cotextual conditions of a device are such that they, and many

others, have within a text, multiple significances that support more than one text-type equally in

which case such texts can be translated according to more than one text-type. Moreover, it is possible

for texts to be ruled by a certain type, but yet for this type to constrain subsidiary portions of a text, in

light of its inherent polysemy, in such a manner that we must understand this subsidiary portion as

106
instantiating a distinct text-type. Thus, for instance, an article in the philosophy of science, while a

text in philosophy, may, owing to a combination of this text-type and its components, indicate that

some subsidiary portion constitutes a discourse that is best treated as an instance of the text-type of

science. None of these considerations are reducible to the individual components of a text in the

abstract, but come to light in combination.

The task of creating a translation is thus very much like recreating a structure, such as a

sculpture, but with distinct materials. In the abstract, a sculpture of a person made of macaroni and

another made with tongue depressors may look the same from a distance: they both anatomically

mirror each other (they will have the same shape arms, feet, head etc.,) but viewed more closely, the

individual pieces of macaroni and tongue depressors will not be exactly equivalent in shape and size.

In some cases, we need to combine several pieces of macaroni together and several tongue depressors

to arrive at the same shape, but so long as this shape corresponds to a crucial component of the

sculpture, identified by a plan, we take these as equivalent given their role in the sculpture, though not

sub species aeternitatis. Our semiotic devices are our bits of macaroni and tongue depressors and our

text-type is the general rule that allows us to extract from the original sculpture made of macaroni, the

shape and pose to be recreated in the replica, made with tongue depressors. The replication of the

sculpture with distinct materials is really what translation is. Philosophers in the linguistic paradigm,

in contrast, have thought that translation is simply a matter of comparing pieces of macaroni with

tongue depressors without concern for what sculpture they are comprising. Disheartened by the fact

that pieces of macaroni and tongue depressors do not completely resemble each other, philosophers

such as Quine and Derrida have drawn negative conclusions about translation. Others have simply

ignored the problem, ethnocentrically assuming that all sculptures must be made of macaroni, or

objects that are virtually identical to their own bits of macaroni that comprise their own sculpture,

thus wilfully ignoring the fact that there are many lands where tongue depressors, and not pieces of

107
macaroni, are what sculptures are made of. Most seem to have ignored the possibility that pieces of

macaroni and tongue depressors can be bent and moulded to make up the difference.

Text-types thus specify constellations of meaning and it is these that are preserved in

translation just as the various parts of the body of an original sculpture are preserved in the recreation

of the sculpture. Meaning, on my account, is transferred in translation, but this transference cannot

occur without the mediating and filtering role of a text-type, which highlights some semantic

modalities of a ST and recreates these with novel devices in a TT.

An important advantage of thinking about text-types as preservative of semantic content is

that it allows us to stick by the orthodox account of translation as a process driven by the goal of

semantic content preservation. There are several types of semiotic conversions. Adaptation,

paraphrase, interpretation are all ways of transforming symbols, but what distinguishes translation—

determinate translation that is—is that it and it alone is semantically equivalent to its original. When

translation is not determinate, we have no reason to be fully confident that putative parings of

semiotic phenomena are semantically equivalent.

Determinate translations yield what I shall call determinate meanings of three sorts: (I) whole

texts are determinate meanings, or if one prefers, determinately meaningful, when they are

determinately translatable according to an institutionally recognized text-type: STs and TTs that are

genuine translations on this account have the same meaning. Thus, for instance, Wittgenstein’s

Philosophische Untersuchungen and Anscombe’s translation, Philosophical Investigations, are each

semantically determinate and moreover semantically equivalent. (II) Text-type features that must be

preserved in a determinate translation are determinate meanings. Thus, for instance, if certain

concepts are essential to the articulation of a text of a certain type, and the text-type demands that

they be preserved in translation, their canonical specification by the relevant institution constitute

determinate meaning. The conceptual content of key philosophical terms as set out by QI are such

determinate meanings. (III) Translation units that are preserved in determinate translation as per TTS

108
are determinate meanings and semantically equivalent to the ST counterparts. (Translation units so

understood are the smallest units of a text to be preserved in translation: as atoms of translations, they

will be semantically equivalent in the ST and TT, though their constituents and syntax considered in

the abstract will likely not be equivalent.) Such units can be as small as words, and as large as

passages of texts, and there is no text-type independent reason to believe that such units must be equal

in size. In each case, they are identified and treated as semantic wholes in translation by the text-type

and the semantics of the constituents of a text. Thus, the following section of Wittgenstein’s

Philosophical Investigation in German and English are semantically equivalent and determinate as

units of this text (and not in the abstract):

I.134 Betrachten wir den Satz: “Es verhält sich so und so”-wie kann ich sagen, dies sei die allgemeine
Form des Satzes?-Es ist vor allem selbst ein Satz, ein deutscher Satz, denn es hat Subjekt und
Prädikat. Wie aber wird dieser Satz angewendet-in unsrer alltäglichen Sprache nämlich? Denn nur
daher habe ich ihn ja genommen.

I.134 Let us examine the proposition: “This is how things are.” How can I say that this is the general
form of propositions?-It is first and foremost itself a proposition, an English sentence, for it has a
subject and a predicate. But how is this sentence applied-that is, in our everyday language? For I got
it from there and nowhere else.

Since translation is only determinate at the textual level when text-types are brought to bear,

it follows that linguistic meaning (and any semiotic meaning that is not itself determinately

translatable according to a text-type) is semantically indeterminate. I call this theory of determinate

meaning Text-Type Semantics, or TTS. TTS is not a theory of all meaning: it does not specify what it

is for something to be meaningful. (Indeed, I think this is probably impossible except at the most

vacuous level: everything, both real and unreal, is meaningful!) It does however specify the

conditions for semiotic phenomena to be determinately meaningful.

Indeterminate meaning, like determinate meaning, is preserved in translation. However,

indeterminate meaning is preserved in a process of translation that falls outside the constraints of a

109
text-type. I call such translation outside of institutionally sanctioned text-type translation schemas. I

would hasten to add that we would do well to also call such translations pseudo-translations, as they

lack the features of a real translation (i.e., they are not translations of texts but text fragments

abstractly conceived, apart from text-types). Translation schemas consist of, among other things, a

definiendum with its definiens. The definiens is a type of translation of the definiendum. Other

examples of translation schemas, or indeterminate meaning, are analyses of ordinary language

concepts provided by philosophers. Other translation schemas may consist in the paring of words with

objects they are understood as denoting (so called ostensive definitions), and others yet might consist

of lengthy, lexicographical explanations of the meaning of a word or expression, which at best

constitute an indeterminate abstraction of the role the definiendum plays in the various text-types of a

culture. Indeed, the more authoritative a translation schema is, the more generally it conforms to texts

of a culture. The main reason that translation schemas yield indeterminate meaning is that they are

incapable of authoritatively informing us how to translate the language that they treat. Translation is

only determinately possible by taking into account the text-type of a text that will guide its

translation: translation schemas say nothing about text-types.

The notion that language is semantically indeterminate is not simply a negative or critical

thesis. On TTS’s lights, it is a recognition that semiotic phenomena that are not texts of definite types

are polysemous and full of semantic possibilities, and that their actual semantic profiles are a result of

their history. Because symbols and words in the abstract are polysemous, we can construct TTs with

distinct devices that are semantically equivalent to STs for such texts are constructed in light of text-

type-theoretic considerations that choose and employ expressions for their ability to mirror some text-

type-identified features of a ST even though the devices themselves are not cross-linguistically

synonymous when abstracted from all cotextual considerations. The immense semantic indeterminacy

of language as a class of semiotic systems is its strength for it implies that language is extremely

versatile in recreating texts of all manner. There is nothing to stop of us from translating a book into a

110
mural, for instance, but to render this mural intelligible as semantically equivalent to a book will take

a huge amount of work that renders it simply not worth it. Experts in the mural would have to

understand how the various components of the mural are translatable into other semiotic systems,

such as language, and would have to be able to decode the text thus for students. It is typically not

worth the trouble to translate Aristotle’s Ethics into a panel of paintings. Philosophers in the Indian

tradition, before writing took off in a big way, translated their philosophies into s tra texts, which are,

for all intents and purposes, verbal paintings that are to be decoded. A line would be comprised of a

few words, each with multiple significances, and appreciating the text-type of philosophy allows us to

organize these significances in a manner that renders them translatable into discursive texts. The

advantage to this mode of comprising philosophical texts was that it economized on length and

allowed for the easy memorization of philosophical texts: words were chosen not on grammatical

grounds but purely for their polysemy, and they were strung together in lines that often had no

grammatical structure. These texts fail to be grammatical on some points, at times not distinguishing

between subject and predicate. They are rather very much like the proposed panel paintings of

Aristotle’s Ethics. We typically do not bother with such labours because language is easy for us, and

it is easy because of its versatility, but with such versatility comes indeterminacy for in the abstract

there is no clear answer to how such versatile devices are to be preserved in translation.

We are led to this conclusion because definitions or even intersemiotic translations of

semiotic devices, such as words, sentences and symbols, do not determine how they are to be

preserved in translation. The missing piece is the text of a definite type: it is only within that

environment that the question of how to translate an expression has a tractable and determinate

answer given text-type-theoretic concerns. To think that meaning is a determinate feature of an

expression is like thinking that a tongue depressor contains within it the information to determine how

it is to be employed in recreating a sculpture, without needing to know what sculpture is being

recreated or what portion of the sculpture the tongue depressor is a candidate component of. Frege is

111
famous for what is often called the context principle: a word only has its meaning within the context

of a proposition (Frege 1980 [1884], x). In its place, TTS offers the following cotext principle: a

semiotic device only has a definite meaning in a text of a definite type.

TTS provides us everything we need in semantics, I think, but it involves giving up the

assumption that meaning is fundamentally linguistic or elucidated by linguistic categories such as

words or sentences. But we needn’t defer to a formalization of TTS to recognize that there is a

problem for the traditional linguistic approach to meaning. Linguistic meaning does not determine

translation, but rather it is the text-type that provides the determinative criteria in translation. If

linguistic meaning has no determinative role in showing how meaning can transcend culture and

language in translation, there is very little reason to think that it warrants the privilege that it has been

granted by philosophers.

In the remainder of this section I shall answer some worries about recognizing the semantic

indeterminacy of language. If my responses to these worries are successful, I will have shown that the

threat to the objectivity of knowledge and meaning is not the indeterminacy of linguistic meaning, but

rather the inverse, namely the possibility that language is determinately meaningful.

III.1.1 Conversation and TTS

One might think that semantic determinacy in language is necessary in order for conversation

and social interaction. TTS may thus seem to be false on this score. However, many have learnt from

Wittgenstein’s Philosophical Investigations that this cannot be correct, for conversational success is

consistent with speakers employing words according to diverse interpretations and rules. So long as in

conversation we employ expressions in a manner that comports with some expectations our

interlocutors have about how our expressions are to be used, we are able to have successful

conversations without sharing determinate meanings.65 Some philosophers have resisted this notion

and have insisted that conversation is a type of communication where the meanings in one person’s

112
mind are somehow transferred via speech acts, through the senses of the interlocutor and into their

mind. This is the Lockean picture of thought and communication known as “telementation.”66 This is

a very naïve view of communication and it is difficult to believe for it ignores the role that diversity

(ethnic, cultural and philosophical) plays in underwriting what people say. Conversation is shared by

interlocutors, but not all of their presuppositions. Many interesting types of conversation occur

exactly because interlocutors operate with differing presuppositions, and even different linguistic

meaning: philosophical conversation is like this. The notion that there is one shared or underlying

linguistic meaning for important philosophical terms, such as “good,” “right,” “beauty,” “real,” or

“knowledge” is belied by the diversity of linguistic accounts of their meaning and varied cultural

histories of peoples within a society and culture: there are often several competing accounts based on

the linguistic evidence, which are not wholly consistent with each other. Each has some linguistic

evidence in support of it, but none is uncontroversial or synoptic. The authorities that persons defer to

in a society to cash out their own ethical theory often diverge radically and the notion that there will

be one underlying literal meaning for philosophical concepts is farfetched indeed. (The notion, for

instance, that a Marxist and a Catholic within one and the same society operate with the same

linguistic meanings in their use of moral terminology, and that they defer to the same authorities to

fill in the details of their moral vocabulary usage, is farfetched, as is farfetched the notion that they

will converge significantly with each other and others on their moral convictions.) Moreover, the

notion that people are simply disagreeing philosophically and that they are talking past each other is

not clear in the philosophical context for disagreements on philosophical matters have a

metalinguistic character to them to the extent that interlocutors desire that others conform their

philosophical language use to their own substantive criteria. (Most language use has this

metalinguistic character, but this is most pronounced in the case of philosophy.)

It was questioned in the previous chapter under the discussion of Davidson’s theory of radical

interpretation whether text-types have any applicability there, and the answer provided there was that

113
they do. We can bring such formats into contexts where we are trying to understand language from

scratch so as to provide a textual frame to interpret speech acts. It is not an ideal condition. Really,

text-types apply to texts. In the context of the functioning of text-type institutions, officers of such

institutions can with great productivity apply the translational principles of the underlying

institutional text-type to interpret each other’s utterances. In this context such an application will be

very felicitous. For instance, in a conference of philosophers, making use of the text-type of

philosophy to understand a discussion or argument between critics will clearly be one of the best

ways to make sense of conversation. But in the radical context it is difficult to see how felicitous such

a strategy would be, but at the same time we have little choice but to try out everything we have in

order to make sense of utterances.

What then of rather domestic conversations those are neither radical nor institutional, such as

a conversation with a cashier at the supermarket: do we require a text-type here? I doubt it. The

interaction is purely pragmatic: the customer declares products; the cashier declares a price, the

customer and cashier settle accounts and if the cashier is satisfied the customer can keep the products

without objection from the cashier. But perhaps the worry is different. Perhaps the worry is that there

is no way to determinately characterize the interaction because it has no semantic determinacy. I

don’t think this follows. We can, after all, abstract from such an interaction a transcript and the

transcript can be the subject of social scientific research. If translating conversations at the grocery

store with a cashier becomes a very important matter for society on the whole, we will no doubt

strengthen our institution of treating transcripts of such conversations as texts of a specific type.

Some scholars have given the social practice of conversation a lot of thought, and

Ethnomethodologists (or “Conversation Analysts”, according to an alternative nomenclature) have

noticed that many ordinary conversations conform to certain patterns (Sacks and Jefferson 1995).

This research may constitute the beginnings of an institution devoted to translating conversations as

their own text-type.

114
 III.1.2 Science, Objectivity and TTS

Perhaps one might think that we render science unworkable if language is semantically

indeterminate. Consider for instance the now famous arguments of Saul Kripke and Hilary Putnam

for the “causal theory of reference” (Putnam 1975; Kripke 1980 [1972]). According to this argument,

meaning, particularly as it is relevant to science, is what a word, typically a natural kind term, refers

to, and the reference of this term is fixed by the history of the term. Particularly on Kripke’s account,

there is a baptismal context where a word comes to name some natural item, and it is subsequently

passed along, through a social chain that constitutes the causal mechanisms of the term’s

dissemination in a language. Speakers come to associate all manner of descriptions with this term,

and the descriptions can lead speakers to identify some referent of a term that is not what it is causally

tied to. The Fregean sense, or the cognitive representation of meaning, comes apart from reference, as

the historical baptism of the term comes apart from the descriptions speakers associate with such a

term. While descriptive content associated with natural kind terms provide some pragmatic facility to

language users, it is distinct from the actual meaning the term as can be guaranteed by science.

Science and empirical investigation on the whole can investigate the provenance of the term and

provide us an account of the essence of the item being referred to, shedding light on the causal

mechanism underwriting its usage in a community. On Putnam’s account, sociolinguistics will shed

light on the “linguistic division of labour” that apportions the disambiguation of the reference of

terms and the proper extension of their respective concepts by way of relevant experts in a linguistic

community, while psycholinguistics will shed light on what competent speakers of a language must

know in order to use terms intelligibly within a community (Putnam 1975, 144-166). If linguistic

meaning were indeterminate, this causal account would not function well, for it assumes that there is

something like a determinate, uncontroversial fact about reference that science uncovers.

115
 Joseph LaPorte in his excellent Natural Kinds and Conceptual Change (2004) characterizes

the theory and its supposed implications for scientific objectivity thus:

According to the causal theory, we do not define natural-kind terms by theoretical descriptions,
which would be expected to change over time with our theories; otherwise the meanings of our kind
terms would change with our theories, as incommensurability theorists say that they do. Rather, we
identify samples of a kind, apply a name to the kind they exemplify, and then discover the essence
of the kind. Theories can come and go without the reference of our natural-kind terms changing at
all, for it is the samples in the world, and not our theory, that determine what our terms refer
to. Thus, given the causal theory of reference, worries about incommensurability do not arise
because worries about meaning instability do not arise. (LaPorte 2004, 3)

According to LaPorte’s careful investigation of the history of cases from biology and

chemistry, Kripke’s and Putnam’s views are interesting speculations, but they do not conform to the

reality of scientific progress. LaPorte shows that scientists do not feel saddled with either the

historical baptism of terms, or with descriptions, and make adjustments to their definitions of terms in

order to make their claims true. “Scientists do come to new conclusions about the essences of

traditionally recognized kinds frequently enough, but their conclusions reflect a substantial measure

of fiat, following a strain to shape the use of old terms in the light of new findings” (LaPorte 2004,

65). As just one example LaPorte relates, consider the question of whether guinea pigs are rodents or

not. Apparently, scientists think they are not, but LaPorte argues that this was not forced upon them

by discovery. Scientists could have maintained that guinea pigs are rodents, if they were willing to

make adjustments elsewhere in scientific theory, for the very notion of a rodent reflects no exclusive

Glade. Indeed, on the strength of the evidence, scientists could have decided that we are all rodents,

or perhaps even that only mice are rodents (LaPorte 2004, 66). In other words, LaPorte argues that

while the causal theory has some rough validity, it cannot secure linguistic stability and as a result it

cannot explain the objectivity of science in the manner that it proposes. This leaves open the threat of

incommensurability (cf. LaPorte 2004, 113).

116
 Incommensurability of scientific theories is the worry opened by Kuhn’s analysis of scientific

knowledge, according to which the meaning and even the truth values of scientific claims are relative

to theories (Kuhn 1970, 206) that are constituents of what Kuhn calls “paradigms.” Kuhn says many

things about paradigms, but he treats them as a shared cultural resource, that give rise to shared

standards of scientific research (Kuhn 1970, 42). The research standards and theories that paradigms

hold are incommensurable with other such paradigms that they compete with (Kuhn 1970, 150).

Communication between paradigms must be thought of as a special case of the effort of members of

differing linguistic communities to try to translate each other’s claims (Kuhn 1970, 175, 202). Kuhn

deferred to Quine’s account of translation (Kuhn 1970, 202, fn.17). But this can only imply that

translation between scientific paradigms is indeterminate. He later sharpened his view and concluded

that incommensurability implies untranslatability: “The claim that two theories are incommensurable

is then the claim that there is no language, neutral or otherwise, into which both theories, conceived

as sets of sentences, can be translated without residue or loss” (Kuhn 1983, 670).

LaPorte dismisses such worries in a few pages. He reasons that difficulties in translation do

not stop people from learning about each others world view. If nothing else one can gain such an

understanding by learning some of the language of one’s interlocutor. The real issue on his account is

not that terms across languages should be intertranslatable but that one should be able to find suitable

sentential correlates across languages, and this can be accomplished, he reasons, even if individual

words cannot be translated across languages (LaPorte 2004, 145-6). He writes:

Translation, in the strict sense Kuhn understands it, is unnecessary for reasoning about statements
made in two languages, and so is the use of older terms. It is enough to mention the older terms and
talk about their conditions of application. This would not be to translate… but it may be used instead
of translation to convey an idea.(LaPorte 2004, 146)

LaPorte is correct to think that Kuhn’s account of translation is unnecessary for reasoning

about claims of scientific theories. But LaPorte has not shown why this is. Worse, he does not dispel

117
the original worry that Kuhn brings to light, namely the criteria of assessing the success of scientific

theories are internal to each theory. While we must recognize LaPorte for his invaluable service in

showing how the causal theory of reference that is supposed to underwrite the objectivity of science

does not underwrite its objectivity, there is still heavy lifting to do to overcome the general problems

of relativism that come in train with thinking of meaning as something language relative.

There is also a deeper practical problem that Kuhn overlooks, and that LaPorte’s easy dance

around the topic of translation avoids as well. If we cannot translate scientific theories, how do we

know when we’ve come across one?

Kuhn’s relativism, like all relativisms that I know of, is provided without any serious effort to

overcome the indeterminacy of and obstacles to translation. In Kuhn’s case, he argues that the

translation of science is impossible. Oddly, sceptical theses about translation are not uncommon

among relativists (cf. Harman 1969). But if we cannot translate scientific texts, there is no reason to

believe that we can determinately identify science across cultures and languages, and distinguish it

from any other cultural activity. But if there is no way to determinately identify science, or its texts,

across cultures and languages, we cannot even begin to make a case for the thesis of relativism,

namely that there is no common arbitrator among scientific theories. Moreover, in the domestic

context, there is no reason for us to think that we can determinately identify science and its texts

either if we cannot determinately translate scientific texts. Certainly, we can find activities and texts

labelled with the term “science” but words have many meanings, and unless we have an objective

manner to identify such texts and their resultant practices in the relevant sense of “science”, there is

nothing disciplined at all about a historical study of scientific progress. In the absence of such an

account of the translation of scientific texts, what Kuhn does is rely upon his intuitions to tell him

what counts as science and what does not. From this intuitive perspective it may seem as though there

is no common arbitrator between scientific theories or practices, but this should not be surprising for

118
Kuhn has not made explicit use of a manner of determinately identifying scientific texts. But for the

same reason, his analysis can hardly be said to be conclusive.

If we can determinately translate scientific texts that articulate scientific theories, the notion

that there is no common platform from which to judge competing theories is undermotivated, for

whatever allows us to determinately identify and translate scientific theories would also provide us an

Archimedean platform to judge scientific theories from. In so determinately identifying scientific

theories and practices by reference to a determinate account of the semantics of scientific texts, we

would be liberating our analysis of science from a perspective that rests solely upon intuitions to tell

us what counts as science and what does not and we would thereby have a platform for the objective

criticism of science.

We know enough about translation by now, in light of the previous chapter and the present

discussion, to know that Kuhn, just like Quine before him, simply got translation wrong. Translation

is not primarily an attempt to match up words and sentences across languages. The entire view that

scientific theories are incommensurable is based upon this spurious, linguistic conception of

translation. Had philosophers had the right account of translation in view, this entire literature may

never have been spawned. For, according to the right account of translation—TTS—all content comes

apart from systems of representation for they are, in themselves, semantically indeterminate. It is only

within a text of a definite type that a scientific theory has the content that it has, and it is by virtue of

the text-type of science and its institutional underpinnings that such determinate content can be

translated into any language. Thus, there is always a language-independent way for a scientific

theory to be articulated in the sense that a scientific theory is not trapped or constituted by a particular

language.

If the target language should fail to have resources that we require for translation, we simply

use old devices in new ways (making use of the devices of text moulding) or we introduce the

necessary technical terminology. So long as the experts in an institution can understand such

119
deviations from provincial term usage, the process is successful. None of this compromises

translation because translation is not an effort to match up words across languages on the basis of

some prior, determinate, linguistic meaning. If, thus, a certain scientific theory is prevalent in a

culture influencing, to a great extent, its use of natural kind terms, we can always use these same

terms differently, in deference to an alien scientific theory, in constructing a TT, in order to

determinately translate the alien text of science. If, however language had determinate meanings that

it brought into texts, this would not be possible.

All types of discourse are characterized by their own text-type, and science is no exception.

The text-type of science, I submit, is concerned with the articulation of descriptive theories in light of

empirical considerations, broadly conceived, that aim at being objectively persuasive. If the translator

of science did not understand that scientific texts aimed at putting forward considerations that were

thought to be objectively in support of their scientific theories, it would be possible to translate a

scientific text as though it were a type of anthropological study of conceptual schemes or perhaps

even a type of literary exploration of images, or a type of phenomenology. But this would be to lose

much in the way of important content in the translation of science. Science typically aims at more

than systemiticity. It aims at being right as far as its descriptive assertions go. Thus, not only must the

translator of science understand the descriptive theory or theories that are articulated in a text of

science, she must also understand that such texts point to, or identify, reasons that could objectively

vindicate such a theory. Understanding the text in this light allows the translator to save the right

content in the translation of a scientific text.

Given this understanding of the very semantics of the scientific text, we can use this criterion

as a means of adjudicating between competing scientific theories. We can thus reject theories that fail

in some respect to live up to the very semantic mandate of science, to provide descriptive theories that

in light of empirical considerations are objectively persuasive. And it so happens that the implications

of this translational criterion match up with widespread institutional expectations in science.

120
 The first rung of empirical considerations that speaks to the objectivity of the descriptive

theory in a scientific text is some type of empirical support in the way of predictive success. If a

descriptive theory does not have any such empirical implications, it is difficult to understand how it

can even be a descriptive theory. For theories abstract from many cases, and thus in putting forward a

certain picture of the world as a description, a scientific theory characterizes times that are yet to

come. This characterization of future cases is simply what is identified as prediction or its testable

implications. Predictive success is thus evidence of descriptive success. Predictive success in the face

of explicit testing is always great. At other times, practical success in the employment of a descriptive

theory can be taken to be evidence of predictive success—or at least, the translator of science must be

aware of this possibility. In the case of traditional forms of medicine, for instance, what one will

likely find is nothing so explicit. Rather, what is usually passed down in such traditions of medicine,

such as Ayurveda, or Traditional Chinese Medicine, is a descriptive theory of the body, and a

normative theory of clinical practice. These normative theories of clinical practice can be regarded in

the eyes of traditional practitioners as a vindication of the descriptions they operate against if

employed over a long time with some therapeutic success. Understanding this is essential to

translating such texts as texts of science.

The next step in assessing the success of a scientific theory, once it has shown itself to have

some predictive success, is its ability to be objectively vindicated. What can objective vindication be?

It can only be vindication independent of the theory itself. In other words, if a scientific theory can be

shown to gain support from scientific theories in other areas of scientific research, then one has a

form of objective vindication—what is known as the “'consilience of inductions” (cf. Ruse 1987,

238). Thus, for instance, if a theory in biology about the progress of evolution can be found to

conform with what is taken to be settled fact in geology, then the biological theory has a type of

objective support. Of course, findings in biology might cause geologists to reconsider some timelines

if any of them depend upon biological accounts of fossil records or both biologists and geologists

121
may revise their views to come to an account that is mutually satisfactory. As LaPorte notes, at many

times revisions in scientific theory may not be mandated by the linguistic facts as such. They may

represent a substantial amount of linguistic fiat. But either way, the ability of scientific theories to

gain support from other scientific theories is an important way in which scientific theories defend and

vindicate their objectivity for in gaining such support, the theory has shown that its pronouncements

should not be taken seriously simply on its own word.

If in the long term several competing theories of science are equally vindicated, scientists can

attempt to systematize the area where these competing theories, and their implications, converge and

overlap. At the point of scientific impasse, the objectivity of a scientific theory is simply what

support it can find in its competitors. Thus, the overlap that can be obtained from this process of

dialectical and mutual understanding between theoretical competitors provides a foundation for

science to move forward. This may seem impossible at first glance for the manner in which each

theory conceptualizes its implications will be theory relative and thus there will not even be a

common stock of observational claims that theorists can agree upon. But this of course assumes a

certain unwillingness of scientists to move forward and find common ways of systematizing their

overlap, which in turn assumes a semantic priority of the various competing theories. There is no a

priori reason that we should think that scientists must be so trapped by their theories. There are

always wider institutional considerations based upon the text-type of science that can provide

common ground for theorists with competing theories to stand on in order to synthesize their findings.

This is one of the benefits of not viewing language as the seat of semantic determinacy: we can revise

the meaning we attach to terms in institutional contexts if we have reason to do so, and institutional

impasses are an example of one such good reason to do so. Thus, the type of impasse that Kuhn often

describes in his work between purveyors of rival scientific theories is not necessary and may

represent a breakdown of scientific rationality.

122
 If Kuhn were really correct, it would be difficult to say whether the Darwinian account of the

origins of life or Creationism is correct: each is its own theory, which makes its own claims, and it

seems there is no common arbitrator: we cannot translate Darwin’s concepts into that of the Book of

Genesis, and the Book of Genesis cannot be translated into The Origin of the Species. Orthodox

scientists and so-called “Christian Scientists” could arguably constitute distinct paradigms, with their

own culture and technical terminology. Educated people tend to think that Darwin provides the

correct account, and the Book of Genesis does not. But what underwrites our deference to scientific

orthodoxy at this point?

“Humans evolved from apes.” Typically, the response one gets from Creationists to this claim

is something like, “How do you know? You weren’t there!” Indeed, none of us were around to see

this process unfolding and there are many claims of science that educated people take to be true but

people could never observe (particularly historical claims of science). The linguistic turn in its

emphasis on language leads us in the first instance to take very seriously the notion that science is

about an individual proposition. Thus, criteria like falsifiability, and testability, come to the fore as

essential tests of individual propositions. However, to test an individual proposition one must assume

many other propositions as well. Thus, philosophers of science since Carnap have thought that the

notion that claims have their meaning in isolation (even when they are empirical claims) is simply too

naïve. What is at stake is the scientific theory. But if the scientific theory is what renders scientific

claims meaningful, how do we adjudicate between competing scientific claims? It is difficult to see

what there is if meaning just comes down to theories and propositions. And thus the claim among

some Creationists that Darwinism is just as much an article of faith as Creationism has some standing

for, at least against the backdrop of the linguistic turn, it is unclear how we decide between the two.67

The Kripke-Putnam account may have seemed to provide a way out of this conundrum, in so

far as it sought to ground the meaningfulness of language as such (and thus science as well) on

scientific facts about language that science would later uncover. But this seems to be the wrong

123
account of scientific development and linguistic meaning. Scientific development is answerable to

scientific fact, not language as such. But we shouldn’t be surprised by this. Science at times

vindicates language, and at other times it does not. If its job was to solely vindicate language, every

racist and culturally contingent mode of labelling the world would make its way into scientific theory.

But then there would hardly be any need for science.

TTS can accept that science alters linguistic meaning and creates concepts to do its bidding

that are not necessarily restricted by the linguistic facts about a language, thus recognizing the

autonomy of scientific considerations in the development of science, but it can help us also out of the

conclusion that scientific theories are each autonomous semantic islands.

For TTS, Darwinism and Creationism are not competing theories hanging in the air: to the

extent that they are competing theories, they are competing theories because they can be expressed in

texts that are translatable according to the text-type of science. Thus, they can both be judged against

this text-type, which leads us to judge descriptive theories not only in light of their predictive success,

but their ability to find support from other scientific theories (hence confirming their objectivity).

When placed in this light, it is obvious that Creationism is an abject failure. It has no obvious

predictive implications, while Darwin’s theory of natural selection does. (Darwinism implies, for

instance, that the makeup of organisms subjected to selective pressures over generations changes).

Darwinism moreover provides explanations for phenomena of interest to biological theory that

Creationism does not. (Darwin, for instance, has an explanation for why the bone structures of

vertebrates are similar while Creationism has none.) Most specific accounts of evolution provided by

biologists in light of known mechanisms of natural selection accord with timescales of the age of the

earth that are consistent with the best going accounts of the age of the universe. Creationism,

particularly if fixed by the Book of Genesis, does not. While all scientists and philosophers of science

can point to such facts in demonstrating how the Origin of the Species is superior to the Book of

Genesis, TTS shows how such considerations in assessing a scientific theory are semantically

124
mandated by virtue of the text-type that we would have to defer to in order to entertain the Book of

Genesis as a text of science. Thus, to the extent that “Christian Scientists” wish so-called “Christian

Science” to be taken seriously as science, their texts semantically sign up for the same treatment that

all scientific texts receive.

Here, I think, is a major strength of TTS. It shows us how a broad range of institutional

considerations can be reduced to a proper account of the semantics of a discipline such as science.

The broad range of institutional considerations comes about, in the case of science, from the fact that

it is a text-type that is concerned with the objectivity of descriptive theories. But objectivity in the

face of competing theories can only be objectivity generated with the support of other theories that

are also descriptive with testable implications. Kuhn’s picture of science thus as a battle of

semantically disconnected worlds makes no sense, not only because it fails to describe the semantic

impetus of science correctly, but it cannot even underwrite the identification of disparate paradigms

as paradigms of science. The only way to do this determinately is to determinately translate texts of

science, but the only way to do this is by a viable account of the text-type of science. With TTS in

view, the theses of translational incommensurability and the relativity of epistemic claims is shown to

be an incredible muddle. They, like all forms of relativism, assume to be able to identify the items that

are the putative points of relativity without also having a common point of reference that they can be

judged against.68

In closing this section, we can note how TTS safeguards the objectivity of institutional

knowledge, including scientific knowledge, in at least three respects. First, it distinguishes

institutionally vindicated claims, such as the best scientific convictions, from opinion at large and

thus opinion at large can be far behind the state of scholarship in a given institution. Institutional

epistemic claims are typically underwritten by specific text-types while ordinary opinions may not be.

Secondly, because the institution of an epistemic enquiry, such as science, is translational, based as it

is on a text-type, it is not itself a type of language game or unified culture. Rather, it transcends

125
particular languages and cultures, including paradigms in Kuhn’s sense. Successful epistemic

institutions will be patently cohesive and we may be tempted to call the institution a community. But

such a “community” will be quite atypical and formal in its constitution. Members of the institution of

science, or any text-type institution, can disagree individually on all manner of philosophical matters

and still have a common text-type-theoretic focus. Aside from such an institutional grounding, they

may share little else. Third, because knowledge must be underwritten by an institutional text-type, it

is possible on TTS’s account for members of a text-type institution to make errors, and indeed it is

possible for many or most in a few generations of such scholars to be incorrect. At some point, errors

in an institution change the very underlying norms of an institution and the error becomes the norm,

changing the underlying text-type that defines the institution. But for many generations, it is possible

for there to even be wide-spread error even within an epistemic institution.

The cost of such objectivity is that we deprivilege the role that particular languages play in

determining meaning, knowledge and translation. I take this to be a small price for such rewards.

III.1.3 Language and Meaning

The critical view that TTS urges us to adopt of language may seem radical. However, in light

of the linguistic paradigm’s utter failure at explaining how meaning can transcend culture and

language through translation, we should view the option presented by TTS as a welcome alternative

as it provides us a way to understand how meaning and knowledge can transcend culture and

language. For the account to work, we must deprivilege the role of language in determining meaning.

Language and any semiotic system must be viewed as indeterminately meaningful in order to

understand how textual translation may be determinate. Anything less would render it difficult to

understand how translation was successful. This may seem like a radical option, but is a moderate

compromise between two untenable extremes.

126
 On the one extreme we have the view that there are meaning facts about a language that are

semantically foundational and determinative and these can be discovered through some special

technical methodology (say, using mathematical logic) just as a chemist is able to discover the

periodic table and the molecular composition of substances in a laboratory. This is probably the

orthodox view of semantics in analytic philosophy. On this view, there are facts of the matter as to

what words in a language mean and contrary accounts are clearly wrong and demonstrably so. Such

facts have prescriptive implications both for language use and all areas of inquiry. Any area of

putative inquiry that cannot be made to square with these meaning facts is hokey and fraudulent.

On the other extreme we find the view that there are no facts about the matter as to what

words in a language mean. Claims about the meaning of phrases and expressions in a language are

speaker’s efforts to influence and negotiate their experiences with others through language: it is all

prescriptive and a matter of dialectical interplay. This is not an orthodox view of meaning in the

analytic tradition. It shows up in the all too fleeting deconstructive aspects of the later Wittgenstein’s

thought. The Wittgensteinien linguist Talbot Taylor holds such a view (Taylor 1997c). And he finds

a kindred view in Roy Harris’ school of linguistics founded in the 1980s first at Oxford known as

Integrationism. Harris describes his position thus:

The integrationist approach to language rejects the ‘language myth’ that has dominated Western
thinking on the subject for centuries past. This myth continues to dominate modern linguistics, whose
orthodox exponents postulate idealized linguistic communities bound together by shared systems of
known rules and meanings. The integrationist agenda offers the prospect of an alternative: a
demythologized linguistics which corresponds more realistically to our day-to-day communicational
experience. High on this alternative agenda are the demythologization of the concept ‘language’, the
demythologization of the connexions between speech and writing, and the demythologization of the
linguistic relationships between individual and society. (Harris 2005a)

Harris contrasts his view with “Segregationism” which holds that communication between parties

presupposes the existence of a shared system of signs, and in our terms, determinate meanings. His

127
view that all signs, including linguistic signs, are products of the communicational process, not its

prerequisites.

Harris’ view seems very similar to the view I am advocating. In his early book, the Language

Myth, he writes:

The language myth is the product of two interconnected fallacies: the telementational fallacy and the
determinacy fallacy. The telementational fallacy is a thesis about the function of language, while the
determinacy fallacy is a thesis about the mechanism of language. Although logically independent, the
two fallacies complement each other. Historically, too, they are closely associated.

According to the telementational fallacy, linguistic knowledge is essentially a matter of knowing

which words stand for which ideas. For words, according to this view, are symbols devised by man
for transferring thoughts from one mind to another. Speech is a form of telementation....

The determinacy fallacy, or ‘fixed code’ fallacy (as it might alternatively be called) provides for the
explanation of how the telementational process works, and indeed of how telementation is possible....
Individuals are able to exchange their thoughts by means of words because-and insofar as-they have
come to understand and to adhere to a fixed public plan for doing so. The plan is based on recurrent
instantiation of invariant items in two respects: form and meaning. Knowing the forms of sentences
enables those who know the language to express appropriately the thoughts they intend to convey.
Knowing the meanings of sentences enables those who know the language to identify the thoughts
thus expressed. (Harris 1981, 9-10)

I have much sympathy for Integrationism as a corrective antidote to the implausible view that

there are determinate meanings that characterize a language. The only way that the notion of

determinacy linguistic meaning can be made sense of is if one abstracts from real language some

idealized language, distinct from natural language that hovers over real language like an unhappy

ghost who does not get the respect it feels it deserves among the living. However, the rejection of the

view that there are semantic facts about a language comes up short too. There is a point of analysis

from which meaning is a function of the activities of persons, but this activity leaves a trail: namely

texts. Texts stand as semantic snapshots of wider semiotic activity, and the meaning the freeze is

fixed not by reference to linguistic activity, but by reference to its translatability according to text-

128
type. Harris, like most who criticize the orthodox views on meaning, holds that signs only have their

meaning within a context. But context, as it is conceived typically these days as defined by a time,

place, and interlocutors, is not the same as cotext. Cotextuality thus stands as a unique dimension in

meaning that is not reducible to wider linguistic activity for it involves institutional norms that can

mediate snapshots of language use. 69

If it were truly the case that there are no semantic facts about a language, then linguists would

never be able to trace the historical change and progression of meaning of expressions in a language.

The Oxford English Dictionary does just this for almost every entry. Not only do lexicographers and

linguists study and report the meaning of expressions through historical development but this is quite

indispensable to training scholars, and producing translations. Even if one’s view is that the meaning

of signs develops in a historical context, there are still social scientific facts about this development

that can be tracked. It just so happens that the semantic aspect of this development is indeterminate,

which is to say that by virtue of the context transcendence of meaning itself, the usage that persons

put to signs does not exhaust their meaningfulness, but forms a partial point of reference for it. The

context transcendence of meaning cannot be tracked if we restrict our attention to language use

among speakers. In other words, the only way to understand how meaning transcends context is by

thinking seriously about translation. But real translation has always proceeded by textual

considerations, and text-types take us in a different direction than language, for their canonical

specification is not reducible to the categories or grammar of any particular language.

Thus, the reasonable position is to hold the middle ground between the view that there are

determinate semantic facts about a language (i.e. that language is characterized by determinate

meanings) and that there are no semantic facts about a language. The middle ground holds that

language, though meaningful, is indeterminately meaningful. Languages develop and change, and as

they do they display new meanings, or new sides of old meanings. At any point we can swoop in and

try to abstract from this historical process some meanings, but this process will always be partial just

129
like any empirical observation. To call such observations un-objective would also be to make an

error. Our empirical observations are not subjective, but neither are they the whole story or fully

objective. The notion that there can be a final story about meaning is probably a mistake. What we

require is not a final story about all meaning, but rather an account of how meaning can objectively

transcend context, and here the text-type is an essential and determinative ingredient. Texts can be

comprised with such devices and on the basis of an appreciation of the meaning of the devices in a

text, along with a text-type to guide the translator in attempting to determine a semantic synergy

among the components of the text that is translatable. Moreover, in recognizing texts as determinately

meaningful, we can recognize their normative and regulative role in language use within

communities. Text-type institutions, such as the various disciplines of the modern academy, perform

an invaluable service in standardizing language use, across great swaths of land by the instruction that

authorities in these institutions provide to members of cultures at large. This normative role that texts

of a definite type play in a society for language usage cannot be underestimated. But for text-type-

theoretic conventions in reading and authoring texts, language would be quite unstable.

TTS thus fills a void in the current philosophical landscape that is sorely wanting. At present

philosophers and linguists believe they must choose between two extremes, neither of which are

satisfactory. The first extreme is the orthodox notion that definite and determinate meanings

underwrite language and the other extreme is the radical view that there are no such semantic facts

about language. Neither extreme does knowledge justice for neither can show how knowledge can

transcend context.

III.2. Why Determinate Translation Must Preserve Meaning

Many philosophers trained in the prevailing linguistic paradigm find TTS alarming. After the

previous three sections I hope TTS will now seem far less threatening. The notion that linguistic

meaning is indeterminate is frightening if we do not have any way to account for the objectivity of

meaning and knowledge. TTS provides us with just such an account. However, for the account to

130
truly work, we cannot view language to be the semantic foundation that translation must always bend

to, but rather as a malleable material for the creation of TTs that are semantically equivalent to STs.

There are limits to how much bending of a TL we can make use of, and these limits are institutional:

if it should turn out that our translations employ the TL in such a manner that even the authorities in a

text-type institution who are fluent in the TL cannot make out what is written, translation has gone off

its tracks. The institutional framework of a text-type provides the human resources to make language

malleable in just the right way. The text-type provides the language independent criteria against

which translations can be assessed as well as the epistemic claims of an institution. All of this works

splendidly to explain and embrace the reality of linguistic and cultural difference in tandem with the

objectivity of knowledge.

From what I have seen, philosophers wedded to the linguistic paradigm will often grant that

translation is a textual process mediated by text-types once the concept of the TEXT-TYPE is explained

to them, but at the same time resist the conclusion that language is semantically indeterminate. To

take this position is to embrace a contradiction, I think, or at least to embrace a position that is deeply

riddled with tensions. This reaction is like being in a state of denial at the loss of a loved one. To

adopt this state of denial is to affirm that we can succeed in translation, but fail to preserve meaning.

Translation has little value apart from its ability to preserve semantic content across semiotic systems.

This is not simply a bullet that we can afford to bite.

If meaning were fundamentally a feature of language in the first instance and if it was there

that determinacy in semantics was to be found, though translation is adequate when produced along

text-type-theoretic lines, then it would be possible to translate a text according to every discipline

relative, text-type-theoretic expectation, and to fail in no respect (that is, we preserved in translation

every feature of a ST that is necessary for the TT to be equivalent to it on text-type-theoretic

grounds), and yet we could not attribute the translation or any proposition abstracted from it to the

author for the text and such propositions would be semantically distinct from their original.

131
Professors could not teach their students about the opinions of figures who wrote in different

languages from the language known to the students, for the best translations of the works of such

authors would be semantic fakes. Science as an international, cooperative pursuit spanning several

linguistic communities would be an epistemic sham, based upon the apparently dubious notion that

scientific texts could retain their meaning in translation. Darwin’s Origins of the Species could only

be accurately talked about in English. We couldn’t talk about it accurately in French, Swahili, or

Korean, for our best translations of the text, which form the basis from which we would abstract

propositions and views, would be semantically distinct from the artefact produced by Darwin. This is

a reductio of the linguistic paradigm. But it shows that the linguistic paradigm in philosophy is vastly

out of touch with reality and practice. Disciplines that span cultures and languages such as the

sciences, and even philosophy, are paradigm cases of successful epistemic institutions and if our

received views on meaning cannot underwrite their operation then we have good reason to reject such

commitments. Even if we do not think that there is any particularly striking substantive knowledge

that philosophy yields, it at least yields knowledge of the history of philosophy that students across

cultures and languages can and do learn about. Indeed, so much of the intellectual history of the West

is bound up with the history of philosophy as it has been accessed in translation. (Where would Kant

be without a translation of Hume, who he read, perhaps only in snippets? cf. Wolff 1960.)

One might object that my argument here to the effect that linguistic meaning is semantically

indeterminate while textual translations according to text-types are semantically determinate is a

weak argument because it attempts to appeal to our intuitions that translation must be possible despite

the apparent trouble it seems to be in. Is TTS based upon an appeal to intuition? Not at all.

My argument is not that we should accept TTS because it appeals to our intuitions or that it

flatters them. Quite the opposite: my argument is that we should be embracing TTS because it

grounds meaning in epistemic institutions, namely their text-types that act as an independent check

against intuition mongering. The notion that determinate meaning is fundamentally linguistic appeals

132
to our intuitions because our intuitions are, or are based on, our generalizations of language use. It is

in our intuitions that we are given seeming meaning facts about language and thus the linguistic turn

flatters this empirical ability of ours. But I think this counts against it. Our intuitions are our

intuitions: there’s no reason for us to think that there is anything objective about what is intuitive. My

argument for TTS in contrast is an attempt to appeal to facts that are not a matter of intuition, but

plain empirical fact. There are all manner of epistemic institutions that transcend cultures and

languages, and operate internationally, and these are where our best or most reliable forms of

knowledge come from. Most every society in the world wishes to benefit from the knowledge that

such institutions provide, and these institutions could not function if meaning were fundamentally

localized to a language, for there would be no way to explain how ideas traverse cultures and

languages through these institutions. The linguistic paradigm cannot explain, for instance, how

scientific knowledge in China, articulated in Chinese texts, is semantically equivalent to English

language texts if they are able translations of them. No theory of meaning can explain the success of

science if it cannot explain the translation of scientific texts such that the good translations are

semantically equivalent to their original. Only TTS can explain this for only TTS recognizes the

semantic role of text-types to mediate the reality of cultural and linguistic difference.

I think the attempt to ground the objectivity of meaning on language is an absolutely muddled

idea. It is absolutely muddled because the meaning of terms and expressions in a language are subject

to all manner of culturally and historically contingent influences. This is a reality. It would be rather

odd if linguistic meaning always correlated with what our best epistemic institutions had to say.

Given the cultural contingency of language, it seems rather strange to try to ground objectivity on it.

But yet, philosophers wedded to the linguistic paradigm often attempt just this. Paradigms are

cultures, and it is difficult often to be critical of one’s own culture and thus I am not wholly surprised

at the resistance that TTS encounters. But it speaks to the rather ill informed status of philosophy

133
today, which does not attempt to square theory that is supposed to underwrite objectivity with natural

facts, such as cultural diversity.

But yet the defenders of the linguistic paradigm may resist the conclusion I am pressing. One

way to resist the conclusion is to beg the question. One might argue, for instance, that the only

justification for the notion that translation preserves meaning is if translation is always literal

translation. (As Quine writes: “the more literal translation is seen as more literally a translation.”

Quine 1960, 75.) One might argue that if one gives up the notion that translation is always an attempt

to preserve literal meaning or truth-functional meaning, one thereby gives up the right to hold that

translation preserves meaning, for literal meaning just is meaning: everything else is pragmatics.

This begs the question against my argument by assuming that only literal meaning is

meaning. I think we need to ask why meaning is restricted to these narrow categories of truth-

functionality or the “literal”. The pressure is on the traditional semanticists to justify keeping the term

“meaning” for such a narrow range of semantic phenomena, when all manners of features of a text

can be preserved in translation given the appropriate text-type.

The prominent justification for keeping “meaning” for the literal or truth-functional these

days is the assumption that these semantic modalities are the most basic and systematic (or,

alternatively, conventional) in a language and by virtue of understanding these, one can understand

other phenomena, covered under the heading of “pragmatics”(cf. Cappelen and Lepore 2005; Salmon

2005). This is the justification provided by what we might call formal semanticists, and I include

Davidson in this group (though I realize that many who consider themselves formal semanticists may

wish to distance themselves from Davidson). But what has not been noticed is that this is not a

semantic argument, but rather a pragmatic argument that tacitly appeals to our sense of what is

necessary for speakers to negotiate language use in a linguistic community defined by a shared

language. Meaning is thus defined not in terms of its objectivity, but in terms of how it helps us

understand the various things that people do with language. Meaning, in contrast, if it is to be

134
objective and of any worth, must transcend contexts, such as speech communities and languages, but

the only manner in which this is possible is by translation. Since translation is only determinate at the

textual level, where text-types are brought to bear, and since translation is the only viable test of the

objective character of meaning, it follows that any “feature” of a text preserved in translation counts

as meaning for any feature so identified by a text-type and preserved in translation is shown to

transcend the provincial contexts of speech communities. There is no basis from the perspective of

translation theory to privilege any type of meaning aside from text-types and their features. When we

have text-types in view, the traditional linguistic conception of meaning is a poor cousin, for

linguistic meaning is an incurably local, language relative conception of meaning. Text-types, in

contrast, show how meaning can transcend any semiotic system. It and only it can show how meaning

is truly objective for it and only it explains how meaning comes apart from systems of semantic

representation, such as languages. The reason it can do this is that it invokes a semiotic intermediary,

namely the text-type that is not reducible to any particular language or semiotic system as such.

One might object that the argument I am presenting here has mistaken natural language for

ideal or real language. “Literal meaning” is not a feature of natural language, but language conceived

as an abstract system. The literal meaning of an expression in such a system is its systematic function.

To shift the focus to ideal languages or abstract languages instead of real natural languages is

to undermotivate the argument for conceiving of literal meaning as foundational to understanding

pragmatic phenomenon, for the pragmatic phenomena that literal meaning is supposed explain occurs

in actual contexts of language use, not in abstract scenarios. The problem however is deeper. The

notion that meaning is best thought of as a systematic property of an abstract system does not save

objectivity: it cements incommensurability. The advantage of the view that I am forwarding is that it

makes the objectivity of meaning trump the contingencies of semiotic systems. As new ideas make

their way into a culture by translation, pressure will be exerted on the meanings of terms and symbols

in a culture to accommodate these new ideas. At times, this influx will cause a creative tension that

135
will lead to much practical changes in a culture, at other times the new ideas will simply add a layer

of significance to a pre-existing symbol. Meaning is objective because it can not only defy language,

but it can also make language bend to it in translation. This is what texts of a definite type do for us.

They stand as stubborn resistance to language use thus guaranteeing the semantic objectivity of the

content it freezes.

The defender of literal or truth-functional meaning is someone who plants their feet on what

they take to be the facts of a language and they try to understand everything in this light: science,

literature, philosophy, are all understandable by virtue of the contingencies of language. This, I think,

makes science, literature and philosophy completely regional. We couldn’t understand or

determinately translate the science and philosophy of speakers of different languages for it would be

articulated with literal meanings that are native to that language, not ours. TTS, in contrast, in making

language take second seat to the determinate meaning of texts shows how language use and its

meanings must at turns be compromised or criticized in order to make room for scientific innovations

or in order to translate or articulate novel perspectives. This speaks to the semantic objectivity and

determinacy of such views, for its determinacy is not contingent upon the culture of our language.

The defender of traditional or “literalist” semantics will likely object to my account and

claim that we cannot deprivilege something like literal or truth-functional meaning for this is the

means by which we come to understand more complicated expressions. Davidson, for instance, has

argued that we understand metaphors by means of the literal or truth conditional meaning that

comprise them ((Davidson 1996 [1978]). There seems to be no way around this. If the literal or truth-

functional is basic in understanding such meaning, it shows that there is something like semantic

determinacy for expressions in a language, independently of questions of whether such expressions

can be determinately translated. We determine such matters of literal meaning with respect to

systematic considerations within a language. Translation will simply have to wait till we have settled

such questions once and for all and if it should be indeterminate because we cannot secure a unique

136
one to one relationship between expressions across languages based upon their literal meaning, c’est

la vie. Failing an appeal to the literal meaning of expressions, we would not be able to identify

metaphors and poetic meaning, and we would thus not be able to preserve such poetic meaning in

translation for we would not be able to identify it in the first place.

The notion that the truth-functional or literal meaning of an expression is central to

understanding all meaning even as it appears in texts is unsustainable. If the lyrical or musical

qualities of a text are features to be translated, they are semantically significant and count as a type of

meaning. These qualities of a text are not understood by reference to their literal content, whatever

that might be, but directly by reference to their musical quality. But even in the case of metaphors,

that do often seem to be intelligible by reference to their literal meaning, we run into the problem that

there is rarely if ever simply one account of the literal or truth conditional meaning of words or

sentences in a language. The notion of literal meaning is foggy. Wittgenstein’s argument that what we

have in language are not clear categories but rather family resemblances is one way in which the

fogginess of literal meaning has been recognized in the tradition. It is almost always the case that

speakers do not agree on the exact formulation of an account of the literal meaning of an expression

(simply consult a few dictionaries to confirm this) and as a result the “understanding” that people

derived by applying their knowledge of literal meaning in interpreting metaphors will diverge, and

their accounts of the metaphor will also so diverge. Or, if the literal meaning is so basic so as to

escape no one (say, a T sentence of the sort “‘Rudolph has a red nose’ is true if and only if Rudolph

has a red nose”) then it becomes quite mysterious as to why some persons are able to authoritatively

explain the significance of metaphors (persons such as scholars of literature) while others who are

perfectly competent users of a language cannot and how they might fail to notice that a sentence is

not simply a false descriptive claim but also metaphorical.70 Literal meaning as such is quite

insufficient to explain such matters. Text-types, in contrast, provide us with cues on what to

recognize as salient in a text, and their cues can help us understand whether a sentence in a text is

137
simply a false descriptive claim or a metaphorical claim, and the determination of such factors will

have a clear implication for how the text, and particularly the translation unit, is to be translated. Such

questions cannot be answered in the abstract, holding before one’s mind a sentence, analyzing it to

discern whether it is a false descriptive claim or a metaphorical claim. Rather, such matters can only

be determinately answered by cotextualizing a sentence within a text of a certain type. The synergistic

unity in a text (and there may be many such unities in a text) is text-type-theoretic and this directs the

translator and expert to properly focus on some features of a text in a manner that would allow them

to explain whether a claim is metaphorical and why. The closer such accounts come to a text-type-

theoretic explanation that would allow for the determinative translation of a text that necessarily

retains metaphorical meaning as one among its text-type features, the closer such an explanation is to

semantic determinacy.

We have in part a response to the following objection: perhaps text-types are necessary to

translate poetry, but they are unnecessary to translate scientific texts. In the case of science, all we

need is an understanding of the literal meaning of expressions, for with this understanding, we can

translate scientific texts, word for word, sentence for sentence, on the basis of their literal meaning.

By now it should be clear that translation is not ever a word for word exchange in interlingual

translation, owing to syntactic differences across languages. Thus, understanding the literal meaning

of words is not sufficient to translate any text. But the objection issues from a deeper state of naivety.

Even if translation in the case of science were a matter of trading literal meanings across languages,

the decision to translate a text as a text of science is a decision that is text-type-theoretic. Thus, text-

types cannot be dismissed on the putative score that in the case of science, they are not required. But,

at any rate, even in the case of science, translation cannot be a trade of literal meaning across

languages because languages are different. If science were fundamentally about literal meaning, it

would be fundamentally linguistically relative given the prevailing account of literal meaning, as the

most broad and systematic role that an expression plays in a language. The importance granted to

138
literal meaning by the linguistic paradigm epitomizes the failure of the linguistic paradigm: ironically,

literal meaning is linguistically relative.

What then about truth-functional meaning? Traditionally conceived, truth-functional meaning

is intimately tied to literal meaning. Truth-functional meaning will thus inherit all of the

provincialism of literal meaning, to the extent that it is a function of literal meaning. This is one of the

lessons of Quine’s Word and Object. Truth-functional meaning might be thought to come apart from

literal meaning traditionally conceived. For instance, François Recanati has distinguished between

three types of literal meaning: the literal meaning of a sentence based upon its conventional meaning

only, the meaning of a sentence that comes about by saturating context-sensitive expressions such as

indexicals with their appropriate contextual referents, and a literal meaning that is directly interpreted

by an interlocutor in a context, which is distinct from the other two sorts of literal meaning.

Metaphorical claims can be in this latter category of “literal” meaning. Thus, for instance, when

someone says “The ATM swallowed my card” the hearer understands this to be a literal assertion that

the ATM did not return the card that the speaker deposited in its slot even though ATM machines

can’t swallow because they do not have mouths. On this account, metaphorical meaning can also be a

type of literal meaning and it is assessed truth-functionally, in light of the type of literal meaning it is.

Thus, on this account, it can be true that the ATM swallowed my card (cf. Recanati 2002). Must

literal meaning thus conceived be central to translation?

It could happen that in a play or text of poetry the sentence “The ATM swallowed my card”

is not a literal assertion of the type that Recanati identifies. Rather, it might have a metaphorical

function that must be treated in translation functionally as the translation theorist Eugene Nida has

identified (Nida 1964, 159-62). None of the types of literal meaning that Recanati identifies will help

us discern this. What helps us discern the semantic functioning of any sentence is that it is in a text of

a definite type. Here, it is the text-type that does the crucial work, and we can apply multiple text-

types in translating the same texts. Contextual meaning of the sort that Recanati recognizes is central

139
to translation only in a vacuous sense that translators must always be sensitive to the various nuances

of language in translation in light of their translational projects. There are often culture and language

wide expectations and conventions in language use and many of these can be charted. Descriptions of

how people arrive at interpretations of utterances in common contexts are possible. Speakers of a

language often rely upon such generalizations that they derive from observing and participating in

language use. There are likely predictable ways in which speakers of a language derive certain

pragmatic inferences from language use within context. The translator can and must make use of all

such information, along with a general appreciation for the grammar of a language—much of which

she has access to by virtue of her linguistic competence. But when and how such meanings are to be

preserved in translation are not set by the rules and generalizations of language use at large. The

incorporation of such considerations in producing a translation manages to show how the meaning

that seems, on the face of it, incurably context-bound, is liberated through translation, but the

conditions for its objectivity are not set by the context but by the text-type and the contingencies of

the text being translated. It is thus only within a text of a definite type that a sentence comes to have a

determinately translatable meaning. Crudely put, text-types are to texts what contexts are to

utterances. The difference is that the text-type determines the significance of units in a text of a

definite type in a manner that alternative determinations by the same text-type are semantically

equivalent, according to TTS. The same cannot be said for the role of context in disambiguating

utterances for here none of the disambiguated sentences are determinately translatable for we do not

know what type of text they are part of. In the context of utterance, there is no text as such. We could

abstract from a context a transcript and treat it as a text of a certain type, but that would be a different

scenario, and it would no longer be a matter of understanding language use contextually, but

cotextually.

But what does this say about truth-functionality? TTS in regarding text-types as the ultimate

ground of the semantic determinacy of particular discourses liberates the notion of truth-functionality

140
from sentences considered in the abstract to propositions abstracted from texts of definite types. We

can thus still recognize the truth-functionality of sentences, but this will not be determined by

systematic properties of language, on TTS’s account, but rather by the very text-type that underwrites

a discourse, vouchsafing its objectivity. Thus, truth-functionality cannot be used to define translatable

meaning, on TTS’s account, so much as it is a function of translatable meaning.

One might object after all of this that if TTS holds that STs and TTs are semantically

equivalent when translated according to a text-type, then works of literature and poetry that are

comprised of dramatically different literal significations and chosen for their functional significance

will have to be treated as semantically equivalent. Thus, we can have a poem in a ST and TT form

that are alike in no ordinary, literal sense, and yet we must regard them as semantically equivalent.

This is too much!

Indeed, this is an implication of TTS. But I think this is good for us. To recognize the

semantic equivalence of a perfectly translated poem or work of literature according to every textual

and discipline relative expectation is to recognize that the objectivity and content of meaning

transcend the limits we and our culture attempt to place on. This is part and parcel of what it is to get

over the provincialism and ethnocentrism that characterizes the linguistic turn. But notice, the same

principle that underwrites this equivalence underwrites the semantic equivalence of scientific texts in

translation that will literally not contain the same literal meanings across translations, though they are

equivalent according to all discipline relative expectations and criteria.

III.2.1 The Case of “Technical” and scientific Translation and the

Question of the Determination of Reference

So far I have argued that success in translation must always be understood as semantic

content preservation, or in other words the goal to produce a TT that is semantically equivalent to a

ST. Here is why this applies across the board: the principle that explains why a TT is equivalent to a

141
ST is not linguistic but text-type-theoretic. If the principle were linguistic, one might on the basis of a

distinction between literal and figurative aspects of a language argue that meaning in the proper sense

is only preserved in the case of literal translation for literal meaning is a foundational or basic aspect

of a language—provided that one could actually make a case that there was something like a

determinate literal meaning that characterizes language (something I have argued is doubtful, or

indeterminate). But as what underwrites treating a TT as equivalent to a ST is not a linguistic

principle but a text-type principle, there is no foothold from the perspective of translation theory to

draw a line between meaning as it is preserved in the case of the translation of scientific texts, or

meaning as it is preserved in the case of literary or poetic texts. Moreover, I also argued that the

notion of literal meaning is a poor grounding for objectivity as it is linguistically relative and thus

those who are motivated to anchor objectivity in literal meaning are really anchoring objectivity in

culturally relative phenomena.

One might argue that in the case of science what one is trying to do is preserve literal

meaning, understood as a type of truth-functional meaning. When we translate a scientific text from

English into Haitian Vodoun Culture Language (for instance) what we wish to preserve are the

“literal meanings” of the various expressions of a ST into the TL. Thus, translation at this point will

be one of preserving literal meaning.

The notion that what we are doing is always preserving literal meaning when translating

science is not quite accurate. It is far better to recognize that what we are preserving are truth

conditions of claims relative to a scientific theory. Scientists after all criticize and refine natural kind

terminology for their purposes, and this may not comport with what is actually basic or systematic in

a language (the tension between what is systematic about an expression in a language and its

divergence from scientific understanding simply speaks to its linguistic indeterminacy). Thus, a

successful translation of scientific texts preserves all the right inferential relationships among

technical terminology and so-called “natural kind” terminology in a ST and scientists will also have

142
reason to want this translation to preserve a relationship to the original context of discovery and

investigation. To this end, scientists will introduce technical terms and may even sharpen or use in a

technical way TL natural kind terms in a manner that would not be basic or common in the TL

language (which is to say, they may make use of terms in the TL as though it were a natural kind

term, even though its meaning might be broader71). But does this mean that translation gets along just

fine without text-types as intermediaries? Can we underwrite this as a purely linguistic process? Not

at all. The various considerations that scientists defer to in understanding the semantics of their ST are

not continuous with what is basic in the SL. Everything, from the refinement of SL and TL

terminology for special or technical employment in a text, to the text-type that allows the translators

of science to assess what relevant content to be preserved in a translation, is institutional and not

linguistic.

There are several features of this scenario that TTS explains, but the usual linguistic accounts

of meaning do not. First, a successful translation of a scientific text into Haitian Vodoun Culture

Language would have to be accomplished by scientists or specially trained technical translators who

were expert in the SL and TL, and able to explain the meaning of the novel terminology to lay

speakers of Haitian Vodoun Culture Language. In the absence of this institutional setting, there is no

way to make the translation stick. The objective is not to render the text intelligible on its own to

average speakers of Haitian Vodoun Culture Language, but to ensure that they need to know nothing

more than the relevant scientific theory and the text-type of science to understand the text. But here,

the text-type is invaluable.

It cannot be the case that every translator who has been successful at translating a text is

aware of text-types and the like. I suspect that most are not. Indeed, what is typical for translators is

for them to be rather unaware of large issues of translation theory. Thus, in the translation studies

literature, one finds silly generalizations about what is necessary in the realm of translation based

upon considerations culled from translating particular types of texts. We examined such bloopers in

143
the previous chapter when examining the case for functional linguistics and translation. What makes

for an objective translation is that the translation must be underwritten by the best translational

practices relative to a text-type. Thus, if our translators of an English scientific text into Haitian

Vodoun Culture Language conform their translational practice to what is appropriate for translating

scientific texts, then indeed text-types are indispensable to explaining why their translations are

accurate and semantically equivalent to their originals. The text-type, if nothing else, explains why

the putatively successful translation of a scientific text into Haitian Vodoun Culture Language

succeeds as a scientific translation, but why it may be a failure as a translation of the poetic or

dramatic aspects of the ST. Translation is always a process of selecting some features of a ST for

preservation in a TT. Translators due to their institutional training may be quite oblivious to the fact

that this is what it is that they are doing when they translate. This does not detract from the fact that

this is what they are doing.

One might think that what one has at stake here is a process of linguistic enrichment only, not

a process of textual translation. But the trouble is that it is only within the Haitian Vodoun Culture

Language text of science that the technical terminology that one introduces will have the type of

determinate meaning that scientists require. It may be the case that such terminology does not become

part of the wider Haitian Vodoun Culture Language or it may be the case that such terminology

comes to assume systematic functions in Haitian Vodoun Culture Language that diverge from the

narrow technical employment that scientists assign them in scientific texts of Haitian Vodoun Culture

Language. It is also likely the case that the technical terminology and technical employments of

natural kind terms in the text of science in English diverge from the basic role that such terms play

within English. Here, it is the scientific theory that is doing the major work in accounting for the

equivalence between a ST and a TT and this is not a creature of language at large, but particular types

of texts, for it is only here that their translatability is ensured. Linguistic enrichment may or may not

144
be successful, but the employment of novel technical terms in a TT can be semantically equivalent to

ST devices provided that they are in the right type of text, with the proper imprimatur.

What then about reference? Does the translation of scientific text have to preserve reference?

There are roughly two traditional accounts of reference. One is the Frege-Russell approach that looks

to description to determine reference. Recently, there is an account provided by Putnam and Kripke,

that looks to sociolinguistics and the deference of referential intentions to experts or baptismal

contexts to explain reference. On the latter account, reference is tied to naming. More recently, there

is a deflationary account from Paul Horwich that understands reference merely as what is delivered

by the comprehension of a term. These accounts are linguistic, in so far as they understand reference

as something determined by the operations of language. If reference is a linguistic notion, then it

seems that the translation of scientific texts, in so far as it attempts to preserve reference, is really

attempting to preserve linguistic meaning. Thus, the notion that what we have in the case of the

translation of scientific texts is text-type-theoretic would seem to be untrue. This is an illusion.

Ideally we do require that reference be preserved in the translation of scientific texts. How

scientists determine reference in the case of scientific texts is always indirect, through testing and the

consilience of inductions. Thus, for instance, determining what the term “electron” refers to is not a

matter of deferring to wider linguistic practice. Rather, it is a matter of identifying the descriptive

theory that this term plays a role in articulating, testing the theory against its predictive implications,

and further confirming the objectivity of the description by the ability of the theory to gain support

from other descriptive theories. On this score it seems that reference in scientific texts is not primarily

a linguistic phenomenon for linguistic reference, if there is such a thing, can be determined by

linguistic factors, such as the descriptions that speakers associate with terms, the experts they defer to,

their referential intentions, the judgments that people employing a term converge upon or perhaps the

function it serves in society. These are not the factors that scientists typically rely upon in order to

determine whether their terms refer or not. Thus, while reference is ideally preserved in the case of

145
scientific translation, it is something determined by scientific methodology. Of course, scientists may

take an interest in some item previously identified in a linguistic practice as discrete (such as a

species) but whether or not scientists will continue to treat such natural kind terms as actually

referring to a kind or as vague and capable of regimentation by other scientific considerations is

determined by scientific considerations, not linguistic considerations. These considerations are, as we

saw in a previous section, based upon the text-type of science not language as such. Thus, the

translation of scientific texts is not an exception to the account I am providing under TTS, but rather

explained by TTS.

This response to the question of the determination of reference generalizes. What does

“Romeo” refer to: “Is Romeo tall?” Does this question have a determinate answer when posed in the

abstract? Of course not. But how do we determine reference in general? Here texts of definite types

and their institutions are invaluable. It is within a text of a definite type, and with respect to such

definite texts, that answers to such questions become tractable. Without texts, even proper names in

society at large would be difficult to determine. There are several Shyam Ranganathan-s in the world.

Without legal documents specifying various characteristics of the various Shyam Ranganathan-s, it

would be quite impossible to determine just who is and who is not referred to (and when) by the name

“Shyam Ranganathan.” How we determine reference will vary with the type of text that we call upon

to help us out in determining reference. In literature, determining what answers to “Hamlet” will not

involve the scientist’s test of descriptive theories in light of their testable implications and the

consilience of inductions, but rather knowledge of literary texts that can only be read and translated

property qua literary texts according to the text-type of literature. Professors of literature would thus

be able to tell us what the name “Hamlet” refers to when we intend to refer to the character of

Shakespeare’s famous play (we might be attempting to talk about a “Beanie Baby” stuffed animal pig

by that name in which case a different set of quasi-institutional considerations will help us determine

reference). Whether “Shyam Ranganathan” refers to me will depend upon a host of legal documents

146
supporting my contention that I am Shyam Ranganathan. The characteristics they specify can only be

understood determinately in light of the requirements of legal text translation.

Texts of a definite type and the institutional procedures based upon them, help us

disambiguate and determine reference in a way that language as such cannot. What then does this say

about the traditional theories of reference? It teaches us that mere baptism and description is not

sufficient for reference. (I can name myself Fred Smith and I can tell others that I am Fred Smith and

they can start calling me Fred Smith: that doesn’t determinately make me Fred Smith—though

contextually, it may do for a time.) We can determine reference in many different ways, but these are

largely in connection to institutional considerations based upon text-types. And these considerations

are not reducible to the language they are articulated in.

With Paul Horwich, we can agree:

…the trouble with the description and causal theories [of reference] is not so much that they are
wrong, but that they are not what we are looking for; they are not really theories of reference.
Properly understood, one of them gives a theory of the meanings of names, the other a theory of
sociolinguistic deference; but neither makes any attempt to tell us about our conception of x referring
to y. (Horwich 1998, 117)

The trouble with our traditional theories of reference is that they do not actually tell us how we

determine reference in all cases. The reason they do not tell us this is because there is not simply one

way to determine reference. Language as such is constituted by the multiple references that

expressions acquire through their usage and development, and the determination of reference is thus

left to textual institutions. The determinations of such institutions can be thought to be objective

because they are not based upon mere cultural contingencies or the convenience of context. Thus, in

recognizing the role for textual institutions in disambiguating and determining reference, we also put

some distance between ourselves and the deflationary account forwarded by Horwich. As an account

of language, perhaps the deflationary account tells us something important (namely that hidden or

mysterious aspects of reference are not necessary for meaningful language use) but we also recognize

147
that there are facts about reference that language users cannot access on their own, and these require

the operation of textual institutions that transcend languages and cultures.

III.3. Text-Types do not Reduce to Systematic or Literal Features of a

Language
To resist the conclusion I am urging, that determinate meaning is really a property of texts of

definite types, and not language, one might try to argue that a systematic understanding of a language

yields knowledge of text-types. If this were so, then we could rely upon something like literal or truth

conditional meaning to unlock all the text-type-theoretic considerations necessary to deal with

translation and meaning. I think this hypothesis is false on the facts.

(1) If my argument in my Ethics and the History of Indian Philosophy is correct, Indologists

have misunderstood the history of Indian philosophy, particularly its moral philosophical content,

because they thought that that their careful study of the Sanskrit language would provide them access

to the philosophical content of Indian philosophical texts. Despite their scrupulous study of Sanskrit

and all aspects of its grammar, including its vocabulary, they were by and large unable to understand

how moral philosophy played itself out in the history of Indian philosophy because they thought that

this would be a manifest linguistic fact. The fact is that language is ambiguous and polysemous and if

one does not understand the textual functioning of key philosophical terms as discourse markers, one

will not be able to understand the philosophical import of such terms in a text, even if one is perfectly

competent to read and “understand” by some account the texts one is studying. The trouble with

language is that it is incredibly polysemous and one will always be able to get something out of

reading a text. Linguistic competence might provide one with that amount of facility with a language

(to allow one to get something out of what one reads) but this is unlikely to be anything text-type-

theoretic and will likely reflect the prejudices that one brings to reading texts.

(2) It is possible for perfectly competent users of a language to fail to appreciate the text-

type-theoretic meaning of a text. It often takes years of special institutional training to recognize this

148
vector of meaning. We philosophers actively attempt to impart this knowledge to our students. If

literal meaning were sufficient to underwrite such text-types, such special training would not

apparently be necessary.

(3) If text-type-theoretic knowledge could be learned from a systematic understanding of a

language or languages, we wouldn’t need courses in philosophy: linguistics would provide us

everything we need to know about philosophy and every other text-type. Indeed, all the various

departments in the modern academy could be folded and linguistics could be instituted as the only

subject offered by a university for an understanding of the systematic features of a language or

languages provided by linguistics would allow us everything we need to know about all the various

text-types that are today apportioned to separate departments. We wouldn’t need natural science or

mathematics as a separate academic pursuits characterized by their own semantic concerns to be

preserved in the translation of scientific and mathematical texts: linguistics would do just fine as the

master academic discipline. Linguistics would provide us with the categories and knowledge to

generate the various text-types. Moreover, each discipline as we understand them today would simply

be a sub-branch of linguistics, emphasizing some regional area of interest, as a course on phonology

or syntax is today such a regional area of interest in linguistics. Medicine and accounting would be

sub-regions of linguistics. According to TTS, in contrast, there can be no master discipline that all

disciplines are reducible to. Each healthy and unique discipline is comprised by its own type of text.

This, I think, conforms to the facts.

(3) The analytic process that is necessary to produce good text-type-theoretic translations is

quite distinct from linguistic knowledge. Linguistic knowledge pertains to individual languages. The

type of analytic process necessary to produce a good translation must mediate differing languages.

The linguistic paradigm assumes that this mediation is achieved through a deeper understanding of

linguistic commonalities between languages. But by now we should know that this is not true and

implausible not only because languages are different, and thus the commonalities that languages share

149
are too week to underwrite translation, but also because the very activity of translating in light of text-

types leads us to think about the content of a text, abstracted from its actual mode of expression. And

we abstract this content in our analysis not by reference to the categories of a language but by text-

type-theoretic concerns. Indeed, if we couldn’t abstract the content of a text from the particular mode

of linguistic expression we would not be able to recreate a TT with distinct linguistic constraints that

all the same mirrors the ST. In other words, if we had to rely upon the systematic aspects of a

language to mediate languages in translation we couldn’t be successful for we would have never

conceptualized the meaning to be preserved in translation as something that could come apart from its

particular linguistic expression.

In the last chapter I had brought up the case of G.E.M. Anscombe’s able translation of

Wittgenstein’s Philosophical Investigations as an example of how translation is not a simple word for

word, sentence for sentence replication across languages based upon some grammatical or literal

equivalence of the words and sentences but involves an evaluation of a text in light of a text-type that

provides the direction for how a TT is to be recreated that will mirror the ST according to its text-

type-theoretic significance. Anscombe deferred to her implicit knowledge of the text-type of

philosophy in her translation of Wittgenstein. The text-type of philosophy, as noted in Chapter 1, is

concerned with the articulation, application, investigation, debate and criticism of theories of a

universal and general nature, canonically articulated with ordinary symbols, such as “real,” or

“meaning”. Anscombe applied this knowledge of philosophy to her analysis of Wittgenstein’s text

and produced a translation that saved the philosophical content of Wittgenstein’s text in translation by

making adjustments that did not rely upon a word for word, or sentence for sentence exchange across

languages based upon some putative understanding of their literal meaning, but on an understanding

of the philosophical important of Wittgenstein’s text. She also relied upon the strategy of text

moulding to constrain its semantic contribution to the overall text. If Anscombe were to have treated

150
Wittgenstein’s text as a text on the grammar of the German language, and not as a text in philosophy,

the adjustments she would have made would have been quite different.

One might object at this point that Anscombe’s translation of Wittgenstein was based upon a

systematic understanding of words in a language. Part of that understanding would have led her to a

systematic understanding of the word “philosophy” and its cognates in other languages. Given

appropriate understanding of this word, she would have realized that the text is a philosophical text (it

is after all in the title: Philosophische Untersuchungen) and this would have lead her to interpret the

text in light of the literal meaning of the term “philosophy” and its systematic relations with all the

words in the original German text. The “text-type” would simply be an abstraction from this process

of linguistic analysis and would pertain to an appreciation of the pragmatic implications of literal

meanings constructed into philosophical texts.

In response we must ask our defender of literal meaning whether Anscombe could translate

this text labelled “Philosophische Untersuchungen” into any language: must the TL always have a

word that literally means PHILOSOPHY? The Indian intellectual tradition has an enormous corpus of

texts (only a miniscule fraction of these texts have been translated into Western languages) and these

texts are philosophical by all reasonable lights. They present arguments for universal and general

theories of ethics, metaphysics, epistemology and logic. But in Sanskrit, there is no obvious word for

“philosophy.” There are some candidates, but none are good synonyms. Indian philosophers never

grouped themselves under a category like “philosophy.” Does this mean that we cannot translate

Wittgenstein’s text into Sanskrit? And what of the reverse? Given that Sanskrit lacks such a clear

term as “philosophy,” how then do we translate philosophical texts written in Sanskrit into English?

Must a culture have a word for “poetry” in order for them to have texts of poetry? I’m not sure why

we should think this. Particularly if translation is a process of preserving some features of meaning

but not all, the notion that the actual text-type must be native or explicit in the very language of a

culture becomes under-motivated.

151
 Here we come back to the core of the problem of assuming that literal meaning must

underwrite all meaning and translation. Literal meaning is the putative meaning of an expression that

is its most basic or systematic significance in a language. It is thus an incurably linguistically and

culturally relative notion. If literal meaning had to underwrite all translation, translation couldn’t get

off the ground for literal meaning is linguistically and culturally relative.

III.4. The Objectivity of Determinate Translation

According to TTS, objectivity in translation is a matter guaranteed by the institutional

oversight of translation in accordance with the relevant text-type. Translation is determined by text-

type-theoretic considerations, but such considerations give rise to institutional knowledge that all play

their part in rendering translation determinate. Translators who are competent to translate a text are

knowledgeable both in the respective source and target semiotic system, but most importantly in the

relevant text-type. There is no way to translate a text objectively without concern for text-type for

there is no way to situate one’s approach to extracting content from a text in light of impersonal

constraints in the absence of a text-type. I can, for instance, decide to translate a text by my own

lights, thinking that my linguistic expertise will allow me to figure out what the content of a text is. I

can afterwards decide, looking at my translation, that the text I have translated is clearly a text of

great importance for philosophy and science, for instance, but if I have not approached the text from

the start in a manner that coordinates with institutional (i.e., impersonal) standards, the resulting

translations will be inaccurate, not because they have not tracked some features of a ST to be

reproduced in a TT, but because they have tracked such features inconsistently and in no principled

fashion and in a manner that is not necessarily consistent with the criteria with which we criticize the

text afterwards. The constraint of a text-type is an invaluable corrective and restraint against the

tendency to produce translations that render TTs in a manner that flatter one’s own view on things. It

prevents us from simply relying upon intuition and instead provides us with a coordinated framework

to produce translations. If our intuitions were always what guided us in translation, translation would

152
be indistinguishable from a purely subjective superimposition of one’s interests and concerns on a

text. We would still be extracting some feature of a text in translation under these circumstances, but

just those that flattered or indulged us, rendering the TT more of an expression of the translator’s

personality than anything else. In contrast, by making the text-type the first constraint in translation,

one defers to criteria for semantic content extraction that one applies to the text on the whole, thus

circumventing the indulgence that comes about by translating according to intuitive evidence, and

sentence by sentence.

Translation is always a matter of choice between alternative ways of transforming symbols.

Translators know this. This is the fundamental problem that they must contend with. The traditional

problem of the indeterminacy of translation, as Quine conceives it, or even the correlatively radical

idea that languages are incommensurate, appear on the scene against the backdrop of the linguistic

turn because there appear to be many ways that we can translate terms and sentences across languages

but no language independent means to choose between competing proposals. It is not as though on the

strength of the linguistic facts alone we can narrow down the range of possibilities of how to translate

words and sentences across languages but that language itself can provide us no criterion in every

case to choose between competitors. Text-types provide us the principled and impersonal means of

making such decisions. And accuracy in translation thus comes about as a result of the principled

approach to translation where choices between alternative ways of rendering words and sentences are

justified and underwritten by the text-type that translators employ.

This raises the problem of the case of the translator, such as Anscombe, who is not a

translation theorist, and is not particularly cognizant about the relevance of text-types to translation.

Indeed, most translators are in this position. Philosophers who are translators simply come to the task

out of a desire to translate a text of philosophical importance, and not because they have a calling for

translation. They usually never think about translation until they have to start translating a text, and

the topic of translation as such remains far from their research.

153
 One does not have to be a translation theorist to get translation right; one needs to simply

translate a text in a manner that accords with a translation guided by the appropriate text-type. In

Anscombe’s case, as we saw, she translated Wittgenstein not word for word or sentence for sentence

(for that is never really possible anyway) but in light of the text-type of philosophy, which counsels

us to look to the theories and reasons articulated with key philosophical terms in a text, and to

understand and translate the text in light of the articulated philosophical theories or reasons. This is

exactly what Anscombe did. Where did she learn how to do this? She gained this knowledge by virtue

of being a trained philosopher. Does every trained philosopher explicitly access this knowledge?

Certainly not. But can we access this knowledge? The goal of this dissertation is to explicate this

knowledge.

Another concern that one might have with TTS is that it seems to render the notion of error in

translation inexplicable. It seems that there is no way for us to make an error for what we are

preserving is considered in itself indeterminate. How does one fail to match up indeterminate

meaning in the first place? What is the determinate fact that can decide between the pairing up of

indeterminate semiotic devices? TTS would be no good as an account of translation if it could not

indicate and explain how it is that we can err in translation. Indeed, erring in translation happens

frequently, I’m sorry to say. If my Ethics and the History of Indian Philosophy is correct, there are

virtually no good translations of Indian philosophy (except perhaps my in press translation of the

Yoga S tra through Penguin that avoids the errors of Orthodox Indology, recognizes “dharma” as a

key philosophical term and translates Patañjali’s self-consciously according to the text-type of

philosophy). But determining error in translation is not something that can be identified sub species

aeternitatis any more than translation can be determined apart from the relevant institutional

knowledge. Experts can err in translation, but for an error to be an error of translation, it must be

identifiable by reference to institutional expectations and knowledge. This will not be knowledge

154
commonly had. It involves an appreciation not only of the text-type, but cultural expertise in both the

target and source semiotic systems.

III.5. Text-Types in Comparison: Philosophy, Science, and Literature,

Broadly Conceived
The text-types of science and many aspects of the literary text-types have already made their

appearance. In the previous chapter, particularly under the discussion of the topic of Functionalism,

text-type features of literary text-types were noted. Poetry, novels and other literary text-types are

defined by their concern for the aesthetic appreciation of the reader: content is crafted to have an

effect. Thus, functionalist translation strategies are often very appropriate for literary text-types.

The text-type of science, I submit, is concerned with the articulation of descriptive theories in

light of empirical considerations, broadly conceived, that aim at being objectively persuasive. This is

what the translator of science must keep in mind when faced with scientific texts regardless of their

cultural provenance. The medical anthropologist and scholar of Ayurveda, for instance, are interested

in selecting from texts they study those features that constitute descriptive theories and the various

ways that count towards objective vindication. The various normative aspects of scientific texts, such

as methodology, procedure, and clinical norms, constitute subsidiary theories that are parasitic upon

descriptive theories, and the normative theories function in many instances as part of the empirical

confirmation broadly conceived for the descriptive theories: that, for instance, tribal healers have been

treating patients for generations according to a particularly normative conception of medical practice

counts as some evidence towards the descriptive success of the underlying theory. This does not

vindicate it against all possible competition, but it is a form of scientific confirmation. It is such

features that interest the translator of science.

Philosophers rarely talk about what philosophy is about. Gilles Deleuze and Félix Guattari

perhaps speak for most philosophers these days:

155
 The question what is philosophy? can perhaps be posed only late in life, with the arrival of old age
and the time for speaking concretely. In fact, the bibliography on the nature of philosophy is very
limited. It is a question posed in a moment of quiet restlessness, at midnight, when there is no longer
anything to ask. It was asked before; it was always being asked, but too indirectly or obliquely; the
question was too artificial, to abstract. Instead of being seized by it, those who asked the question set
it out and controlled it in passing. They were not sober enough. There was too much desire to do
philosophy to wonder what it was, except as a stylistic exercise. That point of non-style where one
can finally say, “What is it I have been doing all my life?” had not been reached. (Deleuze and
Guattari 1994 [1991], 1)

The bibliography on the nature of philosophy is indeed very small and the befuddlement that Deleuze

and Guattari speak to is a direct function of thinking about philosophy linguistically. On this account,

it can only be a type of activity or as a mode of expression that cannot claim any distinct content but

must settle upon a distinct style. However, determinate meaning is textual, and philosophy is

paradigmatically a type of text. The text-type of philosophy is never far from philosophers. We make

use of it all the time and it is particularly salient in our instruction to students. Most of what we teach

when we teach undergraduate students is text-type-theoretic: we teach them how to read and write

philosophy. From this place it is clear that philosophy is a theoretical endeavour, like science.

Philosophers don’t always have theories: they can often spend their lives criticizing theories. But the

discourse is structured by a dialectical interplay between theories and reasons. Unlike scientific texts,

philosophical theories are not descriptive.

This will certainly strike the naturalists as question begging. Naturalists want to conceive of

philosophy as fundamentally an analytic arm of science. But what is important to keep in mind is that

we are not trying to describe the one right philosophical theory, but rather philosophical theories. This

includes non-naturalist theories as well. For us to understand what is common to philosophical

theories, we need to look at the lowest common denominator. The lowest such commonality is the

universality and generality of philosophical theories. Here, the various anthologies, historical texts

that we read, and the diversity of theories discussed by philosophers are data to be taken into account.

156
 In determining the details of text-types, we are on the road to a certain sort of analytic

enquiry but it is not one that most philosophers are used to or are particularly trained in. In the case of

theoretical disciplines, it involves an institutional awareness of the diversity of views expressed

within an institutional discourse, the common terminological currency for such views, and how these

terms function to organize the discourse. The type of analytic activity we require thus is institutional,

historical, and above all, translational. Comparing text-types is useful in order for us to understand the

foci of institutional concern. Such concern might overlap at points, but the text-types that constitute

the underlying norms of various disciplines will form a unique constellation of concerns.

There are some broad stroke differences that are worth noting. Philosophy, for one, is

theoretical, but its primary means of articulating theories are not technical. Rather, philosophy tends

to take very ordinary terminology as its vehicle for theory articulation. Science, in contrast, is

notorious for its creation of technical terminology. This was noted by Kuhn, and translation theorists

concerned with translating science: “The terminology of science is part of its method, of its process of

discovery. It is an essential element in the conceptual framework of the science. [Thus] Faraday

created a new terminology for electricity because the old terminology implied ideas that he had

disproved” (Pinchuck 1977, 177; cf. Wright and Leland D. 2001 [1993]; Hann 1992). This is not to

say that science does not make use of ordinary terminology. Nor is it impossible for philosophers to

invent terminology. (This is generally frowned upon, but tolerated, it seems, if it can be understood

compositionally: “quasi-indexicality,” “text-type” semantics…) Indeed, the very notion of a “natural

kind” term just is a term that science has appropriated for theory articulation. (Thus, “water” is a

natural kind term, but “race” isn’t, or at least, not any more.) But concern and appropriation of

ordinary terminology for scientific purposes is not central to underwriting the type of discourse

science is. Science rather is primarily concerned with the articulation of descriptive theories in light

of empirical considerations, broadly conceived, that aim at being objectively persuasive., and if it

should turn out that ordinary language lacks terms that scientists feel comfortable using as designators

157
of theoretical entities, scientists will simply invent new terms. Science is by no means beholden to

ordinary language. A translator of science must thus understand that scientific theory articulation will

rely upon devices that often shift with theories.

Philosophy in contrast is a very different case. In philosophical texts, it is often words of very

ordinary provenance that play a pivotal role in theory articulation, and the words themselves become

the foci of philosophical disagreement. This is not something that is simply confirmed by knowledge

of the Western tradition of philosophy, but rather repeatedly affirmed when one studies philosophy

throughout the world. (Words like “li,” “tao”, “dharma,” “pram na” that are the stuff of Chinese and

Indian philosophy are not inventions of philosophers.) This fact is a consequence of the institutional

expectations of what philosophy is like regardless of cultural context, but it also points to something

profound about philosophy: it arises out of ordinary life and is in some important sense tied to it:

“good,” “right,” “wrong,” “real,” “meaning,” “knowledge,” “morality,” “ethics,” “beauty,” are the

terms that organize philosophical disagreement in English and these terms are incredibly domestic.

Even philosophical revisionists who want to radically reform language use can be understood as

articulating theories to underwrite such revisions that are in their canonical form articulable with such

ordinary words as “meaning” and “ethics.” Contrast the terms in the sciences that organize debate:

“physics,” “chemistry,” “biology,” “psychology,” “sociology,” “genetics”… These terms organize

debate not by articulating them, but by describing areas of scientific investigation and controversy.

The closest analogues to these terms in the case of philosophy often do double duty of both describing

and articulating debate across languages: e.g., “ethics” and “logic.” The important difference is that in

order to translate science, one must come to the proper categorization of the type of the theory being

articulated in a scientific text, even though the descriptor may be totally absent in the text. Whereas in

the case of philosophy, the key philosophical terms just are the discourse markers for philosophy.

They provide the evidence that what one is reading is philosophical and they also help determine the

158
description of the type of debate at play: they may even function as the descriptor of the debate in

some cases.

On one axis of comparison, then, philosophy does occupy a middle point between science

and literature. With science it shares a theoretical orientation. With literature it shares a strong ground

in domestic and common language. This is confirmed cross-culturally: one often finds works with

both scientific and philosophical merit (cf. Heisenberg 1962), or works with philosophical and literary

merit (such as the Indian epics),72 though it is difficult to find obvious cases of texts with scientific

and literary merit.

The ordinary provenance of key philosophical terms has contributed to a lot of confusion

against the backdrop of the linguistic turn. It has fuelled the absurd view that philosophy is concerned

with linguistic analysis. This is absurd because philosophy, particularly in the canon, comes to us in

translation, and thus if the issues discussed and debated in the texts have any relevance to our

philosophical perplexity now, it cannot be because these texts provide the right account of how to use

words in our language for they are not originally about our language. The linguistic paradigm

supports the illusion that philosophy is about linguistic analysis through its incredibly implausible

assumption that translation is a word for word exchange across languages based upon some pre-

existent linguistic synonymy, sub species aeternitatis, and absent all textual considerations.

Translation in general is not a word for word exchange across languages, and thus it is implausible

that in the case of the translation of philosophy it is a word for word exchange underwritten by

linguistic synonymy.

One might think that translating philosophy is merely a matter of preserving truth values of

sentences across languages and thus the matter of translating philosophy is purely linguistic. But this

is a very poorly informed account of what is involved in translating philosophy. First, it is often quite

unclear what the truth value of philosophical claims in philosophical texts are for it is a matter of

substantive disagreement and thus we need a way to understand their meaning without knowing

159
whether a philosophical claim is correct or not. Secondly, the translation of philosophy, like the

translation of all texts, will lead the translator to have to choose between differing significances in a

text (as we noted above , II.4.6 Indeterminacy of Translation And Difficult Texts pp.80-85). The way

that translators of philosophy overcome these obstacles is by understanding a philosophical text in

light of the theories or reasons articulated by quasi-indexical devices, which are usually represented

by key philosophical terms. This allows them to circumvent the substantive question of truth and

understand the claims of a text in light of the theories or reasons articulated in the text. Thus, the truth

values of claims of a philosophical text are assessed, but not from the view from nowhere, but by

taking the theory and reasons of the text as true for the purpose of translation. This approach also

allows the translator to be able to decide what significances of a ST should be preserved in a TT.

I have brought into our picture the matter of key philosophical terms, and I have suggested

that their conceptual anatomy is defined by QI. But QI in turn is simply a formalization of the text-

type of philosophy. We come to know the nature of the text-type of philosophy through our

institutional training. It is concerned with the application, investigation, debated, criticism and

articulation of universal and general theories (of various axiological differentias) articulated with

ordinary symbols, in light of reasons that are intended to be objectively persuasive.

Philosophers work according to this text-type definition all the time. They teach their students

to track theories articulated by philosophers distinguished by different axiological concerns, and they

impress upon their students that philosophy is not an opportunity to emote or describe one’s likes and

dislikes. Argument, and persuasion that attempts to be objective are the norms of criticism in

philosophy. The universality and generality of philosophical theories is quite frequently a manifest

fact to philosophers. A symptom of this aspect of philosophy is the great weight placed in

philosophical discussions on counterexamples and the importance of philosophical theories, to the

extent that they are put forward, to cover a wide range of cases, including cases that have yet to be

encountered.

160
 This is a synopsis that I shall return to frequently in the remainder of this dissertation. QI is a

formalization of the text-type of philosophy and it is its text-type feature. QI is the general anatomy of

the philosophical concept. But it comes in different flavours, distinguished by such matters as

axiological differentia, valence, direction of fit and extra-theoretic constraint (that distinguish thick

concepts). Each such concept is thus an avatar of philosophy, instantiating the type, but with its own

personality. What distinguishes QI from the domestic accounts of philosophical concepts is that it is

not an attempt to track the meaning of philosophically important words in a language, but rather their

function within the text-type of philosophy. QI thus understands philosophical concepts as devices of

translation and for this reason it can overcome the seeming obstacle of cultural diversity to

philosophical knowledge—both mundane and profound.

III.6. Finding our Feet

In the previous chapter I argued that real translation is determinate at the textual level, with

the mediation of a text-type with institutional support. This view contrasts sharply with the approach

to translation taken by many in the linguistic paradigm. According to the linguistic account of

translation, translation is a pairing up of words and sentences across languages on the basis of a

putative cross-linguistic synonymy. From the perspective of translation theory, one major problem

with the linguistic account of translation is that it fails to take into account text-type-theoretic

considerations that mediate translation. Translating a poem involves a very different constellation of

concerns than translating a text of engineering, and yet again translating philosophy must also be

sensitive to matters of philosophical importance that are its peculiar concern. Words and sentences,

moreover, are steeped in historical and cultural contingencies. Something as simple as syntactic

differences across languages makes the linguistic account of translation implausible. We require text-

types to mediate the linguistic differences across languages, but also the various institutional concerns

we have in translation. Once text-types come clearly into view, it is implausible to deny their role in

translation.

161
 However, recognizing the text-type-theoretic nature of translation has implications for a

theory of meaning. Translation makes little sense apart from the notion that it preserves meaning. To

hold that translation can be successful at the textual level but yet fail at the semantic level is a virtual

contradiction. For then it would seem that we could succeed in every institutionally relevant way in

producing a translation and yet we will have in the form of a TT a semantic fake that cannot be

attributed to the author of the ST. If this were true, all human knowledge that operates at an

international and cross-linguistic level would be impossible. The modern academy would have to

grind to a halt for fear of fraud: no translation would be accurate. But this tells poorly on the

linguistic conception of meaning, for it shows that it is out of touch with what we require to

understand and underwrite our best epistemic practices.

One might think that the way to solve this problem is to argue that while poetry and

philosophy might require text-types for translation, science and other important disciplines do not

require text-types. Rather, here, translation is a matter of paring up words and sentences across

languages on the basis of their literal meaning, or, in terms of their natural kind terms. Of course, by

now, we know that that the very identification of science as requiring its own mode of translation is a

text-type-theoretic decision. Thus, we do require text-types to translate scientific texts as well. But

literal meaning and the metaphysics of natural kind terms are not great ways to vindicate the

objectivity of meaning, for systematic aspects of a language or linguistic devices defined by some

baptismal circumstance are incurably local and linguistically relative affairs. Not only is moral

knowledge that is international, transcultural and trans-linguistic not possible on this score, neither is

scientific knowledge that transcends cultural and linguistic boundaries. If the translation of science

had to rely upon literal meaning understood as the basic or systematic aspect of an expression in a

language to underwrite the objectivity of scientific discourse, or if it had to rely upon cross-linguistic

synonymy of natural kind terms, there could be no way to determinately translate a scientific text.

The Kuhnian argument about the incommensurability and untranslatability of scientific theories

162
simply carries forth the irrationality of the linguistic turn to its logical conclusion. This conclusion is a

simple function of the notion that languages are the primary semantic phenomena characterized by

determinate meanings.

Thus, the real threat to the objectivity of knowledge and meaning is not TTS and its

implication that linguistic meaning is indeterminate, but rather the linguistic turn and the notion that

languages are determinately meaningful. The notion that languages are determinately meaningful

makes change, progress, and determinate intercultural communication impossible for it attempts to

rigidify (in the broad sense) the very means we have of being receptive to the novel and new.

TTS shows us a way out of the quagmire of the linguistic turn and it shows us this way in all

the right ways.

(1) TTS shows how translation has been and can be determinate. It is determinate on the

basis of text-type institutions.

(2) Text-type institutions are distinct from cultures and languages at large because text-types

are not dependent upon systematic or basic features of languages, thus enabling them to

have a semantic identity that is distinct from that of a culture or language.

(3) TTS affirms how languages and cultures can be distinct and radically different along

many axes of comparison, and the fitting subject of social scientific and historical

research.

(4) TTS underwrites the culture transcendence of knowledge via determinate translation of

texts.

(5) TTS underwrites the objectivity of knowledge by recognizing that discipline relative

knowledge is subject to its own criteria of assessment, and not the wider whims and

processes of a culture.

163
 (6) TTS shows that despite the indeterminate nature of linguistic meaning, language has

some semantic stability that is in part a function of the normative point of reference that

texts of a definite type play in a society.

The difficulties of the linguistic paradigm show that the ultimate test of the objectivity and

determinacy of meaning is translation at the textual level: textual meaning is the only type of

determinate meaning. Sentences, expressions, and symbols, only have a determinate meaning in the

cotext of a text of a definite type. Some concepts have a place of privilege for they are the defining

features of a text-type. These too are determinately meaningful, but they too are meaningful by

reference to there text-type-theoretic function. Language, and any semiotic system, no matter how

systematic, is best thought of as indeterminately meaningful.

The misfortune of the linguistic turn is that, in order to attempt to account for some

semblance of objectivity, it must deny or ignore cultural and philosophical diversity. The linguistic

account of translation is the epitome of this denial of diversity: it insists that in order for meaning to

traverse contexts, it must be completely reducible to components that are, none of them, semantically

novel to the target community. Word for word, sentence for sentence, translation is only intelligible

on this account if the artefacts of culture, namely languages and their components, are semantically

equivalent in and of themselves.

If languages are very different, and if determinate meaning is a feature of a language, then no

amount of linguistic analysis, or scientific investigation of the referents of our languages, will yield

objective knowledge for the results of such investigations will always be linguistically relative.

Now it seems that what we have been making a case for is nothing but cultural and moral

Relativism. The Relativist might feel vindicated: see, it is true, moral diversity is roughly continuous

with cultural diversity and there is no common, objective point from which we can decide questions

of ethics.

164
 The problem with the moral relativist is that she assumes that we have some way of

objectively identifying ethical claims across cultures and languages and that it is these that, once

having recognized them, we understand to be relatively true. But she assumes this without making

any effort to show how we can determinately translate such claims. Indeed, prominent relativists,

such as Kuhn, and Gilbert Harman73, have in their own way argued against the determinacy of

translation. Other moral relativists, such as David Wong, have done nothing to overcome the

indeterminacy of translation.74 Relativism only seems plausible if we do not have a way to

determinately translate semantic artefacts across cultures, but this is also its downfall. In the absence

of a way to determinately translate texts, we can at best project our definitions and criteria upon the

world, and perhaps relativism will seem like the truth from this perspective, but it shouldn’t surprise

us, for we embark upon this methodology from a relativistic starting point. If translation is

indeterminate, there is nothing objective about the relativist’s argument for the relativist has not even

been able to determinately identify competing claims that could be relatively true. Relativism is

simply one of the muddled consequences of the linguistic paradigm.

What we require thus is an account of determinate meaning that is not premised on the denial

of cultural diversity, but yet provides the basis for objective, knowledge. Switching our attention from

linguistic meaning to textual meaning is the move that allows us to save objectivity in the face of

linguistic diversity. The move to textual meaning allows for text-type institutions to mediate cultural

diversity via their specific text-type-theoretic concerns, and this text-type in turn provides a cross-

linguistic, cross-cultural platform and criterion to judge the epistemic claims abstracted from texts of

its type.

TTS understands the paradigm cases of meaning to be texts of a definite type, and moral

concepts to be features of such texts, not primarily the properties of languages abstractly conceived.

QI just is a formalization of all philosophically important concepts, including normative and

evaluative concepts, as features of a specific type of text. These concepts, according to TTS, are

165
devices of textual translation. As we shall see, QI is the only plausible contender that can explain how

to translate normative and philosophical discourse, for only it is based upon TTS and only it responds

to the peculiar challenges of translating normative discourse across cultures and languages.

In the interest of space I skip a review of moral semantics since G.E. Moore that could come

after this chapter. In its place I will note that the two main positions in the literature today are Non-

Analytic Naturalism, and Expressivism.

In the next two chapters I shall address the question of how the main competitors in the

literature to QI, namely Non-Analytic Naturalism and Expressivism, fare when presented with the

challenge of translating normative discourse. In the next chapter I shall focus on proposals for the

translation of moral semantics that presume that moral terms across languages must have the same

linguistic meaning in order for them to be translatable. The most prominent of such views is Non-

Analytic Naturalism. In other words, I shall examine theories of moral semantics that assume the

linguistic account of translation. As we know by now, translation cannot be underwritten by cross-

linguistic synonymy for such cross-linguistic synonymy is not forthcoming. If we were to understand

normative discourse translation as it has been discussed in the recent metaethics literature, we would

have to subject normative discourse translation to constraints that would make even the translation of

scientific discourse impossible. After dealing with the obviously linguistic accounts of moral and

normative semantics in the literature, I move to Expressivism, which may be divorced from its

linguistic underpinnings and reconceived as a possible candidate for the text-type of normative

discourse. I show in chapter 5 that Expressivism fails because it attempts to understand normative

discourse as though it were not a species of philosophical discourse. This proves that the only way to

translate normative discourse is as a species of philosophical discourse. QI is the only candidate in the

running that can accomplish this. In chapter 6 I explicate QI in detail and contrast it with linguistic

accounts of key philosophical terms that also attempt to understand their structure after the fashion of

indexicals. In chapter 7 I return to topics touched upon here, and defend TTS and QI against

166
misinformed charges that it is a version of Contextualism. I show in contrast that only a textual

account of meaning can show how meaning is insensitive to context. Having defended QI as the only

way to translate normative discourse, we move on to the topic of philosophical relativism in the

conclusion. As Kuhnian relativism in science is dispelled by TTS as a confused implication of the

linguistic turn, so philosophical relativism can be dispelled by TTS, and QI in particular. QI, I shall

argue, is the Archimedean platform from which we can decide philosophical questions. We gain an

insight into this platform not by ignoring radical cultural and philosophical diversity, but by affirming

it and understanding translation and meaning in a manner that accommodates this brute fact of cross-
75
cultural research.

167
IV. Ethical Naturalism and Linguistic Accounts of Moral

Semantics

“The analytic lesson here for linguistics I will term a linguistic uncertainty principle in
keeping with Whorf's original (and, in retrospect, unfortunate) metaphor: those who
would think that native linguists can directly penetrate to the linguistic coding of
referential ‘reality out there’ by examining their own propositional system—no matter
how ‘deeply’—or by examining others’ with crude approximation—translations of
propositional content[—] unrecognizably distort the object of investigation in the
process… We should ask, in particular, how the seemingly reflective and creative or
‘performative’ functions of language (or, rather, of language use) relate to native
awareness and native ideology. Can we generalize Whorf’s penetrating insights from
the plane of reference to the whole of language function? I think we can discern the
same disjunction between ideology and structure, one, moreover, which assimilates
function to reference and thereby affects the strategy of language use.”— Michael
Silverstein, “Language Structure and Linguistic Ideology”

In chapters 2 and 3, I argued for TTS. According to TTS, the notion that determinate meaning

is best thought of as a feature of languages or bits of languages is the single cause of our difficulties

in understanding translation and cross-cultural, and cross-contextual communication. In the place of

such accounts of semantic and translational determinacy that I identified as “linguistic” in chapter 3,

TTS holds that texts and their types are the paradigm cases of determinacy in semantics. TTS sets the

stage for the picture of normative semantics that we must affirm if we wish to translate normative

discourse. In the next chapter I will turn to the problems with Expressivism and semantic options that

combine naturalistic and expressivist elements. In this chapter I will focus on recent naturalist and

linguistic accounts. Specifically, I will argue that recent naturalist accounts of moral semantics as

well as linguistic accounts cannot help us translate and understand cross-linguistic and cross-cultural

normative disagreements. I hope to show in this chapter that problems that beset classic naturalist

options in moral semantics beset any linguistic account.

The most famous Non-Analytic Naturalist account present in the recent literature is

associated with Richard Boyd. On his account moral semantics is not determined by conceptual

168
analysis, but by a type of empirical enquiry, continuous with scientific enquiry. Moreover, the

phenomena that constitute the meanings of moral concepts are natural on this account, and thus they

constitute the legitimate object of scientific enquiry in their own right. The view in question contrasts

with Moore’s non-naturalism, which defers to non-natural metaphysical properties as the referents of

moral language. The account also contrasts with earlier naturalist accounts that Moore had in his

sights.76 I set out Richard Boyd’s account along with another similar account provided by John

McDowell in the first section. In the second section I review a now familiar argument against Non-

Analytic Naturalist moral semantics, namely Terrance Horgan and Mark Timmons’ Moral Twin Earth

argument. In the third section I take stock of the effectiveness of the thought experiment. The thought

experiment, along with the literature that it has generated, labours under the mistaken assumption that

translational and semantic determinacy can be clinched by providing an account of the meaning of

terms in a language. I have called this a “linguistic” approach to meaning and translation. If the

arguments presented in chapters 2 and 3 are correct, this is a great mistake.

In the fourth section, I argue that any linguistic approach to meaning can be used to generate

its own problems of translation. Specifically, I argue that if the various options in the literature are

correct that moral semantics is really about providing an account of the meaning of certain types of

words in a language, it will be possible for us to have cases where several different languages are

governed by distinct semantics but yet speakers of these languages can be understood as morally

disagreeing with each other. QI can show us how to translate moral discourse in such cases in such a

manner that all parties will come to understand how they morally disagree with their counterparts,

even though they do not share the exact same “semantics” for their moral vocabulary. This aspect of

the thought experiment is very realistic for the notion that languages from radically different cultures

will have words with the same meanings, particularly philosophically interesting terms that have to do

with values, is implausible.

In the fifth section I focus on Richard Boyd’s referential account of moral semantics and

169
argue that it fails us in cross-cultural research.

In the sixth section, I consider afresh the claim that we can do semantics for a language, such

that we could correctly specify the meaning of a word in a language—a view shared by the

naturalistic options in the literature. I argue that this is the wrong way to think about linguistic

meaning and language. Here I consider the peculiar difficulties with determining moral semantics by

reference to language use. This confirms the indispensability of philosophy as a type of text to

mediate the translation of normative discourse.

IV.1. Non-Analytic Moral Naturalism

One example of Non-Analytic Moral Naturalism in the recent literature takes its inspiration

from recent work on natural kinds semantics.

A challenge for natural kind semantics is to explain how some terms are coreferential, even

though their popular understandings present us with divergent pictures. Consider the case of the terms

“water” and “H2O.” “Water” is often taken to mean “clear liquid, that is crucial to life, descends from

clouds, found in streams, rivers, lakes and oceans” while “H2O” has the technical meaning “a

molecule formed only of two hydrogen atoms and one oxygen atom.” Yet, it seems that all things that

are water are also H2O. We came to this conviction via empirical investigation, and not conceptual

analysis. Yet, there is a kind of necessity—a metaphysical necessity—surrounding the identity, or

coreferentiality, of “water” and “H2O.” The necessity is explained by the fact that both “water” and

“H2O” are rigid designators77 for the same item. On this view, there is no possible world where both

“water” and “H2O” are not coextensive.

Several philosophers are associated with this model of natural kinds semantics, but the two

philosophers who are credited with developing this theory are Saul Kripke (1980 [1972]) and Hilary

Putnam (Putnam 1975). The view often goes under the description of the “causal theory of reference”,

where the causes in question are the historical links between word usage and an initial baptismal

ceremony.

170
 One version of Non-Analytic Moral Naturalism seizes on Kripke-Putnam natural kinds

semantics as a way of explaining how it is that moral terms are coreferential with natural terms. The

motivation for the naturalist is plain. Naturalism seeks to explain all things in terms of the ontology

and methodology of the sciences. Kripke-Putnam natural kinds semantics is naturalistic in this very

way: it presumes that it is through science that we come to understand that terms rigidly designate the

same item. Moreover, Kripke-Putnam natural kind semantics explains how there can be a certain kind

of necessity—a metaphysical necessity—in the coreferentiality of distinct concepts, without

understanding that necessity in terms of analytic or semantic connections. If one could explain the

coreferentiality of moral terms and scientific, descriptive terms, then the Naturalist’s task of showing

that moral terms refer only to natural properties would be complete. Moral concepts would thus be

like our ordinary, pre-scientific concepts of “water,” and the knowledge that moral concepts refer to

certain natural properties (be they simple, complex, or functional) would be knowledge that we come

to appreciate, consequent of scientific investigation, just as we came to understand that “water” refers

to stuff that has the structure of H2O, consequent of scientific investigation. Because the necessity

involved in the coreferentiality of moral terms and scientific concepts can be known without appeals

to the analysis of either moral concepts or scientific concepts, one could avoid the types of problems

that Moore’s Open Question Argument raises. On the Non-Analytic Moral Naturalist Model, those

who have mastery of the concept GOOD are not expected to have any knowledge that it refers,

necessarily, to certain natural properties, like pleasure.

Richard Boyd endorses this project when he suggests that moral terms can be understood as

causally regulated by a kind that over time brings it about that the use of moral terms will be true of

the causally regulatory kind (Boyd 1988, 195). Specifically, the kind in question is what he calls a

“homeostatic cluster” property pertaining to human well being and flourishing. A type of

consequentialist moral theory follows from this for Boyd (what he calls “Homeostatic

Consequentialism”), where right action is judged in relation to this cluster property. On his account,

171
our moral judgments are true when we come to refer to the causal mechanism with our moral

terminology. Science, on his account, can thus expedite moral truth by elucidating the causal

mechanisms underlying our moral language usage.78

A second version of Non-Analytic Moral Naturalism takes its cue from Aristotle and

Wittgenstein. Aristotle in the Nicomachean Ethics argues that moral knowledge can only be had by

one who has been raised properly. It is a special type of social knowledge that one gains not from

abstract reflection but through the proper sort of socialization. Reflection on the social norms of the

polis is what ethics as a philosophical activity amounts to. The later Wittgenstein held that semantic

knowledge is a type of social knowledge. Specifically, on his account, to understand the meaning of

an expression is to be able to employ it meaningfully in a form of life. In the work of John McDowell,

these two theses come together. These accounts of moral meaning, while arguably different from the

account provided by Boyd, can be understood as versions of naturalism because they understand

moral meaning and knowledge as what arises through the natural and healthy maturation of a person

within a society. Moral knowledge and meaning on these accounts is a species of cultural knowledge

and meaning. And as these accounts do not imply that moral meaning is a matter of mere definition,

but rather a type of move within a practice, they too can be understood as a species of Non-Analytic

Naturalism—even in the case of a purely Wittgensteinien account of moral meaning, divorced from

the perceptual emphasis of McDowell’s account. Such accounts are naturalistic, for they eschew non-

natural metaphysical explanations, and they are non-analytic for they reject the role of definitions as

determinative in semantic knowledge.

In response to John Mackie’s (Mackie 1977) Error Theory (according to which, moral

judgments are genuine descriptions of the universe, but all of them false for there are no moral

properties that are part of the natural world) McDowell argues that values are best understood as

secondary properties of objects. According to McDowell, “A secondary quality is a property the

ascription of which to an object is not adequately understood except as true, if it is true, in virtue of

172
the object's disposition to present a certain sort of perceptual appearance: specifically, an appearance

characterizable by using a word for the property itself to say how the object perceptually appears”

(McDowell 1998c, 133; cf., Wiggins 1991; Johnston 1989). Aristotle and Wittgenstein enter

McDowell’s picture for, on his account, we cannot come to perceive values without the proper

upbringing. This is because there may be nothing at the level of primary qualities (the subvening

level) that uniquely correlate with the secondary properties (the supervening level) (McDowell 1981,

144-145). To perceive value is to be made sensitive to certain features of the natural world through

the mediation of moral concepts that one gains from one’s society. 79

IV.2. Moral Twin Earth

In a series of papers, Terence Horgan and Mark Timmons (1991, 1992a, 1992b, 1996, 2000a)

ask us to consider a variation on Putnam’s Twin Earth Thought Experiment.

Putnam asks us to imagine a Twin Earth, where the human-seeming inhabitants speak a

language that is homophonic with ours, and that appears to function just as ours does. In Twin Earth

English, there is a term “water” that appears to stand for a substance that looks, acts, and tastes just

like what we call “water” on Earth, except further scientific investigation reveals that the substance in

question on Twin Earth has a molecular structure of XYX and not H2O. Putnam asks us to consider

what our reaction is: do we consider that the substance that the Twin Earthlings call “water” is, on our

understanding, in fact water, or is their term “water” really a term standing for a very different

concept? Putnam argues that Twin Earth “water” cannot be water to us, for, on our understanding,

there is no possible world where water is not composed of H2O; what it is for something to be water,

on our account, is for it to be composed of just this molecule. Water for us, Putnam concludes (using

Kripke’s terminology), is a rigid designator, which designates the same thing in all possible worlds:

H2O (Putnam 1975, 139-149).

Horgan and Timmons ask us to consider a Moral Twin Earth. On Moral Twin Earth, the

inhabitants (who seem just like us) speak a language that is homophonic and interintelligible with

173
ours. Moreover, we find that they use the same phonemes that we use for evaluative terms (“good,”

“moral,” etc.) in a way that is remarkably similar to our use. Our initial reaction is that we find Moral

Twin Earth discourse that uses our phonemes for moral concepts mutually intelligible. Yet, closer

inspection reveals that the predominant sensibilities on Moral Twin Earth are Deontological, while

the dominant sensibilities on our Earth are Consequentialist. Further, when we engage in the

naturalist’s project of determining the natural properties that govern Moral Twin Earth “moral”

discourse, we find that it is causally governed by a different natural property than what governs our

use of moral vocabulary (Horgan and Timmons 1991, 458-460, 1992a, 247-250, 1992b, 164-166).

After setting out the experiment, Horgan and Timmons ask:

Given all these assumptions and stipulations about Earth and Moral Twin Earth, what is the
appropriate way to describe the differences between moral and twin-moral uses of ‘good’ and
‘right’? Two hermeneutic options are available. On the one hand, we could say that the differences
are analogous to those between Earth and Twin Earth in Putnam’s original example, to wit: the moral
terms used by Earthlings rigidly designate the natural properties that causally regulate their use on
Earth, whereas the twin-moral terms used by Twin Earthlings rigidly designate the distinct natural
properties that causally regulate their use on Twin Earth; hence, moral and twin-moral terms differ in
meaning, and are not intertranslatable. On the other hand, we could say instead that moral and twin-
moral terms do not differ in meaning or reference, and hence that any apparent moral disagreements
that might arise between Earthlings and Twin Earthlings would be genuine disagreements—i.e.,
disagreements in moral belief and in normative moral theory, rather than disagreements in meaning.
(Horgan and Timmons 1991, 460)

Horgan and Timmons argue that our intuitions lead us to abandon the view that moral language

rigidly designates natural objects, and instead embrace the view that Earthlings and Twin Earthlings

have a genuine disagreement over substantive moral issues. Specifically, Horgan and Timmons wish

to lobby for an internalist account of moral semantics, where moral meaning is connected in an

essential way with the outlook of speakers.

The Moral Twin Earth Argument has a wider scope than Horgan and Timmons envisaged.

Sean Holland has argued that it can be directed at “dispositional theories of value” (Holland 2001).

174
According to Holland, “[d]ispositional theories of value typically begin by exploiting an analogy with

secondary qualities like colour … A dispositional analysis of a colour term, say ‘red,’ specifies the

meaning of ‘red’ by appeal to the normal or ideal dispositions of normal or ideal subjects to respond

in a certain way (such as judging that an object is red)” (Holland 2001, 181). Dispositional theories of

values hold that value terms refer to events that we are disposed to evaluate, under ideal conditions.

Evaluations, under this analysis, can be thought to be caused by external, natural properties of objects

or events. This is a view put forward by John McDowell (1998c), David Wiggins (1991), and Mark

Johnston (Johnston 1989, 162-169). According to Holland, all of these views are vulnerable to the

Moral Twin Earth argument.

From what has been argued so far, it would seem that dispositional theories of value would be

in trouble when they come up against the “moral” discourse of Moral Twin Earthlings, for

Dispositional Theories of Value also hold that moral vocabulary is causally regulated by certain

natural properties. The real trouble comes into the picture, according to Holland, when dispositional

analyses of colour (which is the model on which the analysis of value is presented) imply that

secondary quality terms are rigid designators of objects that are specifiable by their primary qualities.

These accounts need to stipulate that value terms are rigid designators of natural objects, or else it

would be possible for natural objects to remain unaltered but for their evaluative status to change,

owing to such factors as our preferences; this is a kind of anti-objectivism, and perhaps even

subjectivism, that Naturalist versions of Moral Realism seek to rule out (Holland 2001, 189).

One might argue in response that McDowell’s account cannot so easily be characterized as a

variant of the typical dispositional theory of value, for McDowell argues that moral properties as

secondary properties may not reflect any regularity or pattern at the primary, natural property level

(McDowell 1981, 144-15). McDowell also appears to want to maintain that even in ideal

circumstances, there can be perceptual error of moral properties. However, McDowell’s view is that

moral knowledge is about correctly tracking natural properties in a manner that can only be derived

175
from the right type of socialization and training. Thus, indeed, McDowell can hold that secondary

properties do not reduce exactly to specific primary qualities, but this is because our socialization and

linguistic training gives us an ability to know which configuration of primary qualities correspond to

which configuration of moral secondary properties. The social intermediary is essential on

McDowell’s account, but it is by virtue of it that we can be said to have perceptions as opposed to

projections. Here, McDowell’s view is hardly different from Boyd’s. To get reference of the ground,

Boyd holds that one requires a significant amount of social coordination in place (cf. Boyd 1988, 195)

(to be discussed at length below IV.5, The Explanatory Failure of Ethical Naturalism pp.193-212).

Moreover, on Boyd’s account, the natural qualities that are tracked are not homogenous, but rather a

cluster that is significant in light of the needs of humans. The main difference between Boyd’s and

McDowell’s accounts is that Boyd does not stress perception in his account of how content is tracked

while McDowell stresses perception in his account. But the difference matters little, for the

properties that we perceive, on McDowell’s account, are powers of the objects to produce perceptions

in us—powers that are just as much part of the objects that we perceive as their primary qualities

(McDowell 1998c, 133-6). Thus, our perceptual abilities are intimately tied to both the objects of our

environment and our training. Twin Earthlings, in contrast, have a distinct social practice developed

in light of distinct objects that we have no training in being sensitive to (or perhaps, a different

sensitivity to, in light of our training).

Just as in Boyd’s case, McDowell too requires the mediation of a linguistic practice to deliver

reference: it just so happens that McDowell emphasizes the role of the linguistic practice in mediating

reference more than Boyd does. All the same, if the content of moral concepts is constituted by one’s

social practice, then McDowell too will have trouble explaining how speakers of English and

Twinglish are speaking about the same things if their concepts are derived from distinct linguistic

practices, filled out by distinct natural properties. For, returning to the example provided by Horgan

and Timmons, what accounts for the different views Earthlings and Twin Earthlings take on moral

176
issues (one group being more consequentialist and the other more deontological) is that the respective

parties have judgments that are a function of their unique cultural histories. Thus, when Earthlings

look at Twin Earth’s practices and disapprove, they perceive values in objects that Twin Earthlings do

not, and when Twin Earthlings assess Earth’s practices, they see values in objects that Earthlings do

not. Their fully objective judgments literally track different features, and thus there is nothing for

them to disagree about because they are quite literally meaning different things by their value

concepts as a result of their distinct upbringing and native environment. McDowell would have to

conclude that there is no real disagreement: only apparent disagreement. But he is inclined to put a
80
brave face on such paradoxical implications of his view.

Should we be surprised at this? Not at all. The trouble that Boyd and McDowell run into with

the Moral Twin Earth thought experiment is the trouble that all linguistic accounts of meaning have in

accounting for translation. If meaning is filled out by a linguistic practice, and if practices differ at all,

then translation is rendered puzzling, if not indeterminate. Moral Twin Earth is simply the science

fiction example of the trouble that cultural diversity causes for the linguistic turn.

IV.3. Why the Moral Twin Earth Argument Fails

QI understands moral semantics in terms of devices of translations of certain types of texts,

while Non-Analytic Naturalism in its various ways does not. Thus, I am in sympathy with some

conclusions that Horgan and Timmons wish to derive from the Moral Twin Earth thought experiment,

namely that Non-Analytic Naturalism, and indeed any account of moral semantics that ties moral

meaning to certain properties or items (whether they be natural or non-natural), is the wrong

canonical account of moral meaning, that it cannot explain how to translate moral discourse and that

it fails to explain cross-contextual moral disagreement. However, Horgan and Timmons have not

shown any of these conclusions to be true by means of their thought experiment.

Their Moral Twin Earth thought experiment puts the following dilemma to the naturalist:

either (a) Non-Analytic Naturalism is correct and moral claims across Earth and Twin Earth are not

177
intertranslatable, or (b) an alternative, internalist account of the semantics of moral language is

correct and moral claims articulated in English and Twinglish are intertranslatable. The trouble with

this disjunction is the hidden assumption, namely that translation is about finding words across

languages with the same meaning. Certainly, if the naturalist account of moral semantics as Boyd sets

it out is correct, then Earthlings and Twin Earthlings are unlikely to share the same referents in their

“moral” vocabulary, and ex hypothesi, the meaning of their “moral” vocabulary will not be identical.

Horgan and Timmons take the synonymy of candidate moral vocabulary across languages as

justifying their intertranslatability. Boyd too holds such a position. On his account, people assume

something like his anthropocentric account in their translation of moral discourse (cf. Boyd 1988,

210). So indeed, what we have here is a case where both the critics (Horgan and Timmons) and the

criticized (Boyd) share what we might call a linguistic assumption about translation. The linguistic

assumption is that translation is a matching up of word, for word, across languages, on the bases of

their shared meaning.

We know from chapter 2 and 3 that the linguistic account of translation is unworkable.

Translation if it is at all possible must be possible in the face of cultural differences and the

asymmetry of languages. The major failure of the linguistic paradigm is in conceptualizing translation

as a word for word, sentence for sentence, exchange across languages. Translators have always

translated texts, not languages.

There are many reasons for the non-linguistic nature of translation.

For starters, one cannot determinately translate a language. The very proposal is absurd, and

overlooks the endless productivity of languages, their distinct histories, their holistic, self-referential

nature, the many ways in which languages are bound up with their cultural peculiarities, the

consequent polysemy of words in a language, and their distinct grammatical and stylistic

requirements. The very expectation that it is languages that are to be translated is what leads to the

classical problems in translation theory noted in chapter 2—the most famous of these theses being

178
Quine’s indeterminacy of translation thesis and Derrida’s lost in translation thesis. The problem that

Quine’s thought experiment picks out is not simply academic. Translators often struggle with this

very problem: words across languages cannot ever be said to have the same meaning. There are

always differences, and viewed from the perspective of language as such, it seems as though there is

no principled way to decide between alternative translations. The solution to this problem is to bring

into the picture of translation text-types, which provide a nonlinguistic frame against which to assess

the adequacy of TL resources. The goal is thus converted from attempting to find expressions that are

synonymous across languages sub specie aeternitatis to the problem of identifying the text-type

relative features that must be preserved in translation, and constructing a TT that fulfils this purpose.

The textual turn allows us to accommodate not only differences in the semantic profile of vocabulary

in a language, but also differences in syntax, which makes a word-for-word exchange across

languages unworkable as a translation methodology. It also allows us to introduce new expressions or

use old expressions in a new way, if such innovations are mandated by the textual project of

translation. Certainly, there will be some requirement to attempt to find synonyms across languages,

but the synonymy relationship is judged relative to textual objectives, and not by reference to purely

linguistic concerns.

In light of this argument, all of the literature to date on the Moral Twin Earth problem

(present work excepted) gets off on the wrong foot. Adjudicating between competing efforts to

overcome the Moral Twin Earth problem, given its stipulation that a successful account will explain

translation by reference to cross-linguistic synonymy of moral vocabulary, is like adjudicating

between competing efforts to square the circle, or more fittingly, trying to determine how many

angels can dance on the point of a needle. The trouble is that the Moral Twin Earth scenario raises the

problem of translation in the face of cultural diversity, without any of the respondents to the problem

understanding that the expectation that translation be underwritten by cross-linguistic synonymy is

itself a failure to recognize cultural diversity.

179
 IV.4. Quality Science Fiction for the Analytic Philosopher
In this section I want to continue with a very dubious methodology, namely, relying upon

thought experiments—though, not because I think they are generally a good way to do philosophy.

Indeed, I think that if philosophers actually did more work in cross-cultural research instead of simply

thinking about it, their views on moral semantics would come to seem to them extremely naïve.

However, concocting imaginary circumstances to think about the adequacy of accounts of moral

semantics seems to be the norm in the recent literature. As I wish to talk to this literature, it is

necessary to get dirty in something that is really quite silly. The challenge of course is to construct a

scenario that actually veers away from reality in a useful manner but avoids building in constraints on

solutions that are independently untenable. The Moral Twin Earth thought experiment succeeds on

the first score, but fails on the second.

It succeeds on the first score because it builds into the scenario the assumption that the two

parties from the different planets will speak an interintelligible language. This is patently implausible,

but it allows us to contemplate circumstances of cultural exchange without our having to really know

more than one language.

It fails on the second score because it relies upon an untenable conception of meaning and

translation, according to which the determinate meaning of a word is linguistic, and not textual, and

translation is a matter of matching up words across languages that have the same semantics. This is a

mistake.

The following thought experiment will be unrealistic in the first manner, but it will endeavour

to show the problems with the second nonsensical aspect of the Moral Twin Earth experiment, shared

with the options on moral semantics in the literature. This common assumption is that translation of

moral discourse is only possible if speakers of different languages share the same semantics for their

moral vocabulary. Call this thought experiment the Moral Interplanetary System.

Consider two options in the literature that are supposed to overcome the Moral Twin Earth

180
problem but are yet within the general naturalistic fold. Both options are in the naturalistic fold

because they do not recognize any properties that are not naturalistic. However, they are supposed to

avoid the troubles that Boyd’s account generates by grounding moral semantics in some non-

referential foundation (for instance, in the epistemic role of moral concepts, or in the conceptual role

of moral concepts). The common assumption is that it’s the referential underpinnings of moral

meaning on Boyd’s account that gets him into trouble in the Moral Twin Earth scenario. No one

(before us) seems to have noticed that tying meaning to the reference of terms in a language is simply

a special case of tying meaning to linguistic expressions in a language and that the problems

encountered in the Twin Earth thought experiment could be visited upon any linguistic account of

moral semantics.

One option is presented by Mark van Roojen. On his account, the problem with Boyd’s

naturalism is that it causally ties the semantics of normative terms to reference indexed to their home

planets, thus making it impossible to understand how the inhabitants of Earth and Twin Earth can be

speaking to each other, and not past each other, when they have a normative disagreement. His

solution is to understand “right” as a paradigm moral term, and to argue that its meaning is epistemic,

not rigidly designating: “right” tracks what it objectively makes sense to do, on his account

(van Roojen 2006). Then there is a similar, but distinct option presented by Ralph Wedgwood,

according to which the meaning of moral terminology is given by their role in practical reasoning,

and the paradigm, thinnest of such concepts can be captured by the notion scheme “x is (all things

considered) a better thing for z to do at time t than y.” This term he calls “B” for best (Wedgwood

2001). Wedgwood’s account is quite explicitly act-oriented. van Roojen formulates his account in

response to Horgan and Timmons’ thought experiment. Wedgwood thinks that his approach can

avoid Horgan and Timmons’ criticism of naturalism (Wedgwood 2006). van Roojen’s account is

supposed to ground moral semantics in an epistemic foundation, while Wedgwood attempts to

provide a “conceptual role” foundation (they are indeed similar—Wedgwood’s account specifies

181
times for actions and is thus a clearly act-oriented account, while van Roojen’s account is not act-

oriented and attempts to track a property or feature of the right thing to do, possibly even in the long

run). The assumption is that by ridding moral semantics of a referential grounding, we facilitate

translation.

Imagine one planet, Planet van Roojen, and another planet, Planet Wedgwood. The beings of

these two planets are anatomically identical, and they speak a language that is homophonic and

apparently interintelligible except, it is known for a fact, that their moral terminology is governed by

distinct semantics. On Planet van Roojen, it is van Roojen’s account of right, namely, what it makes

sense to do (period), on Planet Wedgwood, it is Wedgwood’s conceptual role semantics that assesses

what is right in terms of what is best at time t. Now, further imagine that on Planet van Roojen, the

inhabitants are rule Consequentialists and thus commit to nothing, ethically speaking, if it cannot be

said to be the right thing to do in the long run. On planet Wedgwood, they are act Consequentialists

and take practical questions on a case to case basis. Suppose the two groups should meet, and start to

commingle, and further it turns out that the group from Planet Wedgwood has learned that it is best to

sacrifice one person every day to a volcano to placate its wrath. It seems to have worked in the past,

they reason, and every day that they review the facts, they think that indeed it would be the best thing

to do at the crack of dawn. Those from planet van Roojen on the other hand are rule Utilitarians of a

universalist persuasion and think that right in the long term must ensure that no sentient being should

be harmed or compromised.

Do they disagree morally? While the moral semantics of Planet van Roojen and Planet

Wedgwood are similar, they are not identical. Thus, their moral vocabularies have distinct semantics.

They mean something different by their use of moral vocabulary, so strictly speaking, they are talking

past each other when they have an argument of apparent moral significance. Their views could be

roughly inter-translatable, but because the semantics governing their “moral” vocabulary is not

identical, there will be cases of distortion and indeterminacy in translation.

182
 In the original Moral Twin Earth thought experiment there were these outside viewers of this

situation whose intuitions were supposed to adjudicate the situation, floating above the two worlds.

And, let’s introduce these people into our experiment. These are the people from planet Hormin (they

are distant descendents of Horgan and Timmons), who have a language that is homophonic with both

the people in Planet van Roojen and Planet Wedgwood, except that their “moral” terms are governed

by a cognitive Expressivist semantics (to be dealt with in detail in the next chapter). On their planet,

such vocabulary is used to articulate judgments that have all the phenomenological appearance of

representational content, but no real, sui generis content. Further, these people are Particularists: they

don’t have general principles that they appeal to, they are just struck by what is right and wrong by

looking at a concrete situation: they know what is right because it comes with a special

phenomenological glow, and the wrong doesn’t. And let us further suppose that it is these people’s

overwhelming belief that what is right is not to sacrifice a person to a volcano, but rather to a vast

chasm, which if it cannot be found naturally occurring, should be constructed. Are they in a position

to decide whether the people from Planet Wedgwood and Planet van Roojen are having a moral

disagreement? Well, they could certainly evaluate the disagreement between these peoples from the

perspective of their own moral semantics, but that would be to overlook a doctrine shared by all

parties concerned: moral semantics is about the meaning of moral vocabulary in a language. If that is

so, people from Hormin are no better placed to judge the issue. Indeed, they too even seem to have a

practical disagreement with the other peoples, but if we buy into the view that determinate meaning is

relative to terms in a language, then it seems that all parties are talking past each other.

One proposed fix to the moral earth problem, proposed by Heimir Geirsson (2005), is to draw

upon Keith Donnellan’s now famous distinction between the referential and attributive use of

language (Donnellan 1966).

According to Donnellan, neither Russell nor Strawson provide a correct account of definite

description. Both presume that the referential function of a definite description is a part of its

183
meaning, independently of its use in a sentence, and that when the presupposition or implication of

the use of a definite description is false, the truth value of the sentence it occurs in is always affected

in the same way. However, "if there are two uses of definite descriptions, it may be that the truth

value is affected differently in each case by the falsity of the presupposition or implication"

(Donnellan 1966, 282-3). In the case of attributive uses of definite descriptions, Donnellan thinks that

their use can be veridical only if there is a thing that corresponds to the attributive description that is

thought to be part of the meaning of a term. In the case of referential uses of definite descriptions,

their uses, in sentences, may contribute to the over all truth of the sentence, even if, in reality, nothing

corresponds to the description, because it is judged as the best way for the audience to recognize what

one is referring to (Donnellan 1966, 292). According to Geirsson, on this account, for “disagreement

to arise it is not necessary that the crucial terms have a common reference or the same meaning as

long as the disagreeing parties are using the terms to refer to the same thing” (Geirsson 2005, 359).

Could Geirsson’s fix help the Wedgwoods, van Roojens and Hormins understand how it is that they

disagree with each other?

Not in principle. In this example, all we need to do to make difficulties for Geirsson’s fix is to

specify that each planet would have their pool of analytic philosophers, trained at uncovering the

semantics of their terms by inspecting their intuitions about word meaning. If they should have a joint

metaethics conference, they would discover that indeed their moral vocabulary were governed by

distinct semantics, and this would certainly undercut their confidence in thinking that they were really

having a conversation with each other about moral issues, if moral meaning is ultimately about the

semantics of word usage in a language. If they are not even having a conversation about the same

thing, semantically, the attendees may come to doubt that they were ever really talking to each other

on moral issues. And this doubt is enough to undo the workability of Geirsson’s fix, for it relies upon

the belief of all concerned that the mode of talking is the best way to refer in a context. But all those

concerned may come to doubt that the referential mode of interaction even makes any sense or that it

184
is in fact the best way to explain their interaction. But, if such reflexive musings of our tri-planet

metaethicists should go in Geirsson’s favour, try this: reconstruct the thought experiment so that the

parties never meet. Rather, futuristic Star Trek anthropo-noid-ologists (with the powers of

invisibility) study their various cultures. Would they be in a position to decide whether these cultures

disagree, morally? One thing is for sure: Geirsson’s fix wouldn’t help, for the various parties are not

in one context and can thus not judge whether a referential use of language is the best way to refer to

some item of common interest, for, by definition, there are no common items of interest and no

common speech act contexts.

Here we come to one problem with traditional accounts of moral disagreement. They are

always calibrated with reference to similar or shared contexts. But real-life translation of normative

discourse needs to understand how people who have never met could disagree on values. Boyd-style

referential semantics is simply the crudest of such context relative explanations of disagreement, for it

makes out moral disagreement to be barely distinguishable from controversies over how to use a

word, namely “good”, within a linguistic practice: significantly change the practice such that there are

no longer common referents and one cannot even recognize cross-practice disagreement.

C.L. Stevenson’s account, for instance, makes moral disagreement a function of two

disagreements. One type of disagreement is factual and pertains to the dry particulars of the case

(whether, for instance, the sun rose today at 6:54 am). The more central notion of disagreement in

ethics, for Stevenson, is a disagreement in attitude. When two parties disagree in attitude, they have

differing attitudes towards some one object. But this works best for people who can have different

attitudes about common items.81 If people do not share contexts, it becomes difficult to understand

this route of explanation for the matters that they have attitudes towards are not common.

Stephen Finlay has similarly argued that moral disagreement has a two part feature. We can

disagree over moral facts relative to shared standards, or we can disagree in attitude when we do not

share standards, but yet share a context, thus using noncontradictory moral statements as though they

185
were contradictory for rhetorical purposes, facilitated by Gricean conversational implicature (Finlay

2005). In our reconstructed thought experiment, the various peoples neither share a common context

for conversational implicature to do its work, nor shared standards.

Alan Gibbard has recently proposed that ethics is fundamentally about plans, and that the

fundamental disagreement of this sort is a disagreement of plan. To disagree on a plan is to disagree

about what we would do if we were in someone else’s shoes. In otherwords, for me to disagree with

you, morally, I must disagree with what is the best course of action for you to take, if I were in your

circumstance (Gibbard 2003b, 69-70). However, in order to assess whether two people disagree with

each other in plan, one has to be able to access relevant facts about their dispositions relative to

scenarios that they may have never been in. The only way to assess this, given the possibility that the

scenarios that disagreement are relative to may not be scenarios that person has even contemplated, is

if we knew a lot about their broadly philosophical beliefs to begin with. With such a picture, we

could deduce from a person’s broadly philosophical commitments (about what they think is right,

wrong, real, unreal, genuine knowledge, and misleading) what they may be inclined to do in

counterfactual circumstances. Given the status of the problem as set out so far, it is not clear how we

could calibrate such philosophical convictions of persons relative to counterfactual circumstances, for

it seems we are not even sure how to translate the various convictions of the various persons from the

various planets. Or, put another way, if we could independently access such facts, we wouldn’t be

faced with a problem of translation, and we wouldn’t really require Gibbard’s account of moral

disagreement to explain the disagreement (for we would have established how they disagree on

broadly philosophical issues, including their philosophical views on ethics that would explain their

planning dispositions in any context).

In the interest of being thorough, we should note that Moorean, non-naturalistic explanations

of what “good” refers to could not work to coordinate disagreement on moral issues across words and

languages either. For it is a feature of that account that what “good” (in English) means is a certain

186
non-natural property. Even if this were the case, all we would need to do is compare Planet Moore to

our other three, and we would see that the introduction of heavy metaphysics into the mix solves

nothing. It would simply be a Idealistic counterpart of Boyd’s materialism.

This might seem like an odd claim to make. Surely, we might think, because Moore

recognizes non-natural properties, he would not be stuck with having to worry about translation, for

the usual types of naturalistic hurdles to translation would be overcome by countenancing non-natural

properties.

This is tempting, isn’t it: simply call upon non-natural metaphysics and then it seems that the

problem of translation is solved. However, as we noted in the introduction, this move solves nothing.

For even though Good, on Moore’s account, is a non-natural object, languages are natural

phenomena, and thus, even if Moore’s non-natural GOOD were the referent of “good” in a language,

this would be a natural fact about that language. Consider an analogy. It is very plausible to think of

numbers as non-natural objects. We can say many truth things of them, and they are not to be met

with in the natural world. However, the fact that English has a word for π and the means of

articulating this number, even if only approximately (“3.14…”) is a natural fact about English. Some

languages seem to lack the capacity for articulating such numbers altogether (cf. Everett 2005). Thus,

even if we bring Moore into the picture, the language that has a non-natural property as the meaning

of its moral vocabulary would merely be one out of the many planets, and its moral semantics would

be as incommensurable with the others as they are with each other. But this shouldn’t surprise us. If

Platonic metaphysics could solve the problem of translation, we could merely will translations into

existence, without ever having to gain expertise in a target culture and language. One might think that

non-naturalism solves the problem of Moral Twin Earth scenarios because it allows reference to be

realizable regardless of the natural context. But this solution overlooks the real problem: Moral Twin

Earth, and all problems of translation, arise out of the fact that the semantics of natural languages

differ. The confusion in the discussion of this topic in the literature is the notion that problems of

187
translatability come about because of referential semantics. Rather, the real problem is the linguistic

account of translation and the correlative account of meaning.

A defender of Alan Gibbard might object to the manner in which I have set out the problem

here. Gibbard proposes his planning language to clarify issues of ethics, and thus it would be best to

use the Gibbardian language as a type of lingua franca in normativity. The problem is that there is no

reason for any parties to think that the language spoken on Planet Gibbard has any privileged or

special priority. Certainly, the inhabitants of that planet will want to extend their way of looking at the

world, mediated through their planning language, to the problem of interpreting others, but this is

exactly what everyone wants to do, and in each case, they would be prevented by their shared thesis

that translation is an exchange of synonymous words across languages. If, as is generally assumed,

cross-linguistic synonymy underwrites translation, then all parties concerned couldn’t translate each

others’ normative assertions into their own languages.

Let us bring our Star Trek anthropo-noid-ologists back into the picture. Could they come to

the belief that the various planets (add the Mooreans into the mix, who we can assume believe that

what is ethically right is the best possible, which often on their view involves the contemplative life)

have a disagreement on values? Yes, indeed. But, we might ask, how would they do it without also

presupposing a semantics for what moral terms mean? If they did, wouldn’t they just be another race

to be added to the mix, with their own incommensurable semantics for moral terminology?

Moreover, doesn’t the manner in which the whole issue has been set out incline us to think

that indeed there is some type of normative disagreement solely because the languages are

homophonic? If so, isn’t this a cheap trick, that seeks to elicit intuitions that wouldn’t otherwise be

elicited if all the parties were speaking languages that were phonemically distinct? Really, aside from

the use of “right” in the thought experiment, the anthropo-noid-ologists would have no evidence or

reason to think that any views on morality are being articulated unless the views thereby conform to

their own outlook.

188
 No on all scores. Our anthropo-noid-ologists have done their translation theory (chapters 2

and 3). They know that determinate translation is not a matter of matching the meaning of words

across languages, but rather an activity mediated by text-types. Thus, they would need only to pull

from the shelves of text-types the appropriate one for understanding value discourse. And there is

only one that will work: the philosophical text-type. With this text-type, the anthropo-noid-ologists

would employ QI, as the text-type feature of philosophy, to understand the various things that the

various inhabitants say with their terminology. The anthropo-noid-ologists would amass the texts of

these various cultures, and if they could not find such texts, they would transcribe the various claims,

discussions and discourses given by these people. Then, they would look at the pool of texts relative

to each planet for a textual marker that serves as a means of articulating theories that are moral. The

question of what counts as a moral theory could occupy its own chapter (and indeed it did, but is

omitted from this dissertation for considerations of length). One factor in identifying a theory as a

moral theory is a matter of direction of fit, and moral theories as a species of normative or evaluative

theories have a world-to-theory direction of fit (the world is to be judged in light of the theory). How

we determine the axiological differentia of any key philosophical term on the account I am providing

is (roughly, with some caveats to be discussed in chapter 8) by determining the lowest common

denominator among theories that have been articulated with a key philosophical term. If we conduct

this inquiry, I think we must recognize that with respect to the theories articulated under the heading

of “ethics” and “moral” the only common denominator of this wide range of theories (everything

from ethical egoism to consequentialism) is that they are theories proffered for their social

implications. With this perspective, the anthropo-noid-ologists would be able to tell that the

inhabitants of the first three, as well as the Mooreans have moral disagreements on many fronts.

First, they will disagree on what theory they choose relative to such textual markers of moral

discourse. Secondly, they will disagree upon the reasons that justify their choice of theory selection.

Third, from the perspective of philosophy, we can even show how they have a disagreement over how

189
to use moral vocabulary, even though we have accepted that their respective uses are governed by

different “semantics.” For the text-type of philosophy, in treating their moral vocabulary as markers

of a philosophical text, is able to produce translations of the various planets’ tracts on moral

philosophy, and to distribute them. In producing those translations, the target language “moral”

vocabulary will be used to translate the source language moral vocabulary, not because they have the

same linguistic meaning, but because they share a function relative to the text-type of philosophy.

Thus, even if the languages spoken by the various planets were not inter-intelligible, the translations

produced by this method show to all concerned, using their own language, how they have a moral

disagreement with their extra-planetary counterparts. A superficial reading of the texts would leave an

impression with all concerned that the disagreement concerns how to use words in the target

languages, but closer inspection will reveal a rich difference in moral philosophical outlook. We see

thus that translation has been successful relative to the text-type of philosophy because meaning has

been preserved: not only the text-type-theoretic meaning, but even the planet-relative “semantics” of

how to use moral vocabulary makes its way through translation, and this is only possible through the

combination of the semantics of a language and the text-type. In other words, when we use the text-

type of philosophy to translate Planet Hormin’s normative discourse into Planet van Roojen’s, we use

Planet van Roojen’s moral terminology in accordance with Planet Hormin’s semantics for moral

terminology.

Philosophy thus provides a context-independent frame to assess the values of diverse peoples

from diverse backgrounds. We can, after having established these philosophical particulars, move

back to our parochial ways of assessing moral disagreement. We would have all the information we

need to judge how these people would differ in their attitudes to things (à la Stevenson) for we would

have fixed their universal and general normative theory against which they judge the world. We

would know how they could disagree on plans à la Gibbard because we would have enough

information having fixed their universal and general normative theory to deduce, or at least make an

190
excellent educated guess, as to how they would plan differently if they were dropped into shoes they

were not contingently used to wearing. We might even have enough information to judge how such

peoples would react to each other’s speech acts if they were in the same context but yet spoke the

same interintelligible languages—whether they would adopt the attitude that Geirsson thinks will help

them negotiate their disagreement in contexts.

The reason that QI can work to provide such an account of translating the various texts of

these peoples is that it operates at a level over the fray of the usual controversies that characterize

normative discourse. It is an account not of how people should be using their normative words, but

rather a descriptive account of the features of normative texts that translators must keep track of to

produce successful translations. While the preceding thought experiment is silly, it is realistic in one

respect. In real-life, people have disagreements not only on obviously substantive issues in ethics, but

they also have disagreements on metalinguistic matters of how to use moral language, i.e., semantic

disagreements. These semantic disagreements are not wholly extractable from the substantive views

people hold. Often, the accounts that native linguists are apt to give of the meaning of their words are

barely distinguishable from their substantive views (e.g., “morality means being nice to people, not

lying, keeping your promises…”) and such native speculations on word meaning influence how

people use their language. Indeed, the same is true for the metaphysical and epistemic considerations

native linguists provide for their substantive claims. There is a slight tradition in recent philosophy of

arguing that metaphysics has nothing to do with ethics, but that is only because we in our own case

are quite blind to our metaphysical commitments in ethics, such as: what types of beings are moral

patients, what types of beings are moral agents, what types of properties (actions vs. character traits

vs. non-natural properties vs. natural properties vs. mental properties) are tracked by moral theory. 82

Of course, the same is true of epistemic considerations in favour of a substantive view (e.g. “abortion

is bad because the Bible says so and it is the word of God”). Broad philosophical considerations

cannot be easily distinguished from the normative theories that speakers hold and these considerations

191
jointly influence how they use their normative vocabulary. The cumulative effect of such native

speculations and rationalizations of their substantive views is a unique semantic profile for each

language, which shifts according to a variety of considerations, including the philosophical

convictions of a population: this is the linguistic reality that confronts scholars at the crossroads of

languages, attempting to construct bilingual lexicons. If we were to attempt to assess the meaning of

words in a language, given all the linguistic data, we would quite naturally come to conclusions about

the meaning of words in a language that were strongly constrained by such native linguistics. But if

translation is possible at all, it must be possible despite the fact that vocabulary across languages will

have meanings that are responsive to historical particulars and never identical across languages. If the

translation of normative discourse is possible at all, it must be despite the dizzying philosophical

differences native speakers invoke in the rationalization and codification of their substantive views.

Only a text-type can mediate the lexical differences across languages in translation. The only text-

type that can be sensitive to these various philosophical factors that contribute to the codification of a

substantive normative view is the philosophical text-type. The reason for this is simple: it is the text-

type of philosophy that will preserve the various philosophical considerations closely connected with

persons’ normative views in translation. Including such philosophical differences in normative

discourse translation will help us accurately capture the author’s normative views but it will also help

us understand how we morally disagree with people who we have never met. Failing the intervention

of philosophy, there is no way to coordinate or appreciate how we normatively disagree with those

who we’ve never met and do not share contexts or languages with.

The tight relationship between the substantive view and the various philosophical

commitments invoked by speakers to underwrite their substantive views would certainly be a factor

that our anthropo-noid-ologists would take into account. The various “semantics” of moral

vocabulary for the various words could variously be rolled into the theory that we know them to be

articulating or the relevant reasons that they have for selecting the theory, depending upon whether

192
the semantics were, in the minds of the native speakers, normative of how to use moral vocabulary, or

whether they were intended more as descriptions. If prescriptive, they would be rolled into the

normative theory articulated with moral vocabulary. If descriptive, they could be rolled into the

reasons that speakers provide for their theory use and selection. Other similar issues would have to be

determined on the basis of the particular data. (Thus, for instance, in the case of the Particularists

from Hormin, our anthropo-noid-ologists would have to decide whether their particularism was

because of a theoretical commitment to Particularism as a universally prescribed methodology in

moral philosophy, or because they thought that, as a matter of course, each time they addressed a

moral issue, no general considerations could be found and that the context dependent reasons did the

work in arriving at their decision. In such cases, the theory would be left as a blank variable, as

something to be filled by a theory that could satisfy all of the judgments the people come to. Or, the

theory could be a specifically temporary theory, that simply records past findings but dates them so

that they are not taken as standards against which the future is to be judged.) Either way, whether we

assign their native-linguistic intuitions about their moral semantics to their reasons for theory

selection or the theories themselves, once we had the relevant information from studying this people,

we would have plenty to determine how they disagree and, in particular, how they would differ in

practical cases. All we have done in helping ourselves to this explanation is rejected the privilege

accorded to linguistics in semantics, and extended our institutional standards of reading philosophical

texts to interpreting these peoples.

IV.5. The Explanatory Failure of Ethical Naturalism

Science fiction is silly, but it serves a purpose in philosophy, so long as the ridiculous

scenarios retain just the right factors in order to render the thought experiment fruitful. The Moral

Interplanetary System thought experiment in the previous section is fruitful in this respect, for it does

not assume that one requires interlingual synonymy for translation. The reality of cultural diversity

and linguistic difference makes interlingual synonymy impossible in most all real cases of translation.

193
No theory of translation can be successful if it cannot explain how translation is possible despite

linguistic and cultural differences. None of the going accounts of moral discourse translation in the

literature make it over this hurdle. In short, the real science fiction is not the Moral Twin Earth

thought experiment, nor the Moral Interplanetary System experiment that I propose, but the dominant

account of translation in the analytic philosophy literature. The linguistic account of translation is

science fiction because it is not based upon the realities of the empirical world but rather fantasies of

a world devoid of true and profound cultural diversity. Or, rather, what we have is not science fiction,

but tales of denial, in so far as philosophers have proposed accounts of the meaning of moral

expressions in their own language as a condition of the translatability of normative discourse. To

assume that translation is only possible if there are expressions across languages that are in and of

themselves semantically equivalent is to assume that translation is only possible when there is no

cultural diversity.

QI, based upon TTS, comes to the rescue by injecting into the discussion a measure of

realism that is based upon the realities of cross-cultural research. Academics working in the front

lines of cross-cultural dialogue and research know that cultural difference makes for semantic

differences in expressions in a language. Real translators have always had to deal with this fact. TTS

is a general theory of translation that takes reality into consideration and calibrates our expectations of

where and what counts as objectivity in line with the reality of cultural variation. Its answer is that

texts of definite types are semantically determinate, while language is not. Texts of a definite type are

semantically determinate because they are translatable according to trans-linguistic and trans-cultural

textual expectations, while language itself is a process of continual cultural transformation. To think

that objectivity can be founded on language is a mistake in light of the radically contingent nature of

culture, and TTS recognizes this by looking to the very text-type of an institution to ground the

objectivity of epistemic claims. Thus we saw in chapter 3 that the text-type of science provides us

with the criteria to adjudicate between competing claims to scientific success, even though language

194
as such is in a state of flux. The same can be said about any text-type. Philosophy too, I believe,

provides us such criteria. (I shall discuss how philosophy as a type of text can provide objective

knowledge in the conclusion.)

In light of the radically contingent nature of culture and language, what can we make of

Ethical Naturalism, and the notion that objectivity in ethics is grounded upon the contingencies of

certain locales? Let us consider Boyd’s account as a concrete example of such an Ethical Naturalism.

Boyd wants to provide a causal account of the meaning of “good” in English. But given the

radical contingency of cultural phenomenon, it is difficult to establish that the referent of “good” is in

fact causing speakers to respond to it in a certain manner, or that it best explains the meaning of

“good.” This becomes clearer when we attend to the question of how we are supposed to determine

the reference of terms.

There have been many linguistic proposals on how to determine reference. There is the old-

fashioned Frege-Russell approach, where one looks to descriptions as setting out the conditions of

reference. The post-Kripke approaches to reference take a different route. Kripke suggested that we

think of reference as a purely linguistic matter of naming ceremonies, with a social process of term

dissemination, connected by a chain of referential intentions that link back to the baptismal ceremony.

Another way we might determine reference is by leaving it to the scientists or other authorities in a

culture to decide how vague terms should be applied in actual cases. This is the Putnamesque version.

As a type of compromise between the old and new approaches, we could decide that reference is

determined by a convergence of the descriptions that people associate with a term, or by a

convergence of what people more often than not identify with a term.

Boyd proposes a mixture of such accounts to determine the reference as a causal factor:

Roughly, and for non-degenerate cases, a term t refers to a kind (property, relation, etc.) k just in case
there exist causal mechanisms whose tendency is to bring it about, over time, that what is predicated
of the term t will be approximately true of k (excuse the blurring of the use-mention distinction).

195
 Such mechanisms will typically include the existence of procedures which are approximately
accurate for recognizing members or instances of k (at least for easy cases) and which relevantly
govern the use of t, the social transmission of certain relevantly approximately true beliefs regarding
k, formulated as claims about t (again excuse the slight to the use-mention distinction), a pattern of
deference to experts on k with respect to the use of t, etc .... When relations of this sort obtain, we
may think of the properties of k as regulating the use of t (via such causal relations)... (Boyd 1988,
195).

This account of reference, like most all accounts of reference, relies upon linguistic practice

and usage to determine reference. We are supposed to treat the item referred to, on this account, as

causing us to refer to it. However, in order for it to cause any such thing, a society must have a certain

social organization in language use that brings it about over time that we consistently name this item

correctly and thus say true things of it. These mechanisms also involve experts who people defer to in

their usage of terms. In the case of value terminology, we must thus countenance moral experts that

people at large defer to in order to direct their value term usage in order for there to be reference. That

such “experts” chosen by people at large should converge in any suitably cosmopolitan society is

implausible.83 But if there is no such convergence, reference, as Boyd understands it, will not obtain.

Thus, the account assumes a certain implausible cohesion for there to even be moral reference. And if

reference just is meaning (and usually, for the referentialist, reference just is meaning), it seems that

we need an implausible cohesion in society to get moral meaning off the ground. For all of its

problems, Boyd’s account of reference is a plausible linguistic candidate to account for some type of

real world reference of moral terminology. In the case of ethics, people do often mediate their

referential intentions through experts (the Pope, the Dalai Lama, or perhaps the Koran itself) and it is

these authorities who are often thought to play a decisive role in disambiguating the content of moral

concepts.

But societal organization is in a constant state of flux, and varies according to many cultural

factors, including the ideas that are prevalent in a society. Thus, the notion that, for all the social

coordination necessary to get reference off the ground, we have something that necessarily or

196
uncontroversially or metaphysically in the world is causing us to refer to it is doubtful and naïve—

particularly in the case of value terminology for here the beliefs and allegiances of a culture that are

value laden will lead the society over time to identify a certain item with a certain term and thus over

time this society will be able to say many true things about this item, but only because of its latent

evaluative tendencies.

The problems are more serious however. As an empirical hypothesis, Boyd’s causal account

of “good” is very poor, for it cannot establish the causality of reference independently of the values

people hold. Given a population with predominantly anthropocentric values, indeed their usage will

start to converge in the manner that refers to Boyd’s cluster property, but this would be hardly

surprising. The “experts” elected by members of a society to whom they defer their moral language

use to will simply be those people who reflect the values of their votaries. There may be something

materially objective being referred to in such cases, but the idea that this counts as an alethic

justification of the claims being made would be strange indeed, for given any set of values and a

population that is suitably cohesive, one could get an objective reference on Boyd’s account. A

community of sadists could thus come to successfully refer with “good” to some sordid cluster

property, but that would make judgments about good in this community objective. People can always

simply self-select into communities of like minded persons to make their judgments of value

objective, making use of the exact mechanisms Boyd specifies. The notion that there would be

something objective about such judgments would be true enough in one sense (namely that language

use refers to something in the material world) but that this objectivity should have any normative

implications is far from clear.

There are other problems with the hypothesis. If Boyd’s view is truly naturalistic, it must aim

at establishing, empirically, that moral terminology across cultures is responsive to the type of cluster

property he envisions. For this to be possible there must be some independent means of objectively

identifying moral terminology across cultures so that we might empirically inquire about the causal

197
mechanisms governing its operation. But whatever allows us to identify moral terminology across

cultures as a condition of testing Boyd’s theory would itself constitute the semantics of moral

terminology. A condition of the testability of Boyd’s theory is thus that it not be treated as a semantic

thesis. This is a paradox for Boyd’s view. If we were to treat it as a semantic thesis about the

meaning of moral terminology and as a condition of the translatability of moral terminology, we

would be converting what is supposed to be a non-analytic hypothesis into an analytic hypothesis

about the synonymy of moral terminology across cultures. It would thus be unclear what would count

against such a hypothesis.

The problem with objectivity in the case of linguistic judgments is not simply confined to the

case of ethics. Even with matters of interest to science it is not always clear that what people are

referring to has any essential foundation in the natural world. Categories of race, folk categorizations

of diseases—even scientific categorizations of illnesses that are later recategorized as “syndromes”—

not to mention many putative natural kind terms turn out on scientific examination to have no

underlying natural essence to account for the linguistic dispositions of persons to refer to them under

one heading. As noted in chapter 3, often scientists will regiment a natural kind term for the purposes

of scientific discourse in a manner that diverges from what is popularly referred to under the heading.

At times, linguistic practice hits on a kind that will be of interest to scientists, and at many other

times, it does not. Cultures can draw distinctions between kinds that are of no obvious interest to

natural scientists. Racial categorizations are of this type. Genetic variation among humans is such that

racial categorizations of folk taxonomy do not track anything significant at the genetic level. But

categories of race still refer, for all manner of reasons. We must rely upon them as referring terms for

policy issues (to implement programs of affirmative action, for instance) and they refer for no other

reason than that many people must live with the fact that their society treats them first as a token of a

certain racial category, and only secondly, if at all, as full members of society. While racial

categories may not be categories of biology or genetics, doctors may still have pragmatic reason to

198
rely upon them, to the extent that they loosely correlate with populations that have a predisposition to

some diseases—natural facts of these populations that can in no way be thought of as part of the

explanation of what caused people to originally make use of racial categories.

Reference and the objectivity of what is referred to in language is not the problem. The

problem is that reference and objectivity in the linguistic sense come too easily. We can and do use

descriptions (as Frege and Russell thought) to refer when we are having trouble finding the right

words. We do rely upon social avenues of language use to mediate reference. Within contexts, we can

even misuse descriptions to refer, as Donnellan noted (Donnellan 1966). It is the multiple ways that

expressions come to refer that contribute to the richness of language: it is these associations that in

large part constitute the semantics of a language. As argued earlier (III.2.1 The Case of “Technical”

and scientific Translation and the Question of the Determination of Reference pp.141-148) it also

necessitates text-type institutions so that we can objectively disambiguate reference. The objectivity

of such disambiguation is assured because such institutions are not reducible to a culture or language.

We have good reason to believe that “electron” has referents in the natural world not because of the

facts about English (as though we could do physics through linguistic analysis) but because scientists

have shown through testing and the consilience of inductions that physical theories that recognize

negative charges in atoms are credible—theories that could be articulated in any language. “Electron”

could also be the name that a post-modern beat poet might give to her son, and the work of scientists

to show that “electron” refers may not actually play any role in accounting for why this child has the

distinction of being known as “electron.”

To stress how linguistic reference is of indeterminate significance, and how vacuous the

notion of causality in reference can be, we might consider a society that mistakes certain effects of the

sunlight on rock formations for ghosts. Over time this society might develop a practice of referring to

these optical phenomena with a certain term, and tribal elders may play a pivotal part in mediating the

beliefs of persons in this society about these optical phenomena. They would likely come to say many

199
true things about them in the long run: they would know, for instance, when and where these

apparitions appear, and what their distinct visual appearance is. They would be wrong that the

phenomena constitute the souls of departed persons, but that would be simply one false belief among

many true beliefs that they would have about these optical phenomena. Among the true beliefs that

one might think these people hold is that these apparent apparitions are frightening, for indeed, these

people become quite scared when they see this spectacle. And many of the truth things they will say

of these phenomena are true in light of their psychological reactions to them. Does the optical

phenomenon cause them to be frightened? I’m not sure that there is a clear answer to this question.

But more importantly, does their word for “ghost” refer? Certainly, on Boyd’s account of reference, it

seems that it does. Does this mean that ghosts are real? Can we give just one answer to this question?

I don’t think so. Indeed, ghosts are real, but not for the reasons that these people think. Or, we could

put the explanation another way: what are being called “ghosts” in this community are not the souls

of departed persons, but optical phenomena that are a function of sunlight and a peculiar rock

formation. Could we say that the optical, naturalistic phenomena are more fundamental in causing the

beliefs of these terrified people, or is it their cultural outlook? Here, in part, it would be a matter of

counting the various beliefs these people have about these phenomena in order to discern which can

be attributed solely to the physical factors, and which are due to their cultural outlook with the hopes

of discerning which side the preponderance of beliefs falls upon. But individuating beliefs is no easy

matter because they do not come neatly packaged. We can elicit a virtually infinite number of beliefs

from a person on any given topic by asking them about various aspects of what their beliefs are about.

If the society was not inclined due to their lore to believe in ghosts, they would not be inclined to

view such optical phenomena as ghosts. If they were not privy to such optical phenomena they would

not be inclined to think that ghosts are real. I don’t think there is a correct answer here on the strength

of the linguistic evidence alone. Language is semantically indeterminate for such considerations. If

we thought, for instance, that the optical phenomena unambiguously cause them to be frightened, we

200
might think that they are the more fundamental. But if they had a different world view that had no

place for ghosts in it, the optical phenomena would not be frightening and they wouldn’t call them

“ghosts.”

The point I am making here is distinct from the point that Hilary Putnam has made about

reference. Putnam has emphasized that the world is a mereological continuum and that language

divides the world into discrete parts (Putnam 2004, 1987). I think this both gives language too much

credit and not enough credit. Language does not unambiguously divide the world. Rather, it imbibes

multiple semiotic relationships that are a function of the interaction of persons with the natural world,

in part, not to mention persons with their own abstract creations (such as, for instance, fictional

stories) and thus we require non-linguistic, textual means of selecting out of these multiple

significances some that we will take an interest in, for some specific reason. Text-types and their

institutions play this role.

Looking back to our society that thinks that ghosts are real, we could take their lore that

speaks at length on ghosts and we could read and translate them according to more than one text-type.

If we were to treat them as proto-scientific texts, we would read and translate them as texts

concerning descriptive theories that are forwarded on the strength of empirical considerations that are

thought to be objectively persuasive. So understood, we would treat the text as though it were making

the case for the existence of ghosts. Scientists would thus be able to understand the text as picking out

a putative natural phenomenon in the world and they could thus subject the society’s theory about

ghosts to empirical investigation and the consilience of inductions. They would probably find very

early that the theory about ghosts can be falsified very easily by blocking out the sun from the rock

formations (if there were truly ghosts, one might think, the sunlight should make no difference in

their appearance—though the society may have an ad hoc explanation for this). Scientists might

conclude thus that there are no ghosts and thus that “ghost,” and this society’s word for ghosts, does

not refer. Alternatively, anthropologists could study these texts and treat them as documents outlining

201
the people’s world view: their culture’s cosmology. So understood, the scholar would not treat the

theory as one primarily about empirical phenomena, but rather the ghosts would be recognized as

entities in this culture’s literature about the world. This society’s word for ghosts does refer, in

exactly the same way that “Hamlet” refers to a character in Shakespeare’s famous play by the same

title. Just as we can say true and false things about Hamlet as the prince of Denmark, so we can say

true and false things about Ghosts, relative to this culture’s world view. If we were to treat such texts

as philosophical texts, we would have to treat them as setting out universal and general theories on

specific topics for relevant considerations that are thought to be objectively persuasive. We would

likely find, I suspect, that their claims about reality are rather poorly argued for and suffer for their

incurably perspectival nature, just as, I think, we should conclude the same regarding much current

philosophy.

TTS thus provides us a way to wade through the multiple significances of language when it is

pressed in the service of constructing texts. It also shows us how we require institutions of documents

and their translation to help us with reference. Even something as simple as whether “Nirvana” is the

name of a band that was fronted by the late Kurt Cobain requires legal documentation (apparently

members of a band from the 1960s disputed the 1990s grunge band’s claim to the title—the matter

was apparently settled out of court). We can use expressions to refer within contexts, but whether

such reference has any objectivity outside of our conversational contexts requires an institution of

texts to vouchsafe for claims of reference. The point of such institutions is always selective. Even if

the courts, presiding over legal documents, should decide that “Nirvana” is not up for grabs for the

Seattle trio, it would not follow that fans and the band themselves never successfully referred with the

name “Nirvana” to the Seattle trio. Rather, it would only show that for the purposes of advertising,

royalties, and fiduciary matters, the Seattle band could not be successfully referred to with “Nirvana.”

Of course, none of this has any bearing on whether there is a state of final release from suffering

known as “nirvana” in the Indian tradition.

202
 The notion thus that there is just one phenomenon that answers to “reference” is simplistic,

and the notion that the philosophy of language will shed light on this phenomenon is a mistake. We

need textual institutions to help us deal with aspects of semiotic phenomena so that we can select out

of the multi-semantic nature of language just those modalities that we are interested in, relative to

textual concerns.

The Ethical Naturalist wants us to treat ethical expressions like scientific expressions. The

naivety of Ethical Naturalism as it is traditionally forwarded is that it takes words in a language as the

object of scientific enquiry, without an intervening textual consideration. The criticism can be put

differently: Ethical Naturalism assumes that languages are semantically determinate in and of

themselves, and thus it assumes that there is only one way to account for expressions such as “good.”

“Good” like any expression in a language is polysemous, and its determinate significance only comes

to light when we are purposively selective about what it is that we want to extract from an expression

like “good,” in translation. Since translation is only determinate at the textual level, we need to bring

Ethical Naturalism up to speed. Let us therefore understand Ethical Naturalism as the view that texts

with ethical expressions in them should be treated as texts of science.

Notice, once we bring TTS into the picture, the Ethical Naturalist has lost the argument, in so

far as the Ethical Naturalist can no longer claim that there is something peculiarly scientific about

ethical expressions. If we are talking about texts, and not expressions, then one text can be translated

according to multiple types. That we can translate a text with ethical expressions in it as a scientific

text does nothing to detract from the fact that we could treat the text as a text of philosophy as well. I

think however that there are distinct advantages of the philosophical approach to translating ethical

discourse.

If we were to translate a text with ethical expressions as a text of science, we would not

necessarily be interested in its normative implications. Texts of science are primarily concerned with

descriptive theories. Perhaps if the normative implications of ethical expressions are part of what is

203
involved in empirical considerations in favour of a scientific theory, we might think them worthy of

retention in translation, just as the normative aspects of a theory of clinical practice have a role in the

empirical support of a descriptive theory of biology. But this would be quite derivative.

The Ethical Naturalist’s view, of Boyd’s variety at least, is that ethical terms have a scientific

significance in the same way that a natural kind term has a scientific significance. If we treated such

terms thus, and translated texts they occur in as scientific texts, we would have retained their rigid

referential relationship to their referent in translation. We would understand the role of these terms as

occurring with a certain type of naturalistic theory (scientific texts are always mediated by scientific

theories) and we would thus have to translate such a text in light of the text-type of science, treating

the ethical expressions as technical terms articulating a descriptive, scientific theory put forth in light

of empirical considerations that are thought to be objectively persuasive. The concern would not be to

show that ethical expressions necessarily are a matter of cross-cultural debate, but rather to consider

the reference of ethical expressions as worthy of scientific enquiry.

To treat a text with ethical expressions as a text of philosophy, in contrast, is to treat these

expressions as quasi-indexical theory articulators. Thus, we would treat such expressions as

articulating theories of a universal and general nature chosen for their social implications (as I have

suggested) in light of relevant considerations that are thought to be objectively persuasive. Cultures

with distinct ethical expressions that have their own culturally particular referents, judged from the

perspective of their own culture, could still be understood as talking about the same thing, to the

extent that both cultures could be understood as making claims to the effect of what the objectively

correct moral philosophical position is. In contrast, if we were to treat putative ethical expressions as

technical terms in a scientific text, we would have to treat such texts across cultures as talking about

different things, if such terms refer to different items, for they would thus play roles in descriptive

theories that are not concerned with the same phenomena even though the considerations brought

forth for these descriptive theories are presented as objectively persuasive.

204
 Consider the case of Aristotle and “eudaimonia.” Aristotle is well known for his theory of
84
eudaimonia, and “eudaimonia” is frequently translated as “happiness.” However, scholars often

point out that what is typically meant by “happiness” in English is not exactly what Aristotle meant

by “eudaimonia.” But what would be our reason to think this? Aristotle seems to say many things

about eudaimonia that are commensurate with what people say about happiness in English. Thus, it

may seem that the properties these terms are picking out are similar. However, there is reason to think

that they are not the same. Claims that people might make about eudaimonia in Greek are not always

exactly what people would say about happiness in English. It seems that the properties they are

picking out are not exactly the same.

If we bring TTS into the picture, we realize that the question of whether “happiness” and

“eudaimonia” have the same meaning without reference to textual considerations is indeterminate.

The question is rather whether we should translate Aristotle’s texts on ethics as texts of science or

texts of philosophy. The answer is, we can translate them as both.

If we translate Aristotle’s texts as texts of science, we treat “eudaimonia” as a rigid

designator picking out some property in the world and Aristotle’s views on ethics as a descriptive

theory that is forwarded in light of empirical considerations that are thought to be objectively

persuasive, then it is these constellations that the translator keeps track of in creating a TT. Many of

the features of interest in philosophical translation will be dropped and the strictly empirical and

predictive features of the text will be emphasized in translation. The resulting translation into English

could not correctly be read as a text on the topic of happiness. “Happiness” according to the same

text-type-theoretic considerations, would be understood as itself picking out some particular property

in the world, and while the properties of eudaimonia and happiness maybe similar, they would not be

exactly the same. We could thus disagree with Aristotle on whether he has the right theory of

eudaimonia, but this would be based upon the facts of his linguistic practice in relationship to

considerations of scientific theory. This is what is involved in treating “eudaimonia” as something

205
like a rigid designator. It is to treat the expression as defined by historical peculiarities and its role in

a baptism that occurs within a cultural context. Not all cultures will necessarily pick out a property or

cluster of properties that is coreferential with “eudaimonia.” Perhaps eudaimonia rigidly picks out one

property, or may be a cluster, like Boyd’s homeostatic cluster property. “Happiness,” in contrast,

could be treated in texts of contemporary English speakers in the same manner and likely the property

or cluster of properties picked out with this term will be culturally peculiar.

In contrast, we could treat Aristotle’s texts on ethics as texts of philosophy, and we could

treat “eudaimonia” not primarily as a type of rigid designator, but as a quasi-indexical theory

articulator. We could understand the theory of eudaimonia forwarded by various Greek thinkers,

including Aristotle, as a theory about fulfilment, and we could recognize that “happiness” could

function as a quasi-indexical theory articulator of the same type of theory (with a world-to-theory

direction of fit, characterized by the axiological differentia of being a theory about fulfilment). So

understood, when we translate Aristotle’s texts as texts of philosophy, we can understand Aristotle

and us as involved in a debate about the objectively correct theory of fulfilment. We would note that

Aristotle has many views about happiness that we are not likely to agree with. For instance, happiness

for Aristotle is something that comes only at the end of a free man’s life (slaves can never be happy

on Aristotle’s account [cf. Politics 1280a33-34])—neither can children, and I suspect woman are left

out too as they are only barely different from slaves [cf. Politics 1260a10-25, 1263a1-4]) and it

involves not only a life of virtue but some measure of good fortune (Nicomachean Ethics Book I, Ch.

9). Happiness, for Aristotle, is only achievable for the potentially slave owning aristocrat. Our view

on happiness is typically quite different. As one author notes, what we have in the contrast between

Aristotle’s conception and ours is really two distinct conceptions of happiness: Aristotle provides for

a more stringent criteria of what would count as happiness, while we are often more inclined to think

that happiness depends on a number of subjective factors (Kraut 1979).

Just to emphasize, the very recognition of the role of text-types in translation is the

206
recognition that language is polysemous. We can translate philosophy as poetry, and the texts of an

ancient culture as both science and philosophy, likely. Thus, the Ethical Naturalist’s view must be

reunderstood in this updated context as the claim that out of competing ways of translating and

understanding the semantics of ethical discourse, it is better to understand it as a species of scientific

discourse. Why? The naturalist believes that the natural world provides a better explanation for

ethical convictions than does any other manner of analyzing ethical discourse. Is this true? In the case

of Aristotle’s texts on ethics, if we translate his texts as texts of science, we find that we must

understand “eudaimonia” as an originally culturally contingent topic, defined by the contingencies of

Greek linguistic practice. What eudaimonia is about is not something that people from anywhere else

in the world could have thought and written about, unless their culture was incredibly similar to the

Greek’s (in which case they too would have an expression that was coreferential with “eudaimonia”)

or if they were fortunate enough to read scientific translations of Greek texts. Translating them as

texts of science provides us a type of epistemic access to the Greek linguistic practice, to the extent

that “eudaimonia” is a term whose reference is supposedly determined by such a practice. We thus

have a type of window into the world of the Ancient Greeks, and we might begin to understand the

various claims they made concerning eudaimonia. But what we do not have by treating such texts as

texts of science is an explanation of why the Greeks tracked the property they did track with their

term “eudaimonia” and why it was not the property we track with “happiness.” Or, conversely, we

have no explanation as to why it is that we are not tracking eudaimonia and why the Greeks are not

tracking happiness in their linguistic practice. In other words, if we go down this naturalistic route, we

have no explanation of cultural difference.

In reality, the notion that words unambiguously track properties in the world I think is an

oversimplification, and I have argued that in this section and in different ways in chapters 2 and 3.

But if we were to ignore the complexities and treat Aristotle’s texts as though they were a type of

proto science by virtue of their ability to capture aspects of Greek language use, what we arrive at is

207
not much of an explanation at all. “Eudaimonia” remains an incurably culturally relative notion and

we have no explanation for why the Greeks never bothered with any other property that was similar

but distinct. We might be able to provide some type of culturally specific explanation: perhaps the

concept of EUDAIMONIA is linked to other concepts of interest to the Greeks, but all we would have is

a web of concepts peculiar to the ancient Greeks. In effect, we would arrive at an explanation that was

in some sense incommensurable with cultural phenomena elsewhere.

If we translated Aristotle’s texts on ethics as texts of philosophy, and the various other texts

from ancient Greece on the topic of eudaimonia as texts of philosophy, or in light, at least, of the

philosophically interesting texts on eudaimonia by authors such as Aristotle, we would have an

explanation for why the Greeks were tracking the properties that they were with their terms. The

explanation would be philosophical: namely, they had a certain conception happiness, as a universal

and general theory about fulfilment (if this is the correct axiological characterization of such theories)

and we would recognize that the reason we track distinct properties with “happiness” is because we

disagree with Aristotle and the Greeks. We do not think that happiness and fulfilment are only for the

aristocrat. We don’t think that anyone is born for slavery, and we think that an ideal theory of

fulfilment should make room for happiness in a child’s life. Thus, I suggest, the philosophical mode

of reading and translating such texts provides us with an explanation of cultural difference, while

treating ethical texts as scientific texts about descriptive theories about properties in the world does

not. The philosophical mode of translating such texts provides us an explanation where the scientific

mode of translation does not, for it takes into account the opinions and philosophical convictions of

persons as crucial features of cultural explanation. The putatively scientific approach that seeks to

locate the explanation in the natural world does not.

For understanding questions of cultural difference, the philosophical manner of interacting

with texts is far superior for the simple reason that we all live in the same natural world and our

biological constitutions are remarkably similar.85 Humanity has not biologically evolved that much in

208
the past three thousand years. Despite all of this remarkable similarity in the natural facts, cultural

difference abounds. If we are going to find an explanation that does justice to this difference, it has to

take into account the various philosophical choices and opinions of persons, for this is the wild card

in the picture. The philosophical mode of interacting with texts of cross-cultural interest also does

justice to humanity in a moral sense: it does not conceive of persons as mere passive, empirical

reactions to causal factors in the world, but as actors and agents of cultural change.

Ethical naturalism ignores the role of human beings in moral discourse and culture, or, more

importantly, it ignores our moral responsibility as contributors to cultural practices. Ethical naturalism

has not changed much since Aristotle. The notion is that we simply find ourselves in a certain social

circumstance, and that after having been fortunate enough to be raised in a certain way, we can reflect

upon the causes in our life that give us the convictions we have. The notion that we have an important

role as radical critics of cultural practices has no place to stand on in the Ethical Naturalist’s picture,

for human beings on this model inherited from Aristotle and rearticulated in Boyd and McDowell, are

simply functions of their environment. A human being can be a mild critic of their culture, but not so

radical that they can be thought to reasonably recommend a revolution in ways of life and linguistic

practices for any such recommendations would be judged as strictly speaking false according to what

their terms refer to. That the Ethical Naturalist is causally responsible for their environment, culture

and language seems to be rarely recognized. Thus, if lots of people own slaves in one’s society, then

slave owning is just fine for the naturalist such as Aristotle. If we live in a society that subjects

animals to untold levels of cruelty in slaughter houses and everyone treats them as food, then

goodness has to do with a homeostatic cluster property concerned with human thriving according to

Boyd. The passivity of human beings on this model shows up in the blatant chauvinism of Ethical

Naturalism. According to the Ethical Naturalist, for another culture to have moral convictions, they

must have convictions that refer to what the Ethical Naturalist themselves refer to in their culture with

moral terminology. Thus, the notion that another culture could have moral convictions but not refer to

209
what we refer to, or that they should reject the Aristotelian notion that ethics is simply what we learn

after the right type of upbringing, cannot even be contemplated. But this is because the Ethical

Naturalist cannot understand themselves as agents who continue to have ethical convictions as the

natural referents of their judgments change. Indeed, I think the only explanation here is that putting

all the stock in natural referents as causes of one’s moral convictions lets one off the hook for being

responsible for them, individually, and as a member of a society whose cultural practices are radically

contingent. Aristotle could thus say with a straight face that slavery is just fine: he didn’t invent it,

he’s simply calling it as he sees it.

One might think that the justification for viewing nature as placing constraints on our

concepts is that they must be responsive to naturally selective pressures. All societies, one might

think, must have a concept of human thriving that is their ethical concept. This, however, is very ill

informed.

In the Indian philosophical tradition, for instance, there is a term “artha” that designates

material prosperity (making money, having a materially comfortable life, social power and prestige)

and this has traditionally been distinguished from “dharma,” which usually articulates duties, virtues,

matters of sacrifice, cosmic order, integrity and principle—more often than not, this term was

intended to refer to matters of an explicitly non-anthropocentric nature (matters like friendship and

medical care did not at all fall under this highly deontic term). Traditionally, artha was thought to be

a sphere of axiological concern that had to be constrained by the considerations of dharma. Relevant

experts on dharma through Indian history never held that dharma was substantially concerned with

human thriving only and often they thought it involves a substantial amount of self-sacrifice and

concern for non-human beings. The notion that “dharma” ended up picking out Boyd’s homeostatic

cluster property concerned with human thriving over time is very unlikely for the relevant experts and

the bulk of the judgments in the Indian intellectual tradition that employed “dharma” were explicitly

concerned with matters that were contrary to human thriving. Even “artha” does not plausibly refer to

210
the homeostatic cluster property as Boyd understands it in the Indian tradition. Experts on artha

(such as Kautilya and his Artha stra) directed kings to undertake programs of realpolitik to secure

the welfare of their kingdoms: acting according to artha involved sowing the seeds of discord among

one’s neighbours, evaluating relationships in terms of the power and advantage it gains one, and

ruling with an iron fist. (Little wonder that most thought that dharma had to constrain artha). Other

experts thought that artha was the result of Vedic sacrifices to the gods, and thus directed votaries to

perform the appropriate sacrifices. Artha involves besting one’s neighbours, to a certain extent. The

notion that one even has the type of social cohesion to bring about a successful reference of a

homeostatic cluster property à la Boyd in the Indian tradition is very doubtful and moreover the type

of property that might have been picked out by “artha” or “dharma” would have been quite different

from the broad homeostatic cluster property that Boyd envisions. No doubt, both dharma and artha

overlap in part with the anthropocentric property that the naturalist is interested in, but they singly and

jointly diverge in some respects as well.

But we might have some sympathy for the Ethical Naturalist. Her motivation in looking for

reference in the natural world to ground the objectivity of her cultural practices is to make sure that

her ethical convictions are not a mere whim, but in fact responsive to something mind independent.

However, if she cannot contemplate how, for instance, with terms with a different referent, the Jain

ascetic argued that the demands of morality include that one stop acting altogether so as to not harm

other creatures, then she has no perch to stand on to justify why her cultural practices are the best of

the lot.

What does provide us this perch? It has to be a semantics that abstracts from particulars such

that we can understand people as talking about the same thing, even if the natural referents of moral

judgments change according to cultural and philosophical contexts. The only semantics that will

allow for this is the text-type of philosophy. For only it understands meaning in terms of universal

and general theories (i.e., abstractions) that are forwarded in light of considerations that are thought to

211
be objectively persuasive.

Before we see how non-natural philosophical knowledge is possible, the aim is to make the

case for a more modest thesis, namely that we require QI to translate value and normative discourse.

In this section, we have contemplated translating texts of ethics as scientific texts or philosophical

texts. The scientific translation treats ethical terminology as types of rigid designators, while the

philosophical approach treats them as quasi-indexical articulators of theories with a world-to-theory

direction of fit. Of the two approaches, only the philosophical approach can truly be said to preserve

the evaluative and normative content of value and normative discourse for it treats ethical terms as

articulators of normative and evaluative theories (against which the contingencies of the world are to

be judged) and it is in light of this conceptualization of the text that the translator is to read the ST and

produce the appropriate TT, mediated through this normative theory. The scientific approach, in

contrast, conceives of such terms in a purely referential manner. Understanding the referent of such

terms may have normative implications for the persons in the SL community, but these implications

are not preserved in translation as considerations that are to impress the TT reader in the same

manner. In translating the philosophical text in light of theories with a world-to-theory direction of fit

or the relevant considerations presented as objectively persuasive, the normative content of the ST is

preserved in the TT as something that is supposed to equally impress both the ST and TT reader and

for the same reasons. There seems no way around it: the only way to preserve the normative and

evaluative content of a ST is through the text-type of philosophy.

IV.6. Native Linguistics, Intuition and Ideology

In providing this criticism of the literature on moral semantics and moral translation, I do not

wish to trivialize the admirable interest in the topic, and the seemingly tenacious effort of participants

to overcome the problem of cross-cultural translation of normative discourse. Much ink in philosophy

has been wasted on far more trivial topics. However, the prevailing mode of seeking out solutions to

philosophical problems in the Western tradition—the linguistic turn—is not equipped to deal with

212
this challenge. This may seem like a bold claim. However, if the arguments about translation and

semantics in chapter 2 and 3 are correct, we have little choice but to affirm this conclusion. And it is

with a great trepidation that I make such a pronouncement. However, it is not made lightly, nor

without an unusual expertise in the subject.

In addition to working on this very dissertation, which to my knowledge is the first

systematic work on the problem of translating philosophy and normative discourse as such, I cut my

philosophical teeth on this issue by overcoming a significant problem in a major area of cross-cultural

research, namely the problem of translating and understanding the history of Indian moral philosophy

(Ranganathan 2007a). I did this not squarely within the context of philosophy as an institution, but at

the crossroads of philosophy with several other disciplines, including Sanskritology, history,

linguistics, anthropology, religious studies, and archaeology—loosely called South Asian Studies. I

am also, to my knowledge, the only philosopher writing on moral semantics and the translation of

normative discourse who is also nominally a translator of normative discourse (Ranganathan 2007b).

Thus, when I claim that the linguistic turn lets us down when we want to solve the problem of

translating normative discourse, my conviction has not merely been formed from the comfort of my

arm chair.

As normative issues are philosophical, to translate normative discourse is to translate

philosophical discourse. This requires a certain expertise in the very institution of philosophy. All

philosophers have a certain practical authority in the institutional structure of philosophy, but few

make the very institution the topic of their study, and fewer still are likely to clearly observe the

textual structure of philosophy. Undertaking research in an interdiscipline provided me with a certain

perspective that was very helpful in my task of understanding the institutional norms of philosophical

translation.

One problem with most of us philosophers is that we spend all of our research time around

other philosophers. It is very difficult to see the forest for the trees in this context. If one wishes to

213
understand the institution of philosophy, study it at its boundaries, where scholars from other

disciplines meet to study the same text. It is an incredibly lonely and enlightening experience. In my

case, I had no philosophical company, so to speak. But it was here that I began to rediscover what I

knew about philosophy, more clearly, for I appreciated my knowledge of the institution in contrast to

the institutional expectations of scholars from other disciplines. (None of the scholars who I studied

under were philosophers, though they all thought they had something to say about texts that are

philosophical. They were social historians, linguists and historians of ideas who, in some fleeting

moments, betrayed a philosophical sensitivity. But, all the same, their research was grounded not in

philosophy as such, but in philology and history.) Simple exercises like writing term papers for a

historian or philologist (even when writing on a philosophical figure or text) became a mind boggling

exercise that forced one to sharpen one’s knowledge of text-types. At any rate, the exercise, if

undertaken properly, has an anthropological dimension. One becomes a participant observer in

adjacent fields in the humanities and social sciences. One is able to at once have meaningful

conversations with people in other academic worlds while always being mindful of one’s status as

someone from a foreign land. And this experience as a type of outsider is very important for the topic

of normative semantics. It helps us avoid a type of ideological approach to the topic that characterizes

the current literature.

I made it my research project to come to deal with scholarly assessments of Indian ethics.

The dominant view was that Indian philosophers were not interested in ethics. I began to entertain the

hypothesis that this opinion was a result of bad translation.

As I first thought about the question, I thought that some headway could be made if I could

determine whether there was a right answer to what “ethics” and “moral” mean in English. I did

entertain the notion that if we were going to translate an SL into English, then we should be clear on

what it is that we mean by our English language terminology.

As I discussed this problem with scholars from many disciplines (and some of my old

214
philosophy professors) I hit upon an observation that has only been confirmed in the years that I have

“returned” to philosophy. If one puts the question of what the definition or meaning of a moral

concept is to the typical person, philosopher, scholar, or lay person on the street, you will receive a

host of different responses, and each (further inspection reveals) is tightly interwoven with the

person’s substantive convictions in ethics, and a host of other controversial philosophical issues . And

in each case, people tend with great confidence to talk about what “we” mean by a certain term.

Indeed, recent analytic philosophy is replete with rhetorical appeals to an unidentified “we” or “us”.

This is of course terribly problematic for it is factually mistaken. Philosophical, ethnic, and cultural

diversity in societies have been the norm since the beginning of our recorded history. Rare is the case

of geographically isolated cultures that present the type of homogeneity that linguistically-oriented

philosophers presume in their linguistic analyses.

I am happy to note that since I began my study on the problem of translating normative

discourse, many analytic philosophers have begun to question the ability of philosophers to speak to

what people at large mean by concepts. The attack on this issue has come via the connected question

of whether intuitions—spontaneous, seemings that present themselves to us with a certain prima facie

foundational authority—can form any basis for conceptual analysis and even the related question of

whether linguistics can serve as a basis for philosophy.86 However, there are a few factors that

contribute to the illusion of a type of homogenous linguistic base that the analytic philosopher can

study, thus reinforcing this unfortunate practice.

First, the naked truth is that analytic philosophy is not known for its ethnic and cultural

diversity. It is dominated by people of European heritage. The common heritage does in no way rule

out the incredible diversity of views that one can find amongst analytic philosophers, but it certainly

attenuates potential diversity.87 Secondly, the peer review journal structure, with its emphasis on short

arguments, sustains a mode of inquiry that is parasitic upon what are considered accomplishments in

the literature—accomplishments that usually appeal to “intuitions” that are ethnically conditioned.

215
Since we live in a world where academics must publish or perish, this second requirement places a

great pressure on new-comers to appeal to such intuitions. Those who cannot bring themselves to

feign intuitions they do not have will find it very difficult to get published, and may thus not continue

to be one of the club. These factors manage to systematically maintain a relative level of

homogeneity in the answers analytic philosophers are apt to give to questions of conceptual or

linguistic analysis. Third, the prevalent ideology in analytic philosophy that philosophical problems

are all of them abstract and solvable by logical argument provides a cover for intuition mongering to

flourish and reinforces the relative hegemony in the field, to the extent that few stop to question

whether the actual mode of argumentation defers to data that is deeply contingent.

Relative as it is, it is still a type of homogeneity, and this homogeneity is an obstacle to our

ability to translate normative discourse, if philosophers’ intuitions of what normative terms in English

mean are to be the yard stick by which we determine normative discourse translation. The reason

should be fairly obvious. What we are trying to do when we translate normative discourse is

understand how others, who do not necessarily share our language or culture, but yet inhabit the same

universe as we do, can value it differently from “us”. If our accounts of the meaning of value terms

and concepts are the devices that the alien view is to even be judged meaningful or intelligible by,

then it is difficult indeed to understand how the alien could ever be understood domestically except

by a huge distortion.

Henry Jackman is correct that typically:

the types of intuitions involved when philosophers engage in, say, semantic theory are not
intuitions to the effect that something like the descriptive theory of reference is correct.
Rather, they are our intuitions about how our semantic vocabulary should be applied in
particular (often counterfactual) instances.(Jackman 2005, 373)

But the case of translation is indeed the case of how the translator should use terms in counterfactual

situations (i.e., texts not written by us). If the people we are translating do not share our intuitions

216
then the reliance on our intuitions will be very faulty. And indeed, in the case of the translation of

value discourse, we should expect that many of those we translate will not share our intuitions about

how to use value terms, which is to say that people disagree on values because they have, among

other things, divergent intuitions. But then our reliance on our intuitions to translate normative

discourse will certainly produce bad translations.

But should this matter? Should we care that there are a diversity of intuitions? Is this simply a

political issue, and not one of concern for semantics? Why can’t it be the case that most intuitions are

wrong? The semanticist, or the philosopher of language, in contrast, gets it right because their

intuitions are trained.

The diversity of intuitions about the meaning of concepts and terms in a language suggests

that we are actually not very good at being objective about what a word or term means—and, I might

add, it also suggests that getting meaning right is not the point of language. This view is in part

corroborated by the excellent work of the linguistic anthropologist and psychologist, Michael

Silverstein. His research reveals that speakers of a language are often very poor at articulating the

structure of their language (Silverstein 2001 [1981]). A key feature of Silverstein’s findings is that

language use is a dialectical interplay where interlocutors attempt to get others to see the world their

way, and they are so enmeshed in this activity that they fail to recognize what he calls the “meta-

pragmatics” at work, namely the use of language to get others to use language one’s own way

(Silverstein 1993).

According to Silverstein, when we get it wrong, and provide definitions of terms that suit our

particular world view, we are engaging in what Silverstein calls “linguistic ideology.” But, on

Silverstein’s account, this occurs any time speakers attempt to rationalize or justify perceived

language structure and use (Silverstein 1979, 193). To get language right, on Silverstein’s account, is

to point out the pragmatic structures of language use.

I think that Silverstein’s account is right, but only part of the picture. The view that a word

217
definitely has a determinate meaning is always incomplete: it focuses on some semantic features of a

word and attempts to rule out contrary accounts, and for this reason is “ideological”, which is to say,

it is one sided. But it is also ideological in another respect. In specifying that a word has a certain

meaning, we are specifying that people should view the world in a certain way. And thus, it is

ideological in so far as it is expressive of our desire to share our perspective with others. But it is also

more than that. It is also a contribution to the semantics of a language.

Here I find myself in agreement on one point with Jackman. In his paper “Intuitions And

Semantic Theory” (2005), Jackman responds to the charge that accounts of semantics that make the

meaning of our concepts determined by external factors, such as environment, or our society, render

intuitions superfluous in an account of meaning. He argues that this conclusion only follows if one

assumes that intuitions are expressive of meaning. Rather, his view is that they are, at least in part,

constitutive of our concepts. The view I have been defending is largely consistent with this view, if the

typical concepts we are talking about are understood as relativized to a language. Indeed, the intuitions

of philosophers, and language users at large, are constitutive of the meaningfulness of a language for

no other reason than such intuitions influence our use of such words. So, indeed, our very usage—

including claims about the meaning of words—contributes to the polysemy of language.

None of this refutes the fact that there are regularities in language use in communities, and that

such regularities are important. Indeed, text-type institutions play a large role in regulating language

use across communities and even cultures to make the reading and writing of texts practical. Nor does

this argument refute the scientific probability that our language usage is either responsive to selective

pressures on a society, or natural factors in the environment of a culture. Indeed, we would even have

reason to think that when we use words to pick out such factors, we are speaking truthfully. However,

the question is whether such selective pressure or natural factors can be said to explain everything in

the way of semantics. Can they account for all of our intuitions about the meaning of words?

218
 I do not think so. Here’s a reason to think that no such factor could be determinative. The

factors we are considering are those that are in some sense imminent in a culture. They are either the

referents of words, their typical extensions, syntax, or selective pressures: these are the usual suspects

in an explanation of the meaning of a word in a linguistic practice. These may account for many of the

intuitions that people have about word meaning, but they cannot explain intuitions that people have

that come about through a process of cultural exchange, namely, translation. The introduction of texts

into a culture produces a whole new fountain for word meaning intuitions, which is not in any way

directly reducible to the intra-cultural factors we have been considering. Indeed, such novel intuitions

may even come about through the introduction of mistranslated texts that are mistranslated according

to the standards of meaning prior to their appearance.

Our traditional picture of language that is passed down to us in the philosophical tradition is

the picture of the language of the islanders, cut off from the outside world. But the norm in our history

is the opposite: multiculturalism, pluralism, migration, cultural exchange, have been constant factors in

the development of human language. When one takes into consideration all of these real-life factors, I

think it becomes relatively difficult to suggest that there are “central” usages or referents of words, and

that deviations can always be understood in relationship to this centre. Forms of life have almost never

existed except as dynamic, evolving, processes. The centres of meaning are constantly contested, with

the result that there are often multiple centres of word usage.

This description of the productivity and creativity of language might suggest that what we

really have in language is a split between the semantic and the pragmatic. However, as noted in chapter

3, the efforts of the formalist to draw a distinction between semantics and pragmatics of the meaning of

expressions in a language fails. If translation preserves meaning, as I argued in chapter 3, and if any

semantic modality of an expression of a language can be translated in the appropriate text-type, as it

can, then the distinction between semantics and pragmatics as traditionally drawn (as something that

divides language into two) is mistaken. Moreover, if there are in reality several centres anchoring

219
language usage, within a linguistic practice, then the distinction between what is semantically central

and what is pragmatically derivative is not easily drawn.

The formalist project thus of attempting to pin down the meaning of words in a language can

only be understood as an effort to systematize something that is in reality in flux. I’m not sure why

anyone tries, except perhaps for the satisfaction one gains from showing that certain aspects of

language use can be technically represented. But in our concern to deal with meaning, certainly in

relationship to translation, all such efforts are superfluous.88

In recognizing the indeterminate nature of linguistic meaning, I am not recommending that we

be set on a sea of post-modern uncertainty. It is a part of the view that I am advocating that determinate

meaning and knowledge is institutional in nature. Thus, the view I am advocating is that we shift our

focus from language to institutional considerations. Part of the reason why such a shift is called for is

that it is not profitable to look to language to be the lone, bedrock of knowledge. It is subject to

constant change. Texts, in contrast, are corpuscular limits to the semantic shifts and change in meaning

that are a constant fact of real language. It is here that the type of stability we require for knowledge to

flourish is to be found for this is the space where institutions can flourish. All important knowledge is

after all based upon a specific type of text: science, mathematics, philosophy, literature, are each

textual endeavours, and in part why they can be the foundations for knowledge is that they are defined

by their translatability relative to a text-type. Languages change, texts don’t.

If what I say is correct, that our major epistemic achievements are textual in nature, and not

reliant upon the contingencies of language use in a culture, then we must ask hard questions as to why

the linguistic turn has occupied our interest for so long.

If we idealize linguistic communities as culturally cohesive wholes without diversity, then

there may indeed be some possibility for language to be governed by convergent intuitions that might

form the basis for shared meanings that characterize or define a language. But if culture is already

always a ground for contestation, and for inter-subjective negotiation of an ideological variety, the very

220
expectation that we should have determinate meanings in a language becomes under-motivated. The

reality is, plurality and diversity has been the norm through human history. Bigotry is simply a

pathological reaction to this diversity. 89 Any attempt to render language semantically determinate by

understanding meaning in terms of a privileged referent is also, I think, a pathological reaction to

diversity. Given our various text-type institutions, we do not require that language as such will

underwrite objectivity. We have institutional criteria based upon text-types that transcend cultures that

can do that for us.

IV.7. Conclusion
This chapter focused largely upon naturalistic accounts of moral semantics. The star of the

discussion and of the literature is Richard Boyd’s referential account of moral realism. Boyd’s initial

proposal that one could provide a causal account of reference for “good” triggered a wave of articles

from Horgan and Timmons pointing out the relativistic implications of Boyd’s proposal. This in turn

had spawned a flurry of responses to the problem as set out by Horgan and Timmons. But what the

entire literature shares in common is the linguistic account of meaning, which holds that languages

are characterized by determinate meanings, and that translation is a matter of exchanging word for

word, sentence for sentence, across languages on the basis of cross-linguistic synonymy. The reality

is that the radical differences among cultures makes the linguistic account of meaning relativistic, and

the linguistic account of translation unworkable, leading to a host of problems in translation theory,

such as Quine’s indeterminacy of translation thesis.

Horgan and Timmons have accused Ethical Naturalists of a type of chauvinism in ethics (cf.

Horgan and Timmons 1996). But their account of moral semantics is just as chauvinistic, for it

assumes that for any society to have resources to articulate an ethical perspective, it must be because

the expressions in their language are characterized by an internalist semantics. The reality of radical

cultural diversity makes any hopes to underwrite translation by cross-linguistic synonymy unrealistic.

But, as well, the radical diversity within societies makes the notion that social cohesion can

221
underwrite objectivity in ethical matters doubtful.

Ethical naturalists, and their critics, have thus failed to be naturalistic in one important

respect. They have not grounded their proposals on any sensitivity to the social scientific facts, and it

is very obvious that none of the authors writing on this topic, quoted here, have spent much time

seriously thinking about the challenges that translators face.

Ethical Naturalism might seem to have in its favour that it provides a naturalistic explanation

of the beliefs that people have. However, in the case of many beliefs of people, particularly evaluative

beliefs, the notion that the natural world causes members of a society to have such beliefs is not

obvious. Natural factors underdetermine culture difference and thus we must look to the philosophical

convictions of persons in a society to account for their cultural choices. We might think, for instance,

that the likes and dislikes of persons in a society might account for such factors, but human beings

have a remarkable ability to organize their life according to non-hedonic principles, and the only

general explanation that will account for the various choices that constitute a culture, is philosophical.

In other words, it is the metaphysical, epistemic, and broadly axiological considerations of persons in

a society (their ethical, aesthetic and soteriological convictions) that explain why a culture is the way

it is and not another way.

Accessing the philosophical convictions of cultures and societies is not easy. It takes a

tremendous amount of interdisciplinary work, but it also involves a sensitivity to the semantics of the

philosophical text-type. The philosophical text-type allows us this important role in cross-cultural

research because it concerns a constellation of meanings that abstract from empirical particulars, and

concerns theories of a universal and general nature that are presented on grounds that are thought to

be objectively persuasive. Philosophy thus lifts cultural contingencies into a common forum for

discussion in a manner that scientific investigation of cultural phenomena cannot, as it is always tied

to the empirical particulars.

QI in particular can help us translate normative discourse both because it is an account of the

222
text-type of philosophy, but also because it is a formalization of the anatomy of the philosophical text.

It understands moral meaning in a manner that is not reducible to cultural contingencies. Indeed, on

its lights, the only way that one can have anything like determinate moral meaning is by recognizing

that cultures will vary and that the semantic profiles of languages will differ (and be internally quite

indeterminate as well), and that determinate moral meaning is an aspect of a certain type of text. If

objectivity in ethics is to be found anywhere, it can only be on the basis of such a trans-cultural and

trans-linguistic account of moral meaning. Attempting to base ethical objectivity in nature amounts to

attempting to base ethical objectivity in cultural contingencies.

But, for all of this, it may still be possible to entertain a type of scepticism about the

institutional foundations of philosophy, or the notion that there is such a thing as the text-type of

philosophy. The applicability of the text-type of philosophy to understanding diverse intellectual

traditions is perhaps the best proof, in my mind, that it actually tracks something deeply profound and

important to various cultures. As I have argued in this chapter, it cannot be extricated from the very

notion of normative discourse. So, ironically, I think recognizing something seemingly contingent, as

the institution of philosophy, leads us to certain conceptual facts about normative discourse in

translation that cannot be had via linguistic analysis. But for all of this one might be sceptical that

today there is any such single institution of philosophy. Certainly, we can notice riffs, fractures, and

vague border areas in philosophy. But we have the modern academy: an international network of

scholars in various disciplines, concerned with their own types of texts. My suggestion is that

philosophy is one such text-type institution. Scholars of philosophy engage in research at the

international level: they meet with their colleagues at conferences, keep in touch by email, and train

students in graduate programs that go on to populate universities the world over. How people

understand philosophy the world over is remarkably similar. The topics and substance of debate

varies, but the underlying text-type of philosophy as one concerned with universal and general

theories on various topics in light of reasons that are thought to be objectively persuasive is pervasive.

223
The reason that it can be so pervasive is that it at once abstracts from content but also allows for

coordinated disagreement between differing perspectives. The very character of the discourse

provides cohesion to what would otherwise be unconnected chatter. It is easy not to appreciate this if

one remains cloistered in one’s own narrow corner of the Earth, unconcerned with how philosophy is

done half way around the world. But, for such a large institution, riffs are possible. The reality of such

a text-type as a usual template for understanding texts and translating them cannot be doubted.

What might worry us is that the institutional foundation for such a text-type to be used in

scholarship to produce accurate translations and resolve problems of philosophical importance might

be less than secure. This would be cause for concern. If philosophy as a textual discipline is

fracturing, and if there is a tendency to fragment the philosophical text-type into distinct and

unrelated concerns, then we will have no one to blame but ourselves. The move from the linguistic to

the institutional is, in my view, the move from the adolescent in semantics to the adult. For the

adolescent, the peer group and personal identity is everything. The linguistic turn has made these

adolescent concerns into semantic virtues. The adult, in contrast, is grateful for what useful

institutions posterity has left her and takes responsibility for the future of the institutions, lest things

stay the same (which would be bad enough) or heaven forbid, things get worse. The mature adult

understands herself both in terms of her autonomy but also by her responsibility not simply for the

narrow ends of her tribe, but for the fate of the world, for no one but her, and other adults, are at the

steering wheel. If philosophy is fracturing, it is ours to fix. The sooner the better, for this is the only

way that we can resolve moral controversies and gain moral knowledge. For, as we have seen, science

cannot provide a canonical account of moral meaning or translation. This is left to the text-type of

philosophy.

224
 V. Expressivism, Translation and Normative

Disagreement

In the previous chapter we learned that naturalistic theories of moral semantics are incapable

of showing how we can translate normative or evaluative discourse. The traditional criticism of such

theories, particularly Boyd-style Non-Analytic Naturalism, is that it cannot overcome objections from

Terrence Horgan and Mark Timmons’ Twin Moral Earth thought experiment. Horgan and Timmons

believe that their thought experiment shows that only an internalist account of moral semantics, that

understands moral semantics in light of the expressive power of moral meaning, can explain how

persons from diverse societies with differing substantive convictions could communicate with each

other (Horgan and Timmons 1992a, 1991, 1992b, 1996). However, both Timmons and Horgan, as

well as Boyd’s Non-Analytic Naturalism, along with the entire literature that have been spawned by

Timmons and Horgan’s criticism of Non-Analytic Naturalism, assume the linguistic account of

translation, where translation is thought of as a word for word exchange across languages on the basis

of cross-linguistic synonymy. But we know from chapters 2 and 3 that this is simply not consonant

with the challenges of translation.

Horgan and Timmons’ proposed internalist option to overcome the problems of translation is

a new version of an old standard in the moral semantics literature, namely Expressivism. In this

chapter I take a closer look at the possibilities of Expressivism and ask whether it can explain how we

can translate normative and evaluative discourse across languages.

The exploration will proceed in six parts. I shall first set out the doctrine of Expressivism. In

the second and third sections, I will review the objections and concerns about translation from the

literature. In the fourth section I set out TTS, argued for at length in chapter 3, as the response to

worries about translational determinacy. In the fifth section I review Expressivism in its standard

225
form and note that, like Non-Analytic Naturalism, it fails as an account of normative or evaluative

discourse translation as it is not a textual account of meaning, but a linguistic account. In the sixth

section, I will consider revised versions of Expressivism, designed to meet the requirements of

translation. I conclude that Expressivism, in both its native and revised forms, is unable to help us

translate normative discourse and that as a direct result it is unable to properly account for normative

disagreement. But the failure of Expressivism is very instructive. The reason it fails is because it is

dismissive of the philosophical reasons that authors have for their normative claims as irrelevant to

normative semantics. What this proves is that normative and evaluative discourse is a species of

philosophical discourse. The only account of normative discourse and normative discourse translation

in the running that recognizes this is QI. QI, in contrast to Expressivism, does not treat the

philosophical reasons that authors have for their normative convictions as irrelevant to normative

semantics, and is thus able to produce accurate translations of normative discourse.

V.1. Expressivism
Expressivism in its narrow presentation is an account of the semantics of moral sentences. In

its general presentation, it is an account of the semantics of normative sentences. It is a descendent of

a tradition of Non-Cognitivism, though recently there are cognitivist versions as well. There are also

hybrid versions that take normative sentences to be analyzable into descriptive and expressive

content. Let us look at non-hybrid versions first.

Non-Cognitivism in metaethics was the view that atomic moral sentences (and normative

sentences on the whole) are not truth apt. The earliest position in this tradition is Emotivism (Ayer

1946 [1936], 107; Stevenson 1944). Many later proponents of descendent positions, such as

Prescriptivism (Hare 1993, 451) and Expressivism (Blackburn 2000 [1998], 85; Gibbard 2003b, 183),

betray some ambivalence towards the label of “Non-Cognitivism”.90 Simon Blackburn (2000 [1998],

79) and Allan Gibbard (2003b, 18), the leading proponents of Expressivism, want to claim that moral

226
judgments can be true in a minimal sense, without completely disavowing their Non-Cognitivist

heritage.

Truth, according to minimalism, is a metalinguistic device that stands in for the assertion of

sentences. Thus, minimalism is committed to the biconditional “P” is true iff P; moreover,

minimalism holds that the various phrases associated with weightier conceptions of truth, such as

those associated with the correspondence theory, can be claimed by the minimalist as mere emphatic

devices that stand in for assertion. For the minimalist, “P” is true iff “P” corresponds to the facts is

to say nothing more than “P” is true iff things are as “P” says they are. Since P is how “P” says they

are, we can substitute in “P” for the phrase things are as “P” says they are. We thus arrive back at the

original disquotational schema: “P” is true if and only if P (Wright 1992, 24-27).

For Gibbard and Blackburn, the minimal truth aptitude of normative sentences falls short of

rendering normative judgments into beliefs in one sense, if a successful belief’s defining feature is its

world-guided, sui generis content (natural or non-natural) that could provide a non-personal

grounding for its truth or falsity. Beliefs in a traditional, empiricist sense can have that type of sui

generis representational content. For the Expressivist, normative judgments (though we may call

them “beliefs”) do not. In normative matters, talk of moral properties and correspondence to moral

facts are “constructed precisely in order to reflect our concerns” (Blackburn 2000 [1998], 80).91 What

type of thing, then, is a normative judgment? It is attitudinal (cf. Blackburn 2000 [1998], 31, 65, 70,

280), and it relates to straight attitudes as beliefs do to facts (Gibbard 2003b, 81; cf. Blackburn 2000

[1998], 79). By extension, we can say that normative sentences relate to attitudinal judgments as

descriptive sentences relate to beliefs. Following Mark Schroeder, let us call this the parity thesis. 92

As normative judgments are not naturalistic descriptions of any kind (for Expressivists), their

linguistic correlates (normative sentences) are not supposed to be descriptions of mental states or

attitudes.93 Rather, they are vocalizations of our minds (Blackburn 2000 [1998], 50; Gibbard 1990, 7-

8). This is the doctrine of mentalism.94

227
 A final important feature of non-cognitive Expressivism is its commitment to Naturalism.

Expressivists are averse to calling upon non-natural properties in an account of normative semantics.

Expressivism, and its precursors, Emotivism and Prescriptivism, are attempts to capture the

seemingly open nature of normative discourse without resorting to Moorean extravagances (Gibbard

2003b, 192).

Cognitive Expressivism (Timmons 1998; Horgan and Timmons 2000b, 2006), a late-comer to

the debate, has sparked some controversy as commentators try to understand its full import (cf., van

Roojen 2001; Dreier 2004). It is the view that moral judgments are nondescriptive, though they are

genuine beliefs. Normative concepts on this account are not semantically reducible to attitudes, norms

or desires.95 One way to understand the motive for this peculiar brand of Expressivism is to take the

minimalist conception of truth seriously (Timmons 1998, 107; Horgan and Timmons 2006 fn. 16).

The cognitive nature of moral judgments is the fact that they are psychological correlates to perfectly

assertable and truth apt sentences à la minimalism without thereby relying upon descriptivist

underpinnings that would contradict the essential, moral irrealism of Expressivism (cf. Horgan and

Timmons 2006). Normative concepts, on this view, are sui generis, but they lack sui generis

descriptive content that could vouch for their truth or falsity.

Call the versions of Expressivism sketched so far pure Expressivism. In contrast, there are a

number of recent views, prefigured by the work of C.L. Stevenson (1944), that present hybrid

accounts. Normative sentences on these views stand for both a non-cognitive attitude or desire, along

with a propositional, descriptive belief (Alm 2000; Copp 2001; Ridge 2006).

V.2. Indeterminacy, Impossibility, and Non-Equivalence of Translation

It might seem that Expressivism can help translate normative disagreement across-linguistic

boundaries. All we would need to do, it seems, is locate linguistic devices in a second language that

permit speakers to voice their mind, just as, ex hypothesi, our normative terms allow us to express our

minds. Once we have located such terminological resources, and once we have some ability to

228
translate sentences formed with such terms, we can then explain disagreement between us and the

member of the alien linguistic community in the same way that we explain it locally, as an opposition

of outlooks or attitudes.96

Before we can evaluate Expressivism’s success in helping us translate normative discourse,

we need to identify a theory of translation and semantics that will allow us to overcome traditional

objections to translation. Here, I shall address four such objections. To understand their force, we

need to appreciate the importance of semantic equivalence in translation.

Translation is a semiotic process of conversion, which takes an original or source text (ST)

written in a source medium (usually a language) and produces a translation or target text (TT) in a

target medium, usually a distinct language.97 The process aims at success, and success in translation

consists in the creation of a TT with the same meaning as the corresponding ST. Semantic divergence

between a TT and ST is grounds for criticizing a translation. Thus, for instance, if we attempt to

translate the Bible and produce through this process the Communist Manifesto, something would have

gone wrong, for prima facie, the semantic content and purport of these texts —what they mean—are

clearly divergent.

While the practice of translation is normatively regulated by the notion of semantic

equivalence, philosophers and translation theorists have brought equivalence in translation into

question.

W.V.O. Quine has famously argued that translation is indeterminate. According to this thesis,

linguists can, based on the totality of all possible empirical evidence, arrive at translation manuals that

yield translations of sentences that are contradictory with each other, through no methodological fault

of their own. This is not merely the view that translations are underdetermined by empirical

observation, but that there may be cases where contradictory translations are equally licensed by the

evidence and no empirical observation can decide between the two (Quine 1960, 27). Such possible

contradictory but irresolvable translations bring into question the ideal of equivalence in translation.

229
 The Continental philosopher of language, Jacques Derrida, has argued for the impossibility of

semantic equivalence in translation, which I will call the lost in translation argument. Derrida

illustrates his argument by examining Plato’s use of the Greek word “pharmakon” in the Phaedrus

(Derrida 1981 [1967], 98). The term ambiguously stands for many concepts, including MEDICINE and

POISON. Plato skilfully exploits the ambiguity of the term, allowing it to self-referentially function as a

symbol of its own ambiguity, in order to advance the plot of his dialogue Any attempt to translate this

term into another language will at best capture some but not all of the significances of the original

“pharmakon.” Thus no translation of the Phaedrus will be able to retain the full meaning of the

original. Translations cannot be equivalent.98

The translation theorist Mary Snell-Hornby argues that the idea of equivalence in translation

is deeply misguided, and we need look no further than the English word “equivalence” and its

putative German counterpart “äquivalenz” to appreciate how important semantic differences lie

underneath the surface of apparent cross-linguistic synonymy (Snell-Hornby 1988, 16-17).

Then there is the problem of grammatical differences between languages. Languages differ

according to whether they are gendered, how numbers are expressed (whether they have singular,

dual, plural cases, for instance) whether verbal copulas are obligatory, their tenses, and much more.

Accommodating obligatory features of a target language grammar can introduce semantic elements in

a translation that are absent from the original, and it can also remove elements in the process.

Translational equivalence seems threatened by grammar alone.

V.3. Indeterminacy, Translation, Languages and Texts

There are important differences between Quine’s criticism of translation and the one offered

by Derrida and Snell-Hornby. Most importantly, Quine claims that the problem he identifies revolves

around translating sentences, not words (cf. Quine 1996 [1992], 452, 1960, 27).99 Moreover, his

criticism revolves around the indeterminacy of equivalence in translation, not its impossibility.

However, on closer inspection, Quine’s complaint is remarkably similar to Derrida’s and Snell-

230
Hornby’s. To understand how Quine’s concern, or at least one aspect of it, is the same as Derrida’s

and Snell-Hornby’s, we need to look to Quine’s comments about the unsurprising nature of his

findings. According to Quine, the problem arises because of a “conflict of parts seen without the

wholes.” He continues: “The principle of indeterminacy of translation requires notice just because

translation proceeds little by little and sentences are thought of as conveying meanings severally”

(Quine 1960, 78-9). If the parts in question are sentences of languages, and the wholes are the

languages themselves, then the problem of the indeterminacy of translation is the problem of finding

linguistic equivalents for items that have their full range of significance within their native

language—precisely the problem that Derrida and Snell-Hornby catch onto in their focus upon

semantic individuality of terms relative to their native languages.100 Derrida concludes from this that

translation is strictly speaking impossible (Derrida 1981 [1967], 99; Derrida and McDonald 1985, 99-

100). Quine concludes that it is indeterminate because attempts to find equivalents of words and

sentences across languages can be licensed by pragmatic considerations (what Quine calls “analytical

hypotheses,” cf. Quine 1960, 68), and pragmatically justified concordances may lead to conflicts and

perhaps even contradictory translations. Either way, what we have is a concern and disappointment

over equivalence in translation that characterizes not only the major philosophical criticisms of

translation, but also much work in translation studies.101

It would be a grave mistake to conclude from these considerations that translation is

indeterminate. The arguments presented against the possibility of determinacy and semantic

equivalence in translation rest upon a simple but incredible category mistake about what translation is

concerned with.

V.4. Text-Types and Translation

The asymmetry of the full range of semantic properties of words, sentences and syntactic

devices of different languages (such as the asymmetry between “equivalence” and “äquivalenz”) is

undeniable, and within the same language (between contrasts such as “synonymy” and “semantic

231
equivalence”, “meaning” and “semantics,” or even “equality” and “equivalence”) is difficult to deny

as well. But it is only a problem if we think that the translator’s task is to treat a language as the

object to be translated. But this is to commit a major category mistake. Real life translators do not,

nor have they ever, endeavoured such an unlimitable task. Rather, what translators translate are texts.

Texts are not the same as languages. They may be written in languages, though they could be

composed in other semiotic systems, like icons or ideograms. Texts are finite in size. They have

beginnings and ends, and some type of author. Most importantly, texts come in types, defined by

textual institutions.

A text-type is like a genre except it pertains to texts and has widespread institutional support.

Examples of text-types include the natural scientific text, the poem, the novel, and philosophy. In

each case the translator’s job is different and is guided by an account of what is vital to the translation

process, relative to a text-type. A translator can never preserve all the semantic features of a ST in a

TT, but that is not the goal of successful translation. Rather, the goal of successful translation is to

preserve in a TT just those features of a ST judged important by the appropriate text-type. Thus, it is

no conceptual bar to producing successful translations that the materials of different languages are

constrained by cultural peculiarities of idiom, imagery and connotation, grammatical differences, and

differences of terminological resources. Translationally, the significance of the word and sentence

comprising a text is not at all judged as though they make their contribution severally or little by little.

Rather, to properly translate is to understand that words and sentences are in the abstract full of

semantic possibilities and that they are incorporated into texts according to narrow considerations of

text-type. Thus, a TT may have many semantic features not found in a ST, and vice versa, but so long

as they have the same feature judged relative to a text-type, the process of translation has been

successful relative to that type.

For instance, the translator of science is not concerned to reproduce the rhyme, rhythm and

drama of a laboratory report. These are inessential to the text-type of science and may be omitted

232
from a TT without compromising the accuracy of the translation. The translation of science must in

contrast retain the ontological, methodological and theoretical commitments of the ST, along with its

observational findings. The case of poetry is quite different. Here, the various aesthetic features of a

ST are the crucial features to translate. And since many of these features are constituted by cultural

peculiarities of the source language, many translation theorists have argued that the appropriate

translation methodology is to produce a TT that calls upon the target language’s resources in a

manner that is functionally equivalent to the aesthetic devices of the ST. Thus, puns and rhyme are

not to be translated truth-conditionally, but dynamically or functionally, so that a similar aesthetic

effect can be reproduced in the TT, without regard for the truth-functional meaning of such

devices.102

The text-type of philosophy concerns the investigation, argument, debate, criticism,

development and application of theories of a general and universal nature, canonically articulated

with ordinary devices (like “real,” “knowledge,” “meaning,” “right,” and “good”) on the basis of

considerations that aim at being objectively persuasive (i.e., the considerations are not presented as an

inventory of mere subjective opinions). Philosophical texts may make extensive use of technical

terminology or they may be written in a cryptic fashion, but it seems that our confidence in

classifying such texts as philosophy depends on our ability to translate such texts into nontechnical

talk of reality, knowledge, right or wrong, and meaning (i.e., metaphysics, epistemology, axiology

and semantics for instance). The task of the translator of philosophy is thus to preserve the features of

the ST that advance the philosopher’s project relative to these text-type-theoretic goals of philosophy.

Thus, dialectical structure, argument and secondary literary devices (such as character, dialogue plot,

metaphor and humour) are to be preserved to the extent that they advance these goals.

The notion that translation must proceed according to type has been recognized for some time

in the translation studies literature. Today, many expert or “technical” specializations exist in

233
translation studies to address the specific requirements of translating certain types of texts, such as

scientific texts and legal documents.103

What has not been generally recognized in the translation studies literature is the surprising

semantic consequences of recognizing the text-type-theoretic nature of translation, when combined

with an uncontroversial thesis cited earlier as regulative of the process of translation. Recall the first

principle:

translation preserves meaning.

Now add the following uncontroversial principle:

(1) translation must proceed according to text-type.

To affirm these two theses is to be committed to the idea that the meanings preserved in

genuine cases of translation are those identified by a text-type. This commitment has powerful

consequences.

First, since translations that are produced according to a text-type save specific semantic

modalities of a ST in a TT—those specified by a text-type—it follows that such translations are

semantically equivalent relative to a text-type.

Second, translation units (which can be as small as words and as large as passages) that are

deemed essential to translating a text relative to a text-type are semantically equivalent to their TT

counterparts. Thus, for instance, a German passage in Kant’s first Critique that is reconstituted in an

English TT by accommodating the rules of grammatical and idiomatic propriety of English can,

holophrastically, be taken to be equivalent in meaning to the original German passage, though the

constituents of the passage may, when considered on their own, be different in style, punctuation and

even dialectical sequence.

Call the implications of theses (1) and (2) Text-Type Semantics (TTS). TTS is a theory of

meaning that takes texts of definite types to be the best exemplars of bearers of meaning. TTS

234
contrasts with linguistic conceptions of meaning that take terms and sentences, or whole languages, to

be the best exemplars of the bearers of meaning.

What does TTS say about determinacy of translation and meaning? If meaning is preserved

in translation, translation serves as a type of test of meaning, which separates meaning from semantic

noise relative to a text-type.104 Indeed, successful text-type translations can be easier to understand for

they separate out what is semantically essential about a text, relative to type, from what is relatively

inessential.105 But if translation can only be said to be genuine when it is guided by text-types, and if

text-types are only really applicable to texts, it follows that putative synonymy parings of words and

sentences, definitions or analyses of concepts (with one exception to be discussed below) are

translationally and semantically indeterminate—but for a different reason than Quine thought. For

Quine, the indeterminacy of translation results from our inability to guarantee that we could prevent

incompatible translations. For TTS, such term and sentence parings or conceptual analyses and

elucidations—call them translation schemas—are translationally indeterminate because they cannot

guarantee their own utility in translation. Since translation preserves meaning, such translation

schemas can at best yield indeterminate semantic results.

As an example, take a normative term like “good.” Philosophers since Moore have attempted

to provide us with analyses of the concept behind this term, and the options have abounded. Moore

said that “good” stands for a non-natural property (Moore 1903). Naturalists disagree and have

provided various analyses (cf. Boyd 1988). Emotivists told us that “good” is a device by which

speakers expressed their emotive yay! (Ayer 1946 [1936], 112). All of these analyses are fine as far as

they go, but none can guarantee how the term “good” will be best treated in the process of translating

an actual text, of a certain type. And here the possibilities exceed the analyses provided by

philosophers. In a poem, “good” may serve to evoke certain imagery by virtue of its association in the

western context with Plato, and a translator of a poem may choose to translate it with a construction

that has no obvious association with Plato or axiology because it serves a similar metaphorical

235
function in the target language—a decision underwritten by the text-type of poetry. Or, alternatively,

it may occur as a modifier in a construction that serves to identify a natural object, and a translator of

science may produce a corresponding translation of the unit under question that does not make use of

any obviously normative concept at all in its place, but which picks out the object referred to by some

other means. In a translation of a psychology text that involves an interview with subjects, terms such

as “good” may be naturally translated in terms that more clearly express preferences of subjects

without any claim to objectivity. Once again, such a decision can arguably be underwritten by the

text-type of science, which does not take the articulation of normative concepts or theories as an

essential part of its mission. What goes for analyses and translations of terms and concepts goes for

translation schemas that pair sentences together within and across languages. Thus, analyses of ethical

judgments that have been proffered in the literature, from the perspective of TTS, would be

semantically and translationally indeterminate, if they are articulated without concern for text-type.

Defenders of the traditional linguistic approach to meaning would likely claim that the whole

point of providing an analysis of the meaning of a concept that stands behind a term like “good” is to

provide us with an account of the core of its meaning that will allow us in various circumstances to

assess what its take home value is. Such implications would be pragmatic in nature, while the analysis

(whether it be that of the Emotivist, or the Naturalist) is semantic for it provides for a type of anchor

for the varied uses of the term within a linguistic practice.

TTS does not have to deny that there is value to such an exercise, but from its perspective,

there can be no privileged, semantic account of our practice of term usage for no such account would

uniquely explain how a term is to be treated in translation, our ultimate test of meaning. From the

stand point of determinate translation guided by text-types, a term like “good” has multiple “anchor”

uses and the task of the translator is to decide which, if any, is doing the semantic work in the passage

being translated, relative to a text-type.106 Indeed, there may be some sense to the idea that in a

culture, a term like “good” is being causally regulated by some natural property, or that it is a device

236
that allows an author to emote, or possibly, all of the above, but from the stand point of translation,

such analyses are far from determinate.

For TTS, there is one important exception to the idea that conceptual analyses, term or

sentence parings are translationally and semantically indeterminate, and that is the case of the text-

type feature. Text-type features are the defining features of a text-type, which structure texts of a type,

and provide a standpoint to organize the meaning of a text and preserve its text-type relative meaning

in a translation. Because text-type features are intimately connected with text-types, their canonical

specification can guarantee how they are preserved in translation. Thus, they are determinately

translatable and meaningful against the backdrop of TTS.

In the case of the novel, character and plot are arguably two such text-type features. Analyses

of these concepts as text-type features do provide us with a clear picture of how they are to be

preserved in translation for they identify the features of a ST to be preserved in translation relative to

a text-type. They would thus be semantically determinate, according to TTS.

In the case of philosophy, which is the text-type charged with the investigation, argument,

debate, criticism, development and application of theories of a general and universal nature on

various topics in light of considerations that are thought to be objectively persuasive, the concepts

that we use to articulate and investigate such theories (GOOD, RIGHT, REAL, KNOWLEDGE, and

MEANING, to name a few) can be given a canonical specification relative to the text-type of

philosophy. Their canonical specification would provide the foundation for how to translate

philosophy. Unlike analyses of terms such as “good,” or “knowledge” that appeal to our practice,

such analyses of the conceptual content of these terms as text-type features would be semantically and

translationally determinate for they would specify the conditions of their utility in translation, namely

in texts of philosophy.

The main difference between TTS and the traditional linguistic approach to meaning can be

understood with respect to their views on compositionality. The traditional linguistic approach to

237
meaning has largely attempted to conceive of meaning as a bottom-up dynamic without any

consideration for text-types. The bottom layer is thought to be determinately meaningful, and by

virtue of this the larger constructions, such as texts, become determinately meaningful. TTS partially

reverses this. While recognizing that language does exercise some bottom-up pressure, TTS regards

the determinate factor to be the text-type, which is constituted by its ability to facilitate translation.

This exerts a top-down influence in the compositionality of a text, most strongly felt in terms of

institutional standards of how to compose and read texts of certain types. And the result is that

language use, and thus its meaning, changes as new ideas are introduced into a culture via translation

or by innovations of the institutions that govern text-types, such as the various institutions of the

modern academy, or the arts.

TTS may seem like a very strong pill to swallow. It involves rejecting the linguistic paradigm

that underwrites so much recent philosophy. The strongest implication of TTS that seems most

difficult to square with intuitions based upon the linguistic paradigm just is the notion that adequate

translation at the textual level saves meaning, even though, when considered in the abstract, the

components of texts are not any of them semantically equivalent across languages. Why this may be

disconcerting for some philosophers is that it implies that linguistic analysis of concepts expressed by

language or language use has nothing objective to track for it is only in a text of a definite type that

expressions have a definite meaning. Conceptual analysis can be reconceived as an effort to track the

defining features of text-types that underwrite a particular institution, such as philosophy, but the

notion that there is some way to tackle a philosophical problem by looking to concepts that

underwrite language use is undermotivated on TTS’s account.

But this I think should not be viewed as particularly surprising or problematic. For some time

now in philosophy there has been a growing chorus of scepticism concerning traditional modes of

philosophizing that appeal to intuitions or language use or both. Recent studies on the cultural

contingencies of linguistic intuitions give us pause to wonder if there is anything objective about the

238
linguistic orientation of philosophy (cf. Weinberg, Nichols, and Stich 2001). But the scepticism goes

back further to Quine, who pointed out that linguistic meaning is in rough shape. It cannot underwrite

translation, but translation is our only and chief means of demonstrating how meaning transcends

provincial contexts such as linguistic communities and languages.

TTS solves this problem in the right way. Instead of adopting negative, destructive post-

modern theses about the radical non-objectivity of meaning, it shows us how we can understand

meaning as transcending the provincial contexts of languages and speech communities through texts

of a definite type with institutional support. And this is indispensable for life as we know it, but it

does involve deprivileging linguistic meaning.

If a translation has been produced according to every relevant institutional expectation, as per

text-type, then it would be strange to hold that all the same the original and translation mean

something different. If STs and their dutifully produced TTs are semantically distinct, then we could

never accurately speak about the views of any author in translation for no translation, no matter how

good, would be the semantic artefact that the author produced. The modern academy would have to

grind to a halt, as would the notion of international law. We could only speak accurately about the

views of persons who wrote in our own language.

We might wish for the notion of literal meaning to help us out of this problem. Perhaps we

can attribute translations to authors if the TT and ST have the same literal meaning. The problem with

literal meaning, on most accounts, is that it is the most basic or systematic semantic modality of an

expression in a language. Thus, the notion that words across languages will have the same literal

meaning is difficult to make sense of. Literal translation is the bane of translators. It is an incurably

linguistically relative notion. But this is a problem not just for philosophy and poetry, but also for

science and law. One way to think about the necessity of text-types is that they are mandated as a

result of the failure of literal meaning to determine translation.

239
 Here is our choice: we either insist that linguistic meaning is fundamental and textual

meaning is absolutely derivative, or we hold that textual meaning is fundamental and linguistic

meaning as secondary. The former route characterizes the dominant approach to meaning in the

recent philosophical tradition. According to this approach, we attempt to save objectivity by looking

to meaning facts about our language as foundational, and view the problem of translation and textual

meaning as derivative and perhaps expendable. According to my proposal, we view textual meaning,

defined by its institutional translatability according to a text-type as foundational, and allow linguistic

meaning to be in a state of indeterminate flux, when conceived in abstraction from texts. On my

account, translations can always be possible because the indeterminate nature of linguistic meaning

allows it to be the malleable resource we require to construct TTs that mirror STs in the text-type

relevant ways. Translations produced according to every institutionally relevant consideration will be

semantically identical to their original. Professors in the modern academy can accurately teach their

students about the thought of authors who wrote originally in another language. Legal documents can

be translated at the international level, as can scientific documents, and accuracy is wholly possible

despite linguistic differences. Or, we opt for the tradition. On this account, we can have putative

objective knowledge about meaning facts about our language, but we could never determinately

translate our texts into other languages because each language is defined by its own meaning facts.

We would never be able to accurately discuss the views of those who speak and write in another

language, and we would never be able to determinately translate our semantic artefacts for people in

other cultures to read and understand, for our meanings would never be able to find exact matches

with meanings in other languages.

Does the tradition save objectivity? I for one have trouble even calling the linguistic account

of meaning objective. If meaning cannot transcend the narrow confines of a language or speech

community, it is difficult to understand how there can be anything objective to it. All “knowledge” is

incurably local, culturally and linguistically relative on this account.

240
 But we are not stuck with the linguistic account of meaning. We have TTS, which is based

upon our best translational practices. Humanity is not a collection of unconnected villages. It is a

species whose history is intimately bound with translation. TTS shows us how meaning can and has

transcended culture and language. But in order to recognize how this is possible, we need to reject the

linguistic paradigm and the notion that particular languages and their histories present us with

determinate, fixed meanings. The very possibility of conceptual change and intercultural

communication shows that the semantic potentials of our words and sentences are not fixed. And it is

this dynamism that makes objective knowledge that transcends contexts possible.

V.5. Classical Expressivism and the Task of Translation

If TTS is correct, many of the views in moral semantics are of indeterminate significance to

the translator. The reason being that such analyses, whether New Non-Analytic Moral Realism or

Emotivism, are not articulated as text-type features that facilitate the translation of texts, but are rather

accounts of the meaning of concepts or sentences with no reference to text-types.

There is also a further problem for our traditional moral semantics proposals. If these

analyses cannot explain to us how we can translate normative and moral discourse, it is highly

doubtful that they can explain normative and moral disagreement across cultures and linguistic

boundaries. For in order to be able to explain normative disagreement, such accounts must be able to

authoritatively identify normative discourse. In order to accomplish this task, they have to be able to

identify normative discourse in languages other than the ones they are proffered in. The only way to

accomplish this task with any authority is by determinately translating normative discourse. But such

a task is beyond the capability of our traditional accounts of moral semantics for they are proffered

without any concern for text-type, which is an essential ingredient in translation proper.

Where does this leave Expressivism? Certainly, the proponents of Expressivism have

unanimously operated within the prevailing linguistic approach to meaning, attempting to explain

moral and normative semantics by reference either to our actual language (Blackburn 2000 [1998],

241
80, 229, 297), or possible languages (Gibbard 2003b, 25). The notion of expression is fundamentally

a linguistic notion that relates a sentence to a mental state that it expresses. Pure Expressivism holds

that the content of normative sentence is a mental state that does not have any sui generis

representational content. The traditional non-cognitive variety has insisted that this mental state is not

cognitive, despite linguistic appearances to the contrary. The recent cognitivist variety of pure

Expressivism corrects this view and insists that the mental states that are expressed by normative

propositions are cognitive, even though they do not have any sui generis representational content.

Hybrid versions of Expressivism have typically sought to combine the non-cognitivism of

Expressivism with descriptive propositions as jointly the content of normative judgments. None of

these pass translational muster for, as they have been presented, all are translational schemas that do

not give any indication of their conditions of employment in translation. As these accounts, as

presented, cannot determinately function in translation, they are incapable of explaining moral and

normative disagreement across linguistic boundaries.

V.6. Expressivism Revised

Before we deliver a final verdict on Expressivism, it is worth considering whether

Expressivism can reinvent itself, not as an account of normative judgments, considered in isolation,

but as a theory of normative meaning, that does guarantee the conditions of its employment in

translation against the backdrop of TTS. In other words, could Expressivism be a feature of a text-

type that we can use to translate normative discourse?

Before we answer this question, it is worth reflecting upon what this request entails.

Against the traditional linguistic paradigm of meaning, translation is a type of paring up of

sentences and words in one language with sentences and words in another language. This is naïve for

a variety of reasons the most important of which is that linguistic resources in distinct languages are

all distinct. Thus, we need some way of knowing when they are equivalent relative to translational

purposes. This is not a problem for we have text-types to tell us when translations are equivalent.

242
They do this by providing us with semantic principles that reduce a ST to the core elements to be

preserved in translation and by providing us with a means to determine what target language

resources will serve this narrow aim. Translation thus typically proceeds against the backdrop of a

developed knowledge of cross-linguistic concordances and grammars, tried and true on linguistic

grounds, as affirmed by experts (sophisticated enough to set out the multiple semantic modalities of

any particular linguistic device). None of this can determine translation on its own but it provides the

bulk of the resources that translators, skilled in specific types, can use to create accurate TTs. When

native resources run short, text-types can help us inscribe new usages in target language texts and

provide us a backdrop to explain the innovation to the target community via framing comments or by

other institutional means that have the wherewithal to deal with controversies and revise translational

practices if need arises.

Our task in this section is to assess whether Expressivism can fulfil the role of a text-type

feature that translates normative discourse. For Expressivism to be successful in this role, it must

distil normative discourse in a ST to its essence (separating off semantic noise that is not to be

preserved in translation), and then help us judge what target language resources will serve to

reconstruct the distilled normative discourse in a TT.

Expressivism as traditionally presented is not an account of the features of a text that must be

preserved in translation, but rather an account of the deep or hidden underlying structure of moral

utterances, something specified without concern for text-type. But if Pure Expressivism is asked to

make this transition from being an account of the deep structure of linguistic utterances to the

semantic principle that allows us to identify and recover normative discourse for translation, then

what could it have to offer other than its interpretation of normative semantic content as essentially

reducible via minimalism? Recall that for Pure Expressivism, in its non-cognitive and cognitive

forms, the various large sounding phrases that make appeal to various weighty metaphysical

principles and objects in normative discourse are semantically nothing more than emphatic

243
metalinguistic devices of assertion. This is arguably the core of the doctrine today. By this deft move,

Expressivists show us how normative discourse can seem like it has sui generis representational

import, though it has none. Fortunately, for Expressivism, this doctrine can be detached from the

traditional linguistic paradigm. It can be reinstated as an account of what a text-type concerned with

preserving normative discourse in translation must consult both in its identification of the content of

normative discourse in a ST and in its reconstruction of the discourse in a TT. But unfortunately,

Expressivism fairs very poorly in this new role.

Consider two philosophers—the Historical Plato as we know him in the tradition and

Deflationary Plato.112 Historical Plato articulates in his dialogue, the Republic, through the mouth of

Socrates, that the Form of the forms is The Good, and he further states that this Good is an ideal

object that exists as a permanent part of reality. Deflationary Plato, through his dialogue, the

Republic-lite (an exact mirror of the Republic, except for its Deflationary take on all normative

issues), and through the mouth of Socrates, talks about the Form of the forms, and says that it is The

Good, that it is an ideal object that exists as a permanent part of reality. However, he also says that by

this he means only to emphasize that first-order ideal objects are admirable or desirable.113 Our

revised version of Expressivism must treat Historical Plato as though his view on goodness is

consonant with Deflationary Plato‘s and the reason for this is that Historical Plato’s account of the

ideal nature of the Good as a second order ideal object cannot be interpreted as having any sui generis

representational content. It serves only as a redundant, metalinguistic device to emphasise what

Expressivism takes to be contained in the claim “the Form of the forms is the Good”. This is bad

enough. But it gets worse. For in its role as a text-type feature guiding the construction of translations

of Historical Plato and Deflationary Plato, our revised form of Expressivism will not be able to make

any distinction between the translations of the views because it interprets them as semantically

equivalent. We could start out with Historical Plato, and end up with Deflationary Plato. We could

start out with Deflationary Plato, and end up with Historical Plato. Our revised version of

244
Expressivism as the semantic device by which the translator assesses the normative content of a ST

and reconstructs this content in a TT leads to a type of translational indeterminacy. It produces

alternate translations of normative discourse that we each have no reason to be confident about.

In the hope of avoiding this absurd conclusion, the Expressivist might try to find some way to

ensure that the peculiar phrases made use of by Deflationary Plato and Historical Plato that lead us to

distinguish their views are preserved in translation. On the traditional approach to meaning, this is

putatively accomplished by ensuring that every word and sentence in a ST is paired up with an

equivalent word and sentence in the TT. We know that this does not reflect the reality of translation in

light of the peculiarity of languages. Thus, the Expressivist must find some work-around to ensure

that the phrasing that distinguishes Historical and Deflationary Plato is some how reconstituted in the

TT without appeal to this naïve conception of translation.

The most obvious way that we can make sure that we do not mix up Historical Plato and

Deflationary Plato in translation is to give up on Expressivism. We would simply defer to the

normative theories of these respective philosophers to help us identify what they each take to be the

import of their text and use this to guide our translation (what translators have traditionally done).

What were conceived of us mere phrases, verbal disagreements or aesthetic embellishments—in

essence, semantic noise under Expressivism (cf. Blackburn 2000 [1998], 79; Gibbard 2003b, 12-

13)—would be interpreted as essential to the articulation of the respective normative theories and thus

essential to any acceptable translation. Target language resources would be assessed as adequate

relative to their ability to articulate Historical Plato’s and Deflationary Plato’s respective views, not

Expressivism’s view about what is relevant. Absent this natural route, it is unclear how we could

design a text-type principle that would be adept enough to mimic the phrasing of a source text author

in a translation without thereby deferring to the ST author’s philosophy to produce this similitude.

Perhaps what might help the Expressivist is to argue the following. The problem so far is that

we have assumed that the only way to assess the content of a ST is through Expressivism, when its

245
job is merely to preserve a narrow portion of the text, namely the non-cognitive normative content.

Rather, we must look to other semantic principles, such as the one charged with preserving the

metaphysical content of a ST, to preserve the realist phrasing of Historical Plato, absent in

Deflationary Plato. With these in tandem, we can translate both these thinkers in a manner that does

justice to their texts without completely giving up on Expressivism as our account of normative

semantics.

With this suggestion, we move to an adaptation of Hybrid Expressivist views. The traditional

accounts of hybrid Expressivism provided in the literature concern individual sentences (they analyze

them into two separate judgments, one non-cognitive and the other descriptive) and are thus, as

presented, of indeterminate significance in translation. If we allow this account to migrate to our

textual approach to semantics, then what we have is the suggestion that normative discourse can be

distilled into two separate text-type features, one concerned with the non-minimalist, non-cognitive

aspect of the text (emotions, desires) and the other with metaphysical matters. This seems promising

because any viable text-type will have more than one feature that it attempts to keep track of.

Certainly, at least in the case of the philosophical text-type, metaphysical content of STs will be one

such feature that the text-type will track. If we could interpret the metaphysical content of a ST and its

non-cognitive content independently of each other, then perhaps we can translate texts such that

Traditional Plato does not sound like Deflationary Plato. Unfortunately it is unclear how, after

separately reducing the non-cognitive and metaphysical content of a ST into separate cores, that our

translator has anything other than a list of attitudes and desires (such as, “admire formal objects”), on

the one hand, and non-normative metaphysical theses on the other hand (such as, “there is a second

order formal object”). Our problem is that the two distilled camps of attitudes and metaphysical

theses do not jointly (and in some cases, singly) correspond to the actual doctrines of either Plato and

thus there is no obvious way that we can use them to reconstitute TTs that mirror the STs in phrase.

Once again, it is difficult to tell what the work-around for this problem would look like, for absent

246
deferring to the full philosophical projects of Historical Plato and Deflationary Plato we have no way

to mediate translation through a text-type and yet have it mimic the ST in a manner that we find

acceptable.

The Expressivist might wish to exploit the polysemy of language and look to another

Expressivist friendly text-type that will seize upon what Expressivism says is essential in normative

discourse and preserve that in translation. After all, we have been focussing on philosophers and not

all normative discourse is found in philosophical texts. While Expressivism may be of no use in

translating philosophy, perhaps there are some other types of texts that suit its particular outlook.

While TTS does affirm that translation is text-type relative and that there may be many text-types

concerned intimately with the preservation of normative discourse, it is difficult to imagine that there

could be any workable text-type concerned essentially with normative discourse that is so relaxed

about the relationship between explanatory reasons agents (authors or characters in a dialogue)

have for their substantive claims and their substantive claims, as Expressivism is.

This type of dismissive posture towards reasons can be found in the writing of Gibbard and

Blackburn. Gibbard dismisses theoretical disagreement between rival moral theories as merely

“verbal” and not genuine114 and Blackburn calls weighty metaphysical justifications for moral claims

mere “flowers” that semantically add nothing.115 In both cases, what thinkers take to be important to

understanding their normative convictions as they state in their texts are put aside and instead an

Expressivist analysis of the pure, atomic core of normative meaning is put forth.

The defenders of the recent “Cognitive Expressivism” of Terrance Horgan and Mark

Timmons (2006), resists reducing normative concepts to other types of concepts. Perhaps if we were

to try to apply this revised version of Expressivism to the problem of translation we could make some

headway in avoiding the types of problems cited to this point for we would be barred from being

dismissive about authors’ desires to treat ethical concepts as non-reducible to affective concepts. By

now, it should be apparent that any innovations that we might seek to add to Expressivism will result

247
in a semantics that will fail us in the translational process. The reason is simple. Translation does not

proceed sentence by sentence. It proceeds by selectively interpreting the content of a ST down to its

translational essence, relative to a text-type, and then proceeding to construct a TT with different

linguistic resources that are judged equivalent to the narrow content extracted from a text. If we

employ a normative semantics that is at all at odds with the dialectic of reasons and claims that we

find in a normative discourse, then the resulting translation will be inaccurate. For the moral

semantics that we employed will keep us from preserving in translation what falls outside the pale of

its theory, whether these be normative claims or their supporting considerations. At this point, our

defender of Expressivism (whatever version she wishes to defend) might contemplate some type of

work-around so that we can preserve the characteristic phrasing of our ST through translation. But it

is difficult to understand how this can be effective short of deferring to the complex of claims and

reasons (i.e., the philosophy, roughly speaking) of the ST author in guiding our assessment of what

phrases should be preserved. But if we are to defer to the authors views on the matter, then it seems

that we have given up using Expressivism as the semantics by which we judge what is to be preserved

in translation.

If TTS is the correct account of meaning and translation, Expressivism’s troubles in

translation suggest the following: it is not primarily a moral or normative semantics, but rather a

substantive view in moral philosophy about what types of moral truths there are.

QI, in contrast, can perfectly handle the challenge that Expressivism fails. According to QI,

normative and evaluative concepts articulate universal and general theories with a world-to-theory

direction of fit distinguished by a specific axiological differentia, in light of relevant considerations

that authors take to be objectively persuasive. Thus, pressed to translate Historical Plato and

Deflationary Plato, QI will be able to track the relevant differences in the respective texts for it will

defer to the appropriate author and character in Historical Plato’s and Deflationary Plato’s dialogues

to fill in the theory and reasons. QI can also lead us to appreciate the various dramatic events in the

248
dialogue as worthy of retention in translation as they set the stage for arguments and are at many

points woven into the reasons that are provided. They would thus be preserved not because of their

artistic merit but because of their role in theory articulation. QI thus encapsulates the very dialectical

structure of philosophy. Expressivism, in contrast, regards all reason and philosophy as sheer,

redundant noise in relationship to the core of evaluative meaning that it takes to be the mere

expression of the mind’s desires or plans.

V.7. Conclusion
Classical Expressivism cannot translate normative discourse because it illuminates sentences,

not texts. Revised versions of Pure and Hybrid Expressivism that are conceived of as text-type

features of normative discourse cannot determinately translate normative discourse either. No

modification to Expressivism that retains its character as a normative semantics that is critical of

some supporting reasons for normative claim can help us in this endeavour, for any semantics that

takes a stand on the types of reasons that are exclusively relevant to normative concerns will be a

semantics that will fail us in translating normative discourse. This reflects poorly on its prospects for

explaining normative disagreement across cultures and linguistic boundaries. In order to explain

disagreement on normative matters across linguistic boundaries, Expressivism must be able to

accurately identify the content of such discourse across languages. But in order for Expressivism to

accomplish this task, it must be able to accurately translate normative discourse. However, it cannot

accurately translate normative discourse.

What does this say about Expressivism’s efforts to isolate normative meaning apart from

other types of meaning? Expressivism’s misadventures in translation tell us that any narrow attempt

to understand normative discourse as revolving essentially around substantive claims, apart from

supporting considerations, will result in distortions. And this is not simply a problem for the reader.

Translators can only choose appropriate translational equivalents of sentences based upon textual

considerations—such as supporting reasons. Expressivism’s efforts to abstract from such reasons a

249
pure normative sentence, or to understand an entire passage as articulating one underlying judgment

with no sui generis representational content, operates at a level that takes no notice of what goes into

preserving meaning in translation. Not only can people disagree on substantive normative claims, but

they can also disagree on their supporting considerations. And any minimally adequate account of

translating normative discourse must take both into account in translation. This is simply part of our

translational practice of preserving normative discourse. To treat normative and evaluative discourse

as a mere list of normative assertions to be translated is to fall back into the sentential, linguistic

approach to translation which is untenable. Nothing can stop the Expressivist from extracting from

texts some sentence that they take to be representative of the pure normative judgment, but because

this process of abstraction plays no part in our actual translational practices (we know it cannot, for it

only leads us to trouble if we heed it), it is of indeterminate semantic significance.

What then does this tell us about normative disagreement? If from the perspective of

translating normative discourse we must keep an eye on substantive claims and supporting

considerations, and if the complex of such claims and considerations constitutes normative discourse,

then disagreement on supporting considerations constitutes a type of normative disagreement. And

the range of this disagreement can be quite wide: everything from aesthetic choices to the

metaphysical ground of value. In short, the range of disagreement in normative discourse is nothing

short of the range of philosophical disagreement on matters of value.

Thus, it seems, if nothing else, normative discourse is a text-type feature of the philosophical

text-type. In other words, if normative discourse is a part of a philosophical text, it must be preserved

in the translation of such a text as a text of philosophy and failure to preserve this aspect of the

philosophical ST would be a failure of translation. Put yet another way, under no circumstances could

normative discourse be treated as semantic noise in the translation of philosophy. This confirms what

was already apparent in the failure of Expressivism to translate normative discourse, namely that

normative disagreement is a species of philosophical disagreement.

250
 This has an important implication for normative and evaluative semantics. If we want to have

a determinate answer to what normative terms and sentences mean then we should move to thinking

about normative concepts as text-type features of philosophical texts that act as translational filters

that will allow us to capture all the features of normative discourse that we wish to preserve in

translation, including the dialectic between candidate normative theories and putative considerations

that bear upon normative theory selection. The semantics of normative concepts on this translational

account would explain normative discourse across cultures and linguistic boundaries as coordinated

around the objective aspirations of philosophy. It will manage this feat because it does not seek to

illuminate normative semantics by reference to controversial metaphysical theses, but by reference to

what we need to keep track of to translate normative discourse. Expressivism would be simply one

among many moral philosophical positions we could translate in this manner, and not in any sense the

foundation of normative discourse.

QI upstages Expressivism, for it is the only account of normative and evaluative discourse in

the running that treats normative and evaluative concepts as a species of philosophical concepts. As

we shall see in the next chapter, accounts of normative and value concepts in the literature that have

some superficial similarities with QI fail to capture the philosophical aspects of value and normative

concepts. But this is not surprising: philosophy is a type of text or discourse and only QI is an account

of the text-type features of such a discourse.

251
VI. Quasi-Indexical Account of Thick and Thin Concepts

“Now the first thing that strikes one about all these expressions is that each of them is
actually used in two very different senses. I will call them the trivial or relative sense
on the one hand and the ethical or absolute sense on the other. If for instance I say
that this is a good chair this means that the chair serves a certain predetermined
purpose and the word good here has only meaning so far as this purpose has been
previously fixed upon. In fact the word ‘good’ in the relative sense simply means
coming up to a certain predetermined standard. Thus when we say that this man is a
good pianist we mean that he can play pieces of a certain degree of difficulty with a
certain degree of dexterity. And similarly if I say that it is important for me not to
catch a cold I mean that catching a cold produces certain describable disturbances in
my life and if I say that this is the right road I mean that it's the right road relative to a
certain goal. Used in this way these expressions don't present any difficult or deep
problems. But this is not how Ethics uses them.” — Wittgenstein, Lecture on Ethics

In Chapter 2 I showed that a proper account of normative discourse translation must be text-

type-theoretic, for translation is only determinate at the textual level thanks to the role of text-types

and their institutions. In Chapter 3 I argued that recognizing the text-type-theoretic nature of

translation has implications for meaning. Specifically, the only way to recognize the role of text-types

in translation is if we deprivilege the role of linguistic meaning and recognize the determinate role of

text-type-theoretic meaning. It is by virtue of the text-type that the polysemy of all semiotic resources

has a determinate contribution to make to a text of a definite type. If we want to translate normative

and evaluative discourse, and if translating this discourse means preserving normative and evaluative

concepts as essential features of such discourse, we need to locate the text-type for which such

concepts are essential. I have argued that value and normative concepts are essential features of the

text-type of philosophy. In chapter 3 I spent some time comparing the text-type of philosophy to

contrary text-types. In chapter 4 we saw that the linguistic accounts of moral concepts in the literature

are simply not up to the task of explaining how normative discourse translation is possible, for they

assume a semantic similitude of languages to underwrite normative discourse translation. This ignores

the reality that languages develop their semantic profiles within unique historical and cultural

252
contexts. It also ignores the fact that philosophical differences among differing cultures will lead to

very different semantic profiles for their philosophically interesting terms. Only QI as the text-type of

philosophy can help us appreciate how we can overcome such cultural and philosophical diversity by

virtue of being the text-type feature of the philosophical text-type. In chapter 5 I examined whether

Expressivism could play the role of the text-type feature responsible for translating normative and

evaluative discourse. The conclusion that was reached was negative. Expressivism cannot help us

translate normative and evaluative discourse because it conceptualizes moral meaning in a manner

that abstracts from wider philosophical commitments that an author or agent might have. In so

conceiving of value as an atomic state of mind, abstracted from wider philosophical commitments of

persons, it fails to preserve essential content that any institution concerned with normative and

evaluative discourse translation would want preserved. QI in contrast can rise to this challenge

because it treats normative and evaluative meaning as a species of the text-type of philosophy.

With the two main contenders in the moral semantics literature shown to be failures at

normative discourse translation, QI is the only plausible account in the running: it succeeds exactly

where linguistically-oriented accounts of moral semantics (such as Non-Analytic Naturalism) and

anti-philosophical accounts of moral semantics (such as Expressivism) fail. In this chapter I shall fill

out some details of QI and compare it to similar accounts of philosophical concepts in the literature.

QI will distinguish itself as an explicitly textual account of philosophical meaning. Its basis in TTS

sets it apart from competitors.

VI.1. Review of TTS and the Semantic Foundation of QI

QI is what I have called a “text-type feature” of the philosophical text-type. It is the feature of

a philosophical text that must be preserved in translation in order to secure the right type of

equivalence between the philosophical translations of a ST into a TT. The semantic relevance of text-

types, in turn, was argued for in chapters 2 and 3. The idea of “text-types” arises out of the translation

studies literature (cf. Holmes 1988 [1972], 74-76; Laviosa-Braithwaite 2001, 277-278; Neubert and

253
Shreve 1992). A text-type is like a genre, except it has widespread institutional recognition and

pertains to structural features of texts that exceed what is usually understood under the heading of a

“genre”. Examples of text-types include the philosophical text, the novel, the scholarly text on natural

science, and text-types can include among them sub-text-types. In chapter 2 it was argued that

determinacy in translation is only possible on the textual level, and moreover that the determinate

translation of a text is only possible when a text has been subsumed under a text-type, conserved and

improved upon by a text-type institution. (In other words, there is no determinate way to translate

words or sentences considered outside of texts or without reference to their function relative to a text-

type.) In chapter 3 I responded to the claim that while the theory of translation presented in chapter 2

(what I call the Liberated Equivalence conception of Translation or LET) accounts for translation, it

tells us nothing obviously about semantics. In response I argued that LET presents an implicit

criticism of traditional semantic theories that problematize translation and thus it cannot be

understood as semantically neutral. Rather, a full defence of LET contains within it a theory of

semantics that I call TTS. According to TTS, if meaning is preserved in translation, and if translation

is only determinate on the textual level when text-types are brought to bear, then there are only three

categories of determinately meaningful semantic phenomena: (i) texts that stand as STs and TTs to

each other, (ii) text-type features, and (iii) cotextually defined translation units within texts, be they

words, sentences or passages. All other semantic phenomenon, such as words or sentences considered

in abstraction from texts of definite types, have a very limited type of meaning, which I dubbed

“indeterminate meaning”. Though we can attempt to provide translations for words and sentences that

would lay bare their meaning, there is no determinate way to translate any particular word or sentence

apart from a text or text-type. And as meaning is preserved in translation, the meaning that one arrives

at through such a process of indeterminate translation is at best indeterminate.

254
 The idea that two texts can be semantically equivalent to each other as two translation units

can be equivalent is relatively easy to comprehend, for the equivalent items have a tangible quality to

them. But how are text-type features equivalent?

Text-type features typically have two aspects: an inscriptional aspect, and a semantic aspect.

The inscriptional aspect changes according to contextual factors, such as the language that a text is

written in, or according to cotextual considerations of how such a text-type is expressed in texts. This

is not literally preserved in translation. What is more straightforwardly preserved in translation is the

semantic aspect of a text-type feature. The semantic aspect not only helps explain and define the

genre of a text-type, but it also performs an essential regulatory function in the semantics of texts.

Texts, as noted in chapter 2, are polysemous, with many competing features, not all of which can be

preserved in a translation. By subsuming a text under a text-type, we split a text into two categories:

text-type features, and non-text-type features. The former are essential to all texts of the type, while

the latter are contingent, and consist in, among other things, bits of language that have many

conventional, indeterminate translations and interpretations. The semantic aspect of the text-type

feature plays the role of constraining and regulating the contingent, nontext-type features of a text so

that they subserve the text-type. Understanding this regulative function allows a translator to properly

translate a text, and allows the savvy reader to grasp the take-away content of a text. Indeed, it is the

reader’s grasp of the regulative function of the text-type feature that makes her savvy, in comparison

with the novice who is coming upon an instance of a text of a certain type for the first time.

The Quasi-Indexical account of thick and thin concepts is an account of the semantic aspects

of the text-type features of philosophy. Inscriptionally, the text-type features of philosophy are

instanced by words, typically nontechnical vocabulary, which I call key philosophical terms. Salient

examples of such words are “good,” “right,” “ethical,” “real,” “knowledge,” and “beauty.” The

semantic aspects of these text-type features, in contrast, are semantic rules that connect the key

philosophical terms with various variables to be filled in by contextual and cotextual factors.

255
Foremost among such contextual variables is a philosophical theory that an author consults in order to

select the referent of her use of a key philosophical term. Thus, for instance, when a philosopher uses

the word “real”, she defers to a metaphysical theory (identified in part by the semantic rule governing

“real” in philosophical texts) to inform her choice of the referent of “real.” An Idealist may thus have

a different referent in mind and in text than a Materialist, though both defer to the same semantic rule

in choosing the referent of “real”. That such contextual differences between the Idealists and

Materialists are consistent with a common meaning of “real” is essential to understanding and

translating philosophical texts across contexts, such as languages and cultures.

Because the semantic rule governing key philosophical terms explicitly defers to contextual

input in order to determine reference, I have called this account of the semantics of key philosophical

terms the Quasi-Indexical account, for indexicals are the paradigms of words whose meaning defers

to contextual input to determine reference. As such key philosophical terms come in two varieties—

thin and thick (most notably in the case of value concepts)—QI must be able to account for the

relevant difference between thick and thin terms such as cruel” and “evil.” While the history of the

term “concept” is closely associated with a linguistic and psychologistic account of meaning, I shall

call the semantic aspect of a text-type feature the concept, and the inscriptional aspect the term. Two

key philosophical terms can, therefore, instantiate the same concept when they have the same

semantic aspect. While the psychological reality of philosophical concepts is a murky issue, they

function like traditional concepts, contributing to the truth value of the propositions they form.

VI.2. Indexicality and Philosophy in the Literature

VI.2.1 Kaplan’s Theory

QI is an account of the semantics of key philosophical terms (such as “right,” “wrong,”

“moral,” “ethical,” “evil,” “real,” and “knowledge”) that is modelled loosely on the now classic

account of the semantics of indexicals, provided by David Kaplan.

256
 According to David Kaplan, indexicals take on a referent according to a speaker, time and

place, and what he calls the character of the indexical. On the notion of character, Kaplan writes:

[This] kind of meaning, most prominent in the case of indexicals, is that which determines the
content in varying contexts. The rule, “‘I’ refers to the speaker or writer” is a meaning rule of the
second kind. The phrase ‘the speaker or writer’ is not supposed to be a complete description, nor is it
supposed to refer to the speaker or writer of the word ‘I’. (There are many such.) It refers to the
speaker or writer of the relevant occurrence of the word ‘I’, that is, the agent of the context… Let us
call… [this] kind of meaning, character. The character of an expression is set by linguistic
conventions and, in turn, determines the content of the expression in every context. Because character
is what is set by linguistic conventions, it is natural to think of it as meaning in the sense of what is
known by the competent language user. (Kaplan 1989, 505)

Kaplan also provides a formalized account of character, according to which it is a function of

contexts to contents (Kaplan 1989, 506).

Kaplan’s account of indexicals is not the only account, but it is perhaps the most influential to

date.116 The useful aspect of Kaplan’s account is the idea of a semantic rule that governs the usage of

certain words, such that the rule along with contextual factors yields reference. This is the aspect of

Kaplan’s thinking on indexicals that has been influential in metaethics, for it appears as a means of

accommodating an obvious relativism about the content of people’s moral judgments. If people differ

with respect to the evaluative theories they employ, it will follow that their evaluative judgments will

differ in content. As the value theories that people employ vary according to context (whether the

contexts be cultural, or axiological, as for instance the difference between an ethical and aesthetic

context), the contents of the judgments expressed with value words will differ accordingly.

VI.2.2 Indexicality and Ethics

James Dreier was perhaps the first to recognize this application of Kaplan’s thought to the

problem of ethics, and moreover he recognized the utility of this approach in translating ethical

discourse. He writes:

257
 If a term in some alien language has the character of “good” we can sensibly translate it as “good.”
The alien community may not have the same motivations as we have, so the term we translate as
“good” may not have the content of “good” as we use it. The character, what stays the same between
one context and another, determines the content in a context, and then the content determines the
truth conditions of the sentences in which the term occurs. If something like this is right, then moral
terms have a two-level semantics of the kind David Kaplan proposes for indexicals. Indexicals have
as their primary meaning, according to Kaplan, not a content, but a character, a rule for determining
the content given a context. (Dreier 1990, 8)

Dreier is on the right track here, I believe. But there are many problems with his formulation.

Dreier called his application of Kaplan’s account of indexicals to the problem of moral

semantics “Speaker Relativism.” According to Dreier, ethical propositions are a function of

motivational systems to propositions. The function in question is the character that governs moral

term usage. “Good”, for instance, will roughly have the character of “highly evaluated by standards

of system M” and the sentence “x is good” will have the propositional content of “‘x is highly

evaluated by standards of system M,’ where M is filled in by looking at the affective or motivational

states of the speaker and constructing from them a practical system” (Dreier 1990, 9). According to

this view, the resulting propositions expressed by evaluative sentences have straight forward truth

conditions. Two speakers with differing motivational input can differ about whether x is good, and

both speak truly, so long as they accurately describe their motivational states (Dreier 1990, 7).

Speaker relativism according to Dreier comes in two versions. According to the “crude,”

individualistic version, the motivational system that is the input into a character that yields content is

the motivational system of the individual. Dreier also recognizes another variant of Speaker

Relativism, where the standards that the context provides are not determined by the individual alone,

but by communal or societal standards. This latter approach can explain to us how a Marquis de Sade

can be consistently wrong on moral issues. In both cases, Dreier believes that internalism, the view

that “to accept (sincerely assert, believe, etc.) a moral judgment logically requires having a motivating

reason” (Dreier 1990, 6), is supported by Speaker Relativism, for in both cases, it is not a value

258
theory, or even an ethical theory, but a “practical system” comprised of motivations of an individual

or an individual as a function of a society that count as the input into the evaluative character that

yields a referent for moral vocabulary (Dreier 1990, 21-22). Dreier recognizes that his societal

version of Speaker Relativism opens up the possibility that, in some cases, deviant individuals may

lack the right motivating connection between moral conviction and action due to aberrant psychology.

I’m not entirely sure if this is even consistent with his commitment to Internalism. However, Dreier

believes it is, for the version of Internalism that he wishes to champion appears quite weak. In the

face of such deviant cases, he writes that “internalism can be saved by qualifying it so that it posits a

necessary connection between the believed good and motivation only in the normal cases” (Dreier

1990, 10-11).

One problem with Dreier’s account is that he takes onboard Kaplan’s linguistic construal of

the semantic rule governing usage that yields reference—what Kaplan calls the “character”—in order

to explain the meaning of value terminology. From the perspective of TTS, if moral terminology is to

be preserved in translation and to thus have a determinate meaning, it must be a text-type feature of a

specific type of text (please see chapters 2 and 3 for a more detailed argument in defence of this

thesis). Thus, whatever semantic rule there is that specifies its textual significance must be textual

(not linguistic) in nature. I have no doubt that the character that Kaplan identifies for the English

word “I” is linguistic in nature. In other words, from what I can see, “I” is not a text-type feature of

any specific text, mandating its preservation as a text-type feature through translation, but the

character of “I” is a rule that informs speakers of English in their linguistic interactions. Whether the

first person, personal pronoun survives translation will depend upon other contingencies of the text it

occurs in, along with the text-type that is applied to it.117

While I agree that there is a structural analogy to the semantic rule governing “I” and value

terminology, there is an important difference: value terminology is the inscriptional aspect of text-

type features of philosophical texts, whereas “I” is not the text-type feature of any salient text-type.

259
We can make sense of the idea of “value discourse” and its essentially philosophical nature. We

cannot make sense of the idea of “I” or indexical discourse—given the contingencies of our text-type

institutions. And if we could, the character of “I” would not be a linguistic rule, but a textual rule.118

Dreier cannot truly be faulted for failing to appreciate the textual aspect of the semantics of

moral terminology, for he is not proffering his argument against the backdrop of TTS. However,

Dreier does not seem to be proffering his account of speaker relativism on the basis of a general

theory of semantics at all.119 Rather, he appears to be providing an account of the semantics of moral

concepts in isolation from general questions of meaning. In other words, Dreier’s argument suffers

from what Cappelen and Lepore call “semantic opportunism.” I believe that this is a grave problem.

For it means that such an approach is not mandated by any general and plausible theory of semantics,

and thus it does not have a semantic justification, in the broadest sense. Worse still, as an

opportunistic account, it will possibly contradict or conflict with a plausible theory of semantics.

Whatever the merits of QI, it cannot be charged with being semantically opportunistic, for the

argument I am presenting for QI is based upon TTS. QI, as I conceive it, is the implication of TTS

with respect to the question of philosophical semantics.

If Dreier’s proposal is not to be semantically opportunistic, it must be derivable from a global

theory of semantics. The only such theory that I am willing to defend is TTS, for TTS, unlike

competing accounts, is able to explain how determinacy in translation is possible. If our goal is to

explain how value discourse can be translated, we would do well to take TTS as our semantic

grounding. Unfortunately, Dreier’s Speaker Relativism does not comport well with the requirements

of an account of the text-type features of philosophy.

From the perspective of TTS, Dreier’s proposal is relativistic in the wrong way. Philosophical

concepts preserved in translation must be amenable to a certain type of relativism. The relativism in

question must make room for individual authors to select a philosophical theory, be it ethical,

aesthetic, epistemic, or metaphysical, and for this theory to guide an author in the authorship of her

260
text, and to inform and to guide a translator in her effort to save the philosophical content of the text

in translation. This room for theoretical choice is part and parcel of the philosophical text’s structure

as the text-type dedicated to the application, articulation, investigation, argument, debate and

criticism of philosophical theories, which are theories of a general and universal nature articulated

with ordinary, non-technical words. Given the very nature of the philosophical text-type, the

soundness of philosophical theories and extrapolations from such theories is a subject for debate.

However, on Dreier’s account, ethical claims are self-fulfillingly true, so long as they are properly

formed by inputting the relevant set of motivations into the character of moral terms, whether such

theories be determined relative to the individual or the society that the individual lives in. There is

thus no real room for philosophical debate. If our account of philosophical concepts is to be the

regulative factor in determining the semantic content of a text, Dreier’s Speaker Relativism would

make profound disagreement and argument irrelevant to the text-type of philosophy, for these are not

supported by Speaker Relativism. This is clearly a mistake. Whatever we as philosophers might think

about the validity of much moral philosophical debate, it is undeniable that the practice of philosophy

and the text-type of philosophy are structured around this possibility. The latter is what a text-type

account of key philosophical terms must track, and preserve in translation, not what we ought to think

is plausible in the realm of moral argumentation. Its role is to help us preserve in translation what a

philosopher thinks is relevant to their philosophical project as far as this is to be found in their texts,

not what is objectively relevant according to any other account. And philosophers typically believe

that considerations beyond those recognized by Speaker Relativism are relevant to their projects.

Another problem with Dreier’s account is that it does not generalize to all philosophical

fields. It is at best an account of value semantics, and at the most narrow an account of moral

semantics. Instead of proffering a theory as the input value that, along with the character of a value

term, yields a propositional content, Dreier chooses to identify a psychological “practical system” of

motivations as the input. One problem for the generalizability of Dreier’s account is that it is not at all

261
obvious how motivational or practical systems can function as inputs in the semantics of the concepts

of metaphysics and epistemology.

Dreier would likely respond to this charge that his account was never meant to generalize,

and that if it were to be co-opted into an account of the text-type features of philosophy, it could at

best serve as the model for evaluative philosophical concepts. But even as an account of the semantics

of value concepts, considered as text-type features of the philosophical text-type, there are problems

with his motivational approach. Motivations might motivate philosophers, but clearly the text-type of

philosophy has theories as the input informing usage and reference of key philosophical terms such as

“ethical,” “moral,” “justice,” and “virtue.” Unlike motivational, practical systems, whose sole end is

to motivate action and behaviour, theories have their primary purpose in explanations proper to their

province. Ethical theories explain ethical matters (i.e., they tell us when and why something is ethical

or evil, virtuous or vicious, obligatory or forbidden) just as physics theories explain physical matters

and mathematical theories explain mathematical matters. Commitment to a theory can constitute a

type of psychological practical system, with motivational consequences. But we confuse the issue

when we conflate a theory with an individual’s commitment to the theory. Failing to make this

distinction will, for instance, make it difficult to understand how philosophy can offer an explanation,

be it moral or epistemic.

Dreier might object to this criticism, and argue that motivational systems can indeed provide

explanations: they can explain what an individual is motivated to do, and to the extent that

accomplishments are highly evaluated by a motivational system, they explain why something is

“good.” But, in response, we must protest that this falls far short of the type of theoretical variety we

find in the philosophical literature concerning ethical issues and explanation. Plato, for instance,

never held that x is Good because it is valued highly by his motivational system, but rather that x is

Good because it instantiates the ultimate metaphysical reality, The Good, and in virtue of its

superlative status it motivates those who understand it. However, if we were to treat Dreier’s proposal

262
as indicative of the semantic aspect of key philosophical terms—the conceptual aspect of “good” that

regulates philosophical texts and guides the translator in the reproduction of a semantically equivalent

text—we would have to translate Plato, when he uses “good”, as though what he meant was that the

item he evaluated as good is good in virtue of being highly evaluated by his motivational system. This

can’t be right for it gets Plato’s view backwards. It is even more implausible in the case of a

philosopher who rejects metaethical internalism for in this case we would have to translate her as

though she were committed to the doctrine. Philosophical translation must preserve the full range of

philosophical explanations of moral phenomena, and not simply the narrow, psychologistic account

that Dreier has in mind. This is because we need to preserve the substantive views found in STs in

translation via an account of philosophical concepts.

An improvement on Dreier’s Speaker Relativism is David Phillips’ “Middle Ground Moral

Semantics”. Phillips, like Dreier, proffers a Kaplanesque account of moral semantics. The

improvement in Phillips’ account is his suggestion that the contextual inputs into the character of

moral terms are not motivational systems but standards or systems of norms. This is an improvement

as this comes closer to the demands of an account of the text-type features of philosophy that must

countenance theories as the contextual input into characters that yield reference for key philosophical

terms.

Phillips also shifts from Drier’s presentation by calling the type of context-sensitivity at play

in moral terminology, whereby moral terms take on different referents in different contexts, “Quasi-

Indexicality”—a nomenclature I quite like (Phillips 1998, 142). However, Phillips’ view suffers from

many of the problems of Drier’s. He writes:

Begin with judgments of the form “X is good.” According to the semantic account, the truth
conditions of this utterance are given by: X meets the relevant standards. What the relevant standards are
is determined by context…. [This] gives judgments involving “good” straightforward truth-
conditions. (Phillips 1998, 141)

263
 Phillips takes pains to argue that his account is not relativistic, for his account says nothing

about the propriety of normative systems, whereas relativists, on his account, are typically committed

to the idea that there is no common arbitrator between competing normative systems (Phillips 1998,

150). However, given Phillips’ account of the truth conditions of a proposition “X meets the relevant

standards” expressed by the value sentence “x is good”, it is difficult to see how his account is not

relativistic. Despite putative differences as to the validity of value systems, on Phillips’ account, one

can employ any normative system, even a substandard normative system in moral talk, and speak

truthfully when one says “X is good” so long as one accurately assesses that X meets the standards

that one defers to. While Phillips makes room for the possibility that we can get the content of our

own standards wrong by allowing that the standards are normative systems not to be conflated with

one’s psychological state (Phillips 1998, 146), his theory does make clear-headed moral speech self-

fulfillingly true. “X is good” on this account is true for me simply by virtue of being uttered by me

when I’m clear-headed about my own convictions. This type of relativism does not sit well with the

text-type of philosophy.

Phillips attempts to avoid this conclusion by arguing that there is a distinction between the

activity of invoking normative standards in ordinary moral judgments (which presumes their validity)

and justifying them (Phillips 1998, 143). However, it’s unclear to me how we can attempt to justify

our moral standards without invoking them as well, particularly if the very semantics of moral

terminology demands that we defer to our normative commitments in order to deliver reference.

Perhaps Phillips’ idea is that we can offer epistemic or metaphysical justifications for a particular

moral theory, without actually ever employing moral terminology in the justification, thus avoiding

offering self-fulfilling moral justifications for a moral theory. However, if we generalize this model to

all philosophical terms, as TTS demands, epistemic and metaphysical judgments will turn out to be

self-fulfillingly true, so long as one accurately defers to one’s theoretical commitments. I think the

264
conclusion to be drawn here is that Phillips presents us with no way of avoiding the unacceptably

relativistic implications of the Dreier, Kaplanesque account of moral semantics.

The relativistic underbelly of Phillips’ account is made clearer in an earlier paper, where he

argues that relativism can be secured by what he calls the “indeterminacy thesis” according to which

the “concept of morality is insufficiently constraining to narrow us down to a single adequate system

of moral norms or standards” (Phillips 1998, 398). Phillips’ Middle Ground Moral Semantics, for

better or worse, fits this criterion of an indeterminate concept of morality for it does not narrow down

the range of possible alternative normative standards.

In summary, while a certain amount of relativism is necessary in an adequate account of the

text-type features of philosophical texts, the type of relativism is the room for authors to freely choose

their theoretical commitments. However, as the philosophical text-type is committed to the

articulation and debate of such theories, the relativism of philosophy cannot be obviously alethic or

substantive. The relativism of philosophy must be compatible with the prospects of genuine debate

and disagreement, for the text-type of philosophy is structured around such prospects. (This is

precisely the prospect ruled out by Dreier’s and Phillips’ relativism.) If semantics were a linguistic

phenomenon, it would be possible to dismiss such institutional aims of philosophy as irrelevant to the

semantics of philosophical concepts. But as philosophical semantics is textually grounded in a

specific text-type institution, such institutional considerations are crucial. Specifically, if key

philosophical terms are the text-type features of the philosophical text-type, their semantic aspect

(i.e., their conceptual content) must paint an accurate picture of the goals and aspirations of the

philosophical text. Failing that, it will not be a genuine account of the structure of the philosophical

text-type. Failing that, it will not be able to facilitate the accurate translation of philosophical texts.

Speaker relativism is certainly not the platform of philosophy. If it were, there would be no need to

preserve argument in philosophical texts. Philosophy would merely be a species of autobiography,

265
and it would be translatable as autobiography. Every instructor of philosophy knows that this is not

what we expect from our students in their philosophy essays.

VI.2.3 Indexicality and Epistemology

It is important to note that indexical accounts have been proffered for the meaning of a key

epistemological concept, namely “know.” While these accounts share many problems with the

corresponding accounts of ethical concepts, they have distinct problems.

Stewart Cohen has been the most explicit in advocating an indexical account of “know”

(Cohen 1986, 1987, 1988, 1998, 2005), though Keith DeRose has forwarded such an account by

simply describing it as “Contextualist” or “context-sensitive” (DeRose 2005b, 2004, 2005a, 1995),

while others have chosen other nomenclatures (cf. Lewis 1996)120. According to Cohen,

A better way to view matters is to suppose that attributions (or denials) of knowledge are indexical or
context-sensitive. The standards that apply are determined by the context of attribution. The truth
value of a knowledge attribution will depend on the… standards that apply in the context of
attribution. In general, the standards in effect in a particular context are determined by the normal
reasoning powers of the attributor’s social group. Thus, I may correctly deny knowledge to S where a
member of the moron society correctly attributes knowledge to S. (Cohen 1986, 579)

For Cohen, whether someone can be said to know that P can only be determined by deferring to

standards that are contextual in nature. Cohen’s account is somewhat like Dreier’s second, societal

version of Speaker Relativism, in deferring to social (and not individual) standards—others, like

David Lewis (1996), provide individualist accounts. However, a common theme that runs through the

work of those who would provide an indexical account of knowledge attribution is a desire to avoid

paradoxes of scepticism, according to which common sense knowledge claims (“I have a hand”) seem

to have prima facie validity, as do sceptical criticisms (“You do not know that you are not a brain in a

vat”) (Davis 2004, 257). Cohen’s solution to this paradox is roughly to argue that (a) a sceptic is

someone with a very high standard against which knowledge claims are evaluated, and (b) if

266
knowledge claims are to be judged by context-sensitive standards, it is possible that some knowledge

claims (for instance, ordinary “common sense” claims about the reality of us not being brains in a vat)

can survive sceptical scrutiny if the relevant standards are lower than the sceptic’s. Alternatively,

DeRose argues that, normally, low epistemic standards are operative in ordinary conversations, but

when the sceptic interjects sceptical criticisms, the context changes and the higher standards prevail

(DeRose 1995, 35-6). This is only a limited victory for the sceptic, for it does not dismiss the

possibility of knowledge altogether.

DeRose argues that such indexical accounts of “know” are supported by ordinary language

practices (DeRose 2005b). Like Cohen (2004), DeRose believes that the indexical account is superior

to an alternative, context-sensitive account called “Subject Sensitive Invariantism” or SSI (DeRose

2004). According to this view, the standards that knowledge claims are judged against do not shift on

the basis of the attitudes and expectations of speakers, but rather are grounded in the background

circumstance of knowing subjects. Cohen explains:

Suppose I say ‘John knows his car is parked in lot 2’. Then I am the speaker/ascriber making the
ascription and John is the subject of the ascription. Contextualism says that the truth-value of a
knowledge ascription is sensitive to whether error possibilities are salient to the ascriber. So if the
possibility that John’s car has been stolen is salient to me, then I speak truly when I say ‘John does
not know his car is in lot 2’. This remains true even if this possibility is not salient to John. SSI
denies that there is context-sensitivity in the above sense of ascriber/speaker sensitivity. Rather, SSI
says that the truth-value of a knowledge ascription is sensitive to whether error possibilities are
salient to the subject of the ascription. On this view, the truth-value of a knowledge ascription is the
same regardless of who is doing the ascribing (thus the name ‘subject sensitive invariantism’). So if
the possibility of error is salient to John, then he fails to know his car is in lot 2 at any ascriber
context. So if we say that John knows, because no error possibility is salient to us, we speak falsely if
some error possibility is salient to him. (Cohen 2005, 206)

Recent versions of SSI can be found in the works of John Hawthorne (2004) and Jason Stanley

(2005), who calls it “Interest-Relative Invariantism.”

267
 From the perspective of TTS, Subject Sensitive Invariantism cannot function as the text-type

of philosophy, for translating philosophy cannot involve presuming, as a regulative feature of

philosophical translation, that there is a subject sensitive, invariant epistemology expressed in all texts

and by all philosophers. TTS, of course, could have no objections to Subject Sensitive Invariantism as

one among many theories in normative epistemology, that prescribes to us the conditions under which

we would be justified in counting someone as having knowledge. TTS is committed to being able to

translate all possible views. But TTS cannot take on SSI as the canonical, text-type-theoretic account

of the concept KNOW for this cannot allow us to translate the unique philosophical perspectives of

diverse philosophers the world over. In this respect Contextualism in its individualistic form (à la

Phillips), for all of its Speaker Relativistic problems, is more amenable to the translator’s interest, for

at least it brings attention to the theoretical commitments of the author.

Jason Stanley has put forward a type of Subject Sensitive Invariantism in semantics. A

distinct feature of this view is that it conceives of much language as concealing a logical form that

makes room for contextual input via “unarticulated constituents” (Stanley 2000). One of Stanley’s

aims is to explain phenomena such as when someone claims that “there is no more cheese” and this is

understood not to mean that the world is bereft of cheese, but merely that there is none in one’s fridge

(cf. Stanley and Williamson 1995). On Stanley’s view, the background circumstance (relative to the

practical interests of the subject) provides a restriction on the domain of the quantifier, and thus

constrains the truth of the sentence (Stanley 2000, 419), just as the background circumstance of an

epistemic subject relative to her interests furnishes the conditions according to which we can judge

her to have knowledge for SSI.

Stanley’s view in semantics thus complements his view in epistemology. And while Stanley

does appeal to seemingly esoteric features of language buttressed by recent linguistic theory, the

evidence to which he appeals to make his case, both for epistemology and semantics, is ultimately

ordinary language—specifically, what seems acceptable to say in English, and what doesn’t. If a

268
sentence sounds strained to the competent speaker of English, Stanley takes that as evidence that the

concepts employed in the sentence have been misapplied or misconstrued (cf. Stanley 2004, 125).

Thus, Stanley’s route of justifying SSI and DeRose’s route of justifying Contextualism are of a piece.

In articulating my argument for TTS in chapters 2, and specifically 3, I criticized linguistic

approaches to meaning. Linguistic approaches to meaning take languages to be the primary bearers of

meaning, and take semantics as charged with explicating the meaning of components of a language.

This approach to semantics was criticised as incapable of explaining translation. The main reason is

this: if semantic equivalence is a criterion of translation and if meaning is primarily linguistic in

nature (and not textual), translation will not be possible for no two languages are exactly the same.

Proponents of systematic semantics, who take a theoretical approach to linguistics and semantics,

might take this to be an ineffectual criticism, for they conceive all human languages as instantiating

the same, underlying structure. However, translation is not simply concerned with preserving an

underlying syntactic structure. Rather, translation concerns the preservation of the whole gamut of

semantic modalities, that stretch from the aesthetic, to the culturally dependent, to institutional text-

type features that structure and organize translations—phenomena that are not purely linguistic or

reducible to grammar or literal meanings of expressions. If translation is the ultimate test for

determinacy in meaning, linguistic theories of meaning, whether they are narrower theories concerned

with articulating the deep structures of languages, or “ordinary language” theories of meaning, can at

best enlighten us about semantic modalities that are indeterminate in significance.

Thus, while an appeal to ordinary language may seem like the ultimate arbitrator in questions

of semantics, from the perspective of TTS, ordinary language (if there is such a thing in today’s

multicultural societies) is far from conclusive. The technical reason for rejecting ordinary language

considerations is that they are a semantic fiction. There are no determinate, semantic rules of

ordinary language that can be preserved in translation into another language. We can translate a text

written in ordinary, non-technical, widely used language, but in subjecting it to a text-type we strip

269
the language of the text of its role as sole arbitrator of meaning. A text written in so-called “ordinary

language” (without technical terms) can all the same be inaccessible to the “ordinary” language user

if they are ignorant about the text-type of the text. (I could, for instance, rewrite this chapter replacing

all the technical abbreviations with drawn out phrases composed of terms from ordinary language. I

suspect the argument would still be unavailable to the average master of “ordinary language.”) If

“ordinary language” were a text of a certain type, specifying features of such “ordinary” texts to be

translated, we could, under TTS, make sense of it as a semantically determinate constellation of

meanings. But if there were such a text-type, considerations relative to it could have no direct bearing

upon questions of how to conceive of the semantics of the text-type features of philosophy, such as

“know” or “good”. However, it is difficult to understand how ordinary language could be a kind of

text, for it is presented as culturally specific and context dependent. It is by definition something that

cannot be translated in another context that would be unordinary by its lights. Ordinary language as a

text-type would be an instance of an untranslatable text-type, which essentially is an incoherent text-

type. The fact that there is no department of ordinariness in the modern academy ought to assure us

that “ordinary language” is not a text-type.

These considerations, and those discussed in detail in chapters 2 and 3, ought to cast a fatal

doubt on the view that there is a principled thing called “ordinary language” that can perform the type

of language policing function that philosophers wish to grant it. The only reason that the idea of

ORDINARY LANGUAGE has had any staying power in the literature is because it is a type of reified

ethnocentrism, lacking semantic foundation but backed by a seductive, cultural solipsism that ignores

our transcultural, transnational text-type institutions.

A confirmation of this political and ethnocentric nature of “ordinary language” is that any

novel idea translated into our “ordinary” language will sound strange because it is not ordinary by our

provincial standards. My own work on the history of Indian philosophy has confirmed this lesson to

me time and again. In my Ethics and the History of Indian Philosophy (Ranganathan 2007a), I argue

270
that “dharma” is in fact a thin moral term. Among the obstacles to appreciating this thesis are the

ordinary language sensibilities of Western trained scholars. “Dharma” has been used for many items

that are not usually thought of as morally significant in the Western tradition. Buddhists, early Vedic

thinkers, and Jains, use “dharma” to designate universal, cosmic laws, principles in general, their own

theories, and in the case of Buddhism, “dharma” was used freely for items in its ontological scheme,

such as constituents of reality that obeyed the principle of co-dependent origination. Against the

backdrop of QI, we can understand that “dharma” was used by these various philosophers to

articulate theories with the same common range of concern that we find with theories articulated with

“ethics” or “moral” in the Western tradition. But it sounds odd to Anglophonic ears to talk of

“ethicals” that comprise the universe as constituents of reality, or “morality” as a cosmic law. Of

course, such uses of value terminology are not unheard of in the Western tradition—Plato thought

that the ultimate reality, or the Form of forms is the Good, and Plotinus thought that matter is Evil,

because it is several steps removed from the Good. Likewise, Utilitarians have taken happiness to be

goodness as such. But the Indian uses of thin moral terms (of which “good” is not clearly an instance,

owing to its axiological ambiguity) seem odd. Deferring to ordinary language as the gate-keeper of

translation and semantics thus serves, in this case, to do nothing but insulate ourselves from alien

views on ethics that offend our sensibilities. Thus, in short, it is no surprise that some constructions

will seem odd to ordinary language sensibilities, but this is simply a reflection of a kind of

xenophobia in the world of ideas, and not really proof of anything determinate in semantics.121

These considerations shed light on additional problems plaguing the indexical account of

“know” proffered by Cohen and DeRose. To the extent that this account is similar to Dreier’s and

Phillips’ account of moral terminology, it is unacceptable as an account of the text-type features of

philosophy. If Cohen and DeRose are right, an epistemologist speaks truly about someone having

knowledge so long as she speaks in conformity to the relevant contextual standards of knowledge.

However, on any specific account, whether it is Cohen’s societal speaker relativism, or DeRose’s

271
sceptic sensitive account, the range of interesting philosophical debate on epistemic matters, as

evidenced by the history of philosophy, shrinks. And if we were to build either Cohen’s or DeRose’s

theories into a text-type account of “know” we could only translate philosophical views that are

consonant with their relativism. For the text-type feature is what we look to, to help us sort out the

ambiguity and potentially conflicting nature of texts in determining their meaning relative to a text-

type. Thus, if Cohen or DeRose is correct, we, as translators, would have to attempt to determine the

relevant standards to an author’s work as the contextual input, disregarding their own philosophical

views on the nature of knowledge, if it should diverge from the putative contextual standard. But this

would do violence to most of the history of Western epistemology, which has only recently taken

Contextualism seriously at all. To take Cohen or DeRose’s view seriously would be to build into the

transcontextual and transcultural concept of KNOW considerations that are particular to what Cohen

and DeRose take to be their context as defined by their “ordinary language.” This would be a great

mistake.

VI.2.4 Indexicality, Key Philosophical Terms and Translation

In summary, the indexical accounts of moral and epistemic terms that we find in the literature

have a common deficit: they overlook considerations of what the translator of a philosophical text

would have to keep in mind in order to accurately produce a translation of a text employing such

terms. Rather, they are provided on the basis of considerations far removed from the question of what

makes a translation determinate. As I have argued in chapters 2 and 3, the question of translation is

central to the semantic enterprise, for translation is our best test of the determinacy of meaning.

272
 VI.3. TTS and Quasi-Indexicality

VI.3.1 The Basic Model and its Variables

QI, as I have been presenting it, is the text-type feature of philosophy—specifically, it is a

model of the semantic aspects of key philosophical words. It is not an attempt to capture what people

ordinarily mean by their use of moral terminology (though, as we shall see later, it can help us in this

regard as well). The question of what people ordinarily mean by almost any concept is a bad question,

I believe. It is particularly misleading in the case of philosophically significant concepts. Ask people

what they mean by “ethics” or “know” and you will receive as many answers as there are substantive

views on the topic (within a culture at least). The only point at which answers to such questions

become informative is on a grand scale, in our effort to find a pattern to the diverse uses of the term in

a culture, taking into account the diversity of such uses. And even at this point, the utility of the data

can only be significant in light of a text-type institution for which “ethics” or “know” is a text-type

feature. Only by reflecting upon the institutional norms that constitute the practice of philosophy, and

seeming difficult cases, such as the problem of translating “dharma” from the history of Indian

philosophy, do “ordinary” answers to the questions of “what does ‘ethics’ mean” become significant.

QI, I submit, is what our institution of philosophy is structured around, and what allows for

the translation of philosophical texts. QI strikes me as the glue that holds together our international

institution of philosophy—an institution that spans several contents and at least three major world

traditions (the Western, Indian and Chinese). It is by virtue of QI that texts as divergent as Aristotle’s

Ethics, the Lao Tzu, the Upanishads and Quine’s Word and Object, can all be taught, without

absurdity, under the common rubric of philosophy in the modern academy.

The discussion in the past section brought to fore the following points:

(a) though there is a type of relativism that is accommodated in the philosophical

text-type that allows authors to choose the theories they wish to defend, apply, or

273
 develop, it is not a relativism that allows for philosophical claims to be self-

fulfillingly true, for this contradicts the argumentative aspirations of philosophy;

(b) the concepts behind key philosophical terms—that is, the semantic aspect of a

key philosophical term that is common to various inscriptions—cannot be

indeterminate in Phillips’ sense—that is, the philosophical concept cannot be

completely indifferent between differing philosophical theories, for this would

allow for the absolute relativity of philosophical theories, which would in turn

undermine the textual objectives of philosophy;

(c) indexical accounts of key philosophical concepts, that make room for contextual

input, such as what theory a philosopher wishes to argue for, must be

generalizable to all key philosophical concepts;

(d) theories, not motivations, must be the relativistic input into the characters of

value and philosophical concepts.

Taking into consideration these lessons, I believe that the key text-type feature of philosophy has the

following general form:

ordinary symbol s, refers to an item x that falls (within/outside) universal and general theory t, that
has a (theory to world/world to theory) direction of fit, which is selected for (fill in axiological
differentia), in accordance with the relevant considerations.

This is the basic model of QI. The character is the textual rule of interpretation that allows the

translator to determine not only the intended referent of the usage of a key philosophical term, but

also the philosophical content of a text that must be preserved in translation. The italicized portions of

the characters are variables to be filled in by contextual and cotextual considerations. Contextual

considerations include such matters as a philosopher’s choice of philosophical theory. Cotextual

considerations are those that arise out of the text itself. Thus, for instance, when a philosopher

presents an argument, she often will respond to objections. Here, the theory that informs the objection

274
articulated with a key philosophical term can be determined on the basis of the semantic modalities of

a text and not simply the personal commitments of the author as they are ex hypothesi different from

those of the author.

The symbol variable s refers to the word that a philosopher uses as the inscriptional vehicle

for the character. In philosophical traditions there will often be conventions surrounding word-

character pairings that have a pervasive effect on language use in a culture—providing that the

institution of philosophy is strong in the culture. Thus, for instance, there are strong associations

between words such as “real” and “ethical” and their conceptual content, reinforced by the institution

of the text-type of philosophy, for this institution is quite strong in English-speaking countries. It is

possible for philosophical texts to employ novel word-character pairings, particularly when an alien

philosophical concern is translated into a culture’s language. Here, I believe, the institution of

philosophy heavily favours the usage of an ordinary, non-technical word as the vehicle for the

philosophical concept, even if the word in question has had no clear history of association with the

concept within a culture. The quasi-indexicality of the text-type of philosophy arises in many cases

out of the sociological attempt by competing groups to wrest control over some semiotic device to

articulate a privileged perspective.

The axiological differentia of a philosophical concept is the value that specifies what can

crudely be put as the philosophical topic of the concept. Due to size constraints, I have left out of this

dissertation a chapter on the question of the axiological differentias of various areas of interest in

philosophy today.

Under the linguistic paradigm of philosophy, philosophers attempt to determine the

axiological differentia of philosophical concepts by looking to linguistic practice, broadly conceived:

speaker intuitions are often given a premium in this project. This is certainly an interminable pursuit

as intuitions are contingent and hardly objective. However, the move to an institutional-textual

approach mandated by TTS leads us to consider the question in reference to the text-type institution.

275
In the case of philosophy, we must look to the institutional space of philosophy as an international

discipline to settle questions of axiological differentia. As QI is not an account of word meaning in

isolation, but rather an account of textual devices, we need to understand axiological differentiation

with respect to theories articulated with quasi-indexical devices that characterize the philosophical

text-type. Theories are already matters of abstraction that are not directly reducible to culture, and in

the case of theories of philosophy, we are interested in theories articulated with terms such as “ethics”

or “morality” in philosophy that are often culled from disparate cultural sources and come to us in

translation. To settle the question of the axiological differentia of a key philosophical term is to

discern the lowest common denominator shared by theories articulated with various philosophically

important words, such as “ethics,” “morality,” “justice,” “beauty,” “real,” “knowledge,” “logical,”

“right,” “wrong,” “good,” and the like. In some cases, it is quite clear that there is no one common

axiological differentia: I think this is the case for terms such as “right,” and “good,” that have been

used in the service of articulating all manner of axiological concern. (We can use “good,” thus, as a

manner of articulating both aesthetic and ethical theories, as we can with “right”, “wrong,” and

“bad.”). To sharpen my thought on this matter, I try to think about how a conference administrator, in

charge of a large philosophical organization, such as the APA, would attempt to come to

generalizations about what “ethics” concerns, given the disparate projects that lay claim to this term

in the conference she is administrating. The position that such a conference administrator occupies is

a place of institutional privilege and I think from there it is important for us to arrive at institutional

answers to the question of the axiological differentia of philosophical theories relative to

philosophical terms. Philosophers often refuse to assume this institutional position in their thinking

about axiology, rather preferring to think about questions of axiological differentiation in a quasi-

normative manner that lays claim to what “we” mean by such terms, but there really is no objective

answer to the question of what “we” mean, when the “we” refers to our cultural cohorts. As I argued

in Chapter 4, such deference to a “we” should be understood as a type of rhetorical move that adds to

276
the semantics of our language but is not exhaustive of its polysemy. Rather, we need to think

institutionally if we are going to get to the bottom of the semantics of philosophy, and if we want

there to be objective answers in philosophical questions. Thinking institutionally does not mean we

give up our other avenues of answering such questions. In the case of philosophical traditions, such as

the Western tradition, that we have been translating and thinking about for a long time with no great

problem, studying the history of philosophy in translation is a way to further our institutional

knowledge of the axiological differentia of philosophical theories. The canons of philosophy are

records of philosophy’s past in light of the current state of our text-type of philosophy. Studying such

canons as part of the means of settling institutional questions is best when the translations of the

canons are not controversial. However, even bad translations of philosophy in controversial canons

(such as the canon of Indian philosophy) does give some indication of greater institutional

considerations in translating philosophical texts.

The challenge before us is like that of answering the question of what it is to be Canadian.

Commentators, politicians and lay Canadians often wrestle with this question, looking for some

property or values that characterize Canadianness. This is how philosophers in the linguistic paradigm

answer questions of what it is for a matter to be “moral,” or “ethical.” There really is no objective

answer to this question. However, there is an objective answer to the question of what institutionally

makes someone a Canadian (citizenship, acquired by birth or through naturalization, and not stripped

by formal processes). In the same way, there is an institutional answer to the question of what makes

a topic a moral issue: originally articulated with “moral” or “ethics” in the institution of philosophy,

or naturalized through a process of translation in such a manner that some theories, in translation, are

articulated with “ethics” and “moral.” The common denominator of all such theories is thus the single

differentia of moral or ethical issues.

This is a tractable problem. In the case of languages for which we have long-standing

conventions of translation that are now part of the bedrock of the institution of philosophy, we

277
identify a corpus not by language but by texts. Thus long-standing conventions of translating Plato

and Aristotle into various modern languages that pair up key philosophical terms must be taken into

account. The parings of key philosophical terms across languages that a corpus has been traditionally

translated across (good::agathos, ethic::ethika, justice::dikaiosune) constitutes the terminological

point of reference that theory articulation has to be judged against and from here the axiological

differentia is determined. To the extent that there are long-standing conventions of key philosophical

term translation that are not controversial among experts, a philosophical corpus judged from within

any given language will provide us with sufficient evidence to determine the axiological differentia of

various key philosophical terms in translation.

I shall now specify what I think will be the outcome of this institutional investigation if we

take the time to investigate the issue (which, as I say, has been omitted from this dissertation for

length considerations. Contrary to some claims, “ethics” and “moral” articulate the same range of

theories in philosophy (the notion that there is something distinct about ethics and morality is a

regional distinction that is not reflected by the diversity of theories that fall under the heading of

“moral philosophy” or “ethics”—two labels that are interchangeable in philosophy today). The

axiological differentia of such theories, I suggest, is that they are chosen for their social implications.
122

Theories articulated with “beauty”, or aesthetic theories, are those that are selected for being

theories on disinterested judgments of experience (cf. Kant 1974 [1790], 39 §2). Theories articulated

with “knowledge” or epistemic theories are those that are selected for their implications for belief.

Theories articulated with “real” or metaphysical theories are those that are selected for being theories

about what is not mere illusion. Theories articulated with “logic” or theories of logic are selected for

being theories about inference. Theories articulated by “meaning,” or theories of semantics, are

chosen for their implications for connectivity (cf. Cooper 2003, 30). Many of these accounts of the

axiological differentia of key philosophical terms will likely seem at variance with so-called

278
“ordinary language” and the substantive goals of particular philosophical programs. But our goal in

specifying the text-type feature of philosophical texts is not to capture ordinary language or to

privilege any particular philosophical theory, but rather to capture the institutional space in which

varying philosophical theories on a certain axiological matter are debated. Many of these suggestions

for axiological differentias will also overlap, allowing metaphysical, epistemological and moral

theories to overlap within one greater, systematic theory. This might seem odd in today’s

philosophical landscape, but for systematic philosophers, one and the same super-theory will be

articulated by several key philosophical terms of differing axiological differentias. Traditionally, this

was the primary mode of philosophical argumentation, and our institution of philosophy, in

conserving texts of systematic philosophers, allows for such overlap.

QI as I conceive it is not wedded to any particular account of axiological differentias, though

it is wedded to the idea that we require axiological differentias in order to differentiate the semantics

of the various types of philosophical concepts. I present the above suggestions as educated

suggestions. The types of considerations that would be relevant to revising them, however, are not

likely to be the types of considerations that are usually provided. If they suffer from a defect, it must

be not excessive breath but excessive narrowness. Our goal in providing such accounts must be to

explain how philosophers with diverse views on epistemology or ethics can all the same be involved

in the philosophical enterprise. Thus, if a very successful or prominent philosophical perspective is

forbidden by any of the above axiological differentias, we would have good reason to revise our
123
account in favour of including the successful or prominent school. It is important to note,

however, that institutional norms that we are deferring to in assessing the axiological differentia

associated with words also exert a strong institutional pressure on authors in the practice to conform

to institutional norms. Thus, while there is room for a great diversity of philosophical views within

the text-type of philosophy, there is also strict pressure not to violate conventions of paring key

philosophical terms with axiological differentias. Thus, “real”, for instance, is unlikely to be used by

279
anyone to be the key term by which they articulate an epistemology, for instance. And translators will

feel a great pressure to accommodate strange philosophical views within the scope of conventional

term-concept parings.

The variable of ‘within/outside’ controls whether the concept is positive or negative in

flavour. Positive moral terms include “ethical,” “moral,” and “kindness.” “Real”, “know,” “logic,”

and “beauty” are also examples of positive philosophical terms. Negative terms include “unreal,”

“ignorance,” “illogic” and “ugliness.” If a term is positive, it refers to an item that is putatively

mentioned in the theory that is used to inform reference. If the term is negative, it picks out an item

that falls outside of the theory.

A related variable is that of direction of fit. David Phillips calls upon the idea of direction of

fit to explain the normativity of value judgments such as “x is good” (cf. Anscombe 1963 [1957];

Humberstone 1992; Smith 1994, 111-119; Velleman 1996). According to Phillips, value judgments

have a direction of fit lacking in ordinary descriptive judgments, which accounts for their distinctive

normative quality. Specifically, value judgments have a world to judgment direction of fit, while

descriptive judgments have a judgment to world direction of fit (Phillips 1998, 142), where “world” is

understood to designate not the totality of all that is real, but rather the contingent, empirical world.

While I think there is much to be approved of in Phillips’ employment of this analysis of normativity,

it is too strong a position. While ethical theories are arguably normative, it is far too strong to suggest

that every judgment employing an ethical term is normative. If direction of fit is to elucidate the

normative nature of some key philosophical concepts, it must pertain not to the judgments that are

generated from such concepts, but the theories employed. Normative concepts, on this account,

employ theories that are taken to have a world to theory direction of fit. In other words, in such cases

it is the contingent world that must measure up to the theory.124 This contrasts sharply with so-called

descriptive theories, where the theory is judged according to its ability to live up to the contingencies

of the world. Thus, in short, value concepts, such as those of ethics and aesthetics, will designate a

280
theory that has a (contingent) world to theory direction of fit. Metaphysical concepts, in contrast, will

specify theories that have a theory to contingent world direction of fit. This is not to say that every

aspect of a metaphysical theory is concerned with the empirical world, but merely that qua

metaphysical theory the contingencies of the world are considerations against which it can be judged

as adequate.125 On this account, direction of fit is a systemic property of a philosophical theory on the

whole, not necessarily reducible to its various constituents—though, I suspect, in the cases of

plausible philosophical theories, there will be a preponderance of support within the theory itself for

its putative direction of fit.

This account might appear to render puzzling the overlap between metaphysical theories and

value theories that is so prevalent in the Platonic tradition and the Indian and Chinese philosophical

traditions (such as in the case of Jainism, Buddhism and Taoism). The “Good” for Plato, for instance,

is both a superlative evaluative entity and metaphysical principle. However, it would seem that the

account I am forwarding would make it impossible for one item designated by a value concept or

metaphysical concept, as the case may be, to play both evaluative and metaphysical roles for the

direction of fit of metaphysical and value theories are polar opposites. This would be a grave problem

if this were an implication of the theory I am putting forward. However, I believe it is not. It is

important to keep in mind that every employment of a philosophical concept, according to QI, does

not necessarily generate a judgment with clear normative or descriptive direction of fit. Thus,

normative concepts can, according to QI, be pressed into the service of generating judgments with

indeterminate normative significance, such as “The Form of the forms is the Good” just as they may

also be employed in generating descriptive claims, such as “That kind man came by today to pick up

our donation for the food bank.” A full understanding of the latter descriptive sentence, or the

normatively indeterminate sentence, would involve a near-thorough appreciation of the value theories

employed in the use of the terms “kind” and “Good,” with all of their normative implications, but the

resulting sentences do not have any clear, explicit normative force. The same may be true in the

281
converse case, where a seemingly non-evaluative, philosophical concept can be used to identify a

value (such as, for instance, “the Form of the forms”). In short, one item, such as the Platonic Good,

can play both an evaluative and metaphysical role, and its normative and metaphysical implications

are brought to fore in differing philosophical circumstances depending upon the textual salience of

philosophical theories under discussion, regardless whether it is identified by a value concept (GOOD)

or a metaphysical concept (FORM OF THE FORMS). This may result in a certain philosophical tension

that is often resolved, within such theories, but construing normative concepts as having an intrinsic

representational function whereby they conform to a necessary, evaluative order that is theoretically

distinct from the contingencies of specific times and places, allowing such evaluative and

metaphysical concepts to be coextensive while maintaining a normative distance between the

contingencies of life and normative theory. This putative representational function would be part of

the content of such metaphysical and normative theories (such as in Taoism, or Buddhism) and not

part of moral semantics in any determinate sense.

A distinctive aspect of QI that sets it apart from other indexical accounts of value concepts is

what I call, for lack of a better phrase, the “objectivity clause”. This is the phrase that is a part of the

character of all key philosophical concepts that states that the referent of the word is selected “in

accordance with the relevant considerations.” The reason that I call this the “objectivity clause” is

that it reminds us of the unique character of philosophical discourse. Philosophical discourse is not a

mere subjective inventory of likes and dislikes, feelings and impressions. It has a dialectical structure,

captured in QI, that aims at objectivity. What this objectivity amounts to is typically a matter of

contention among philosophers, and hence the notion that considerations are “relevant” captures just

this seemingly contentious aspect of philosophical argument in actual cases. Yet, the notion of

relevance also retains the notion of objectivity in so far as the goal of relevance can only be

understood as seeking to avoid the irrelevant. Thus, philosophers usually differ as to what they take to

be philosophically relevant, but a commonality is that such considerations are taken to be objectively

282
persuasive, and not simply subjectively pleasing. This is basic institutional knowledge that

philosophers often have to impress upon their students: a philosophy essay is not a personal journal

entry that one might keep to show one’s therapist, nor is it an opportunity to stand on a soap box and

tell the world about one’s feelings.

The outcome of recognizing the objectivity clause’s place in philosophical concepts is that

that propositions constructed with philosophical concepts cannot be determinately true simply by

employing such concepts to correctly identify items included within an agent’s theory. For while an

agent may accurately represent what their theory holds on the issue of x, the claims may all the same

be false if they are not made on the basis of all of the relevant considerations. What I have called the

“objectivity clause” prevents the churning out of self-fulfillingly true propositions of the sort that

characterize Dreier and Phillips’ account of ethical concepts. It also provides the translator with a

semantic tool to assess what the philosopher she is translating thinks is relevant. For, minimally, the

text authored by a philosopher in conjunction with an accurate assessment of the theory that the

philosopher defers to provides a vivid picture of considerations that the philosopher she is translating

regards as relevant, and these features as they are present in the ST are thus the very features that

ought to be preserved in a semantically equivalent TT. This opens the door to a type of application of

the principle of charity in interpretation constrained by the philosophical convictions of the ST author.

Finally, it is important to recognize that QI does not demand that textual employments of key

philosophical terminology are necessarily guided by a theory. A distinct type of dialectical activity

treats philosophical concepts as matters to be elucidated by philosophical inquiry. Here, the theory

valuable is left blank, and philosophers endeavour to arrive at the appropriate philosophical theory by

shifting through a range of considerations that are regarded by the philosophers as relevant to theory

selection. However, such a dialectical inquiry all the same relies upon the general model of

philosophical concepts that QI identifies, in so far as it operates via the components of the objectivity

clause and the theory variable. In the dialectical case, philosophers merely presume, or argue for, an

283
answer to the question of what is relevant to theory selection and leave theory selection an open

question. This is the inverse of the paradigm case, I believe, where philosophers choose theories and

wade open-endedly into arguments about relevance to theory selection. In the case of the dialectical

philosophical text, the translator must come to an assessment of what the author regards as relevant to

the type of philosophical theory under consideration. But this will be dictated in large by the content

of the text itself, filtered through the lens of QI. Translation will thus consist in recreating a TT that

preserves this sense of philosophical relevance—including, arguably, the dialectical methodology of

the ST employed to arrive at an account of the settled philosophical theory.

VI.3.2 Thin Terms

The model presented so far outlines the semantics of so-called “thin” terms, and with a small

modification can accommodate so-called “thick” terms. On the basis of the above considerations, I

propose that the character of a thin, positive moral term such as “moral” is:

ordinary symbol s, refers to an item x that falls within a universal and general theory t, that has a
world to theory direction of fit, which is selected for its social implications, in accordance with the
relevant considerations.

If this is the correct account of the conceptual content of a maximally thin, moral, positive term, there

may be many possible terms bearing this concept, but only one such concept. On the strength of my

earlier research in Indian philosophy, I am satisfied that “dharma” has this conceptual content. Some

philosophers have attempted to argue that “ethics” and “moral” stand for distinct concepts. From what

I can see, the character is the same in both cases. What we find in writers such as Bernard Williams

who are insistent upon a distinction between these terms at the conceptual level, is a willingness to

employ one and the same character in connection with differing theories, and differing senses of

relevance. The insistence that there is a conceptual difference between MORAL and ETHICS seems to

me to be a brief fad, and not the rule in our institution of philosophy. The translator of philosophy can

284
easily accommodate William’s idiosyncratic use of these terms in his Ethics and the Limits of

Philosophy in her translation of this text by recognizing the distinct theories and criteria of theoretical

relevance employed relative to the two terms “ethics” and “moral” and by allowing these contingent

associations to guide her in translation. In other words, there can be token-reflexive considerations in

the employment of philosophical concepts that are not part of the semantics of the concept, but a

result of cotextual considerations that arise out of the philosophical arguments of authors.126

A negative, thin, moral term, such as “evil” has the character:

ordinary symbol s, refers to an item x that falls outside a universal and general theory t, that has a
world to theory direction of fit, which is selected for its social implications, in accordance with the
relevant considerations.

This character distinguishes “evil” from terms such as “wrong” or “bad” by virtue of the inclusion of

a distinctly moral axiological differentia, which seems to be lacking in the concepts of WRONG or BAD.

There is a tradition that draws a distinction between natural and moral evil. I am unpersuaded

that any such difference is actually reflected at the level of the text-type feature of philosophy. Rather,

such differences, like the difference between uses of “moral” and “ethics,” can be accounted for by

the philosopher’s differing choice of theories in differing contexts, or differing conceptions of what is

relevant to theory selection. In the case of so-called “natural evil,” there is inevitably a violation of an

expectation that authors have about ideal social circumstances. The axiological differentia does not

change, though the cause of the evil appears different. In the “natural” case, it is nature, and in the

“moral” case, it is an agent.

“Good”, often thought to be the paradigm of an evaluative term, seems to be distinct among

axiological terms in not being clearly associated with any particular axiological differentia. Rather,

authors usually have to make a point of explaining the axiological differentia of the theory that they

are articulating with “good” for this is not readily apparent by the mere occurrence of the word.127

Thus, its character, while mostly identical to that of a thin, positive moral term, would lack a place for

285
an axiological variable. It seems that it can, without absurdity, be used in connection with any

normative concern, though here again there is the question of whether claims about goodness, such as

“x is good,” are not self-fulfillingly true. Their truth will depend upon the relevant considerations,

which include, minimally, the qualifications of the selected theory as a theory with a world to theory

fit. In other words, there is a greater pressure to justify the normative credentials of a theory invoked

to explain the truth of judgments of goodness, than, say, theories invoked to explain judgments

employing “ethics” or “moral”.

I would argue that “right” must be regarded as a terminological variant of “good”, and that

its differing usages, such as in the contrast between the right and the good, can be explained by the

injection of token reflexive considerations in what philosophers consider are relevant reasons for

theory selection—the same type of token reflexive considerations that explain differing employments

of “ethics” and “moral” by the same author. This conclusion might seem surprising, however it was

argued that if “right” and “good” tokenned distinctions at the conceptual level, either “right” or

“good” would not be thin terms at all, but thick, or it would seem that “right” and “good” differ with

respect to their axiological differentia. However, neither seems plausible. Also, it is vital to note that

the difference in the employment of these terms in the work of deontological authors who are

associated with a distinction between the right and the good is not as clear as commonly thought

(especially in Kant 1997 [1788], 5:59-81; cf. Ross 1927; Ross 1963 [1930]). There is a slight

convention for preserving “right” for deontological matters, and “good” for consequentialist matters,

but this is a very week convention that philosophers do not universally stick to (even when they claim

that there is an important distinction to be drawn between the right and the good). The considerations

for us after all must defer to all usages of these terms in the institution of philosophy.128

The character of other thin value terms can be generated from the model of “ethics” and

“evil” by substituting the relevant axiological differentia. Thus, “beauty”, being positive in evaluative

modality, has the character of:

286
 ordinary symbol s, refers to an item x that falls inside a universal and general theory t, that has a
world to theory direction of fit, which is selected for its implications for disinterested experience, in
accordance with the relevant considerations.

In keeping with this pattern, and the previous discussion, we can specify the character of

“real” as

ordinary symbol s, refers to an item x that falls inside a universal and general theory t, that has a
theory to world direction of fit, which is selected for its implications as a theory about not mere
illusion, in accordance with the relevant considerations.

QI also explains the text-type-theoretic significance of a range of verb phrases. Verbs, as a

category, are not usually thought of in terms of indexicality. However, QI understands these verb

phrases as explainable in an analogous manner, by changing the character from one that specifies

reference, to one that specifies that an argument either satisfies (in the case of the positive modality)

or fails to satisfy (in the case of the negative modality) the conditions set out in the theory. Thus,

“knows that” has the character of:

ordinary symbol s, indicates that an argument a satisfies conditions specified by a universal and
general theory t, that has a (world to theory) direction of fit, which is selected for its implications for
belief, in accordance with the relevant considerations

where “s” in English is “knows that” and the argument a is either an item de re or de dicto that is

known. “Knowledge”, in contrast, will be given a straightforward, nominal account like “moral” or

“real”, such as:

ordinary symbol s, refers to an item x that falls inside a universal and general theory t, that has a
(world to theory) direction of fit, which is selected for its implications for belief, in accordance with
the relevant considerations.

Finally, “ought to” has the character of:

287
 ordinary symbol s, indicates that an argument a satisfies conditions specified by a universal and
general theory t that has a world to theory direction of fit, in accordance with the relevant
considerations

where “s” in English is the phrase “ought to” and the argument a is some action mandated by a value

theory. As in the case of the character of “good,” I take it that “ought” underdetermines its axiological

relevance. The very fact that philosophers need to make a specific effort of clarifying their topic when

discussing what ought to be done (whether they are speaking of a “moral” ought or an “aesthetic”

ought) suggests not that there are distinct types of “oughts” but that “ought” can be used to articulate

theories of varying axiological significance.

VI.3.3 Thick Terms

So called “thick” terms, prominent in the case of value terminology, but also evident in non-

evaluative matters (consider the metaphysical distinction between the relatively thin concept of REAL

and the thicker concept of ACTUAL), have an additional feature that distinguishes them from so-called

thin terms. On the standard model in the literature, thick terms combine an expressed evaluation with

empirical reference (Williams 1985, 129-131, 1996 [1995]; cf. Tappolet 2004; Burton 1992).129 A

variant of this view is that thick concepts are a non-divisible amalgam of the evaluative and

descriptive modalities of meaning, however indeterminate the modalities may be. TTS can have no

truck with such accounts for they are either semantically opportunistic, or parasitic upon an

Expressivist account of value semantics which, I argued in chapter 5, is not up to the task of

providing an account of the text-type feature of philosophy. As well, they do not generalize to

philosophical terms on the whole. QI, in contrast to Expressivism, provides a criterial account of

philosophical and evaluative concepts. It can thus account for the semantics of thick concepts by

recognizing that they build into the character of a philosophical concept an extra criterion that

constrains reference, which I simply call the extra-theoretical constraint. “Cruel,” thus, on top of

288
being a negative moral term, has the extra-theoretical constraint of referring to items that inflict

excessive suffering:

ordinary symbol s, refers to an item x that inflicts excessive suffering, which falls outside a universal
and general theory t that has a world to theory direction of fit, which is selected for its social
implications, in accordance with the relevant considerations.

Likewise, “kind” could be explained as:

ordinary symbol s, refers to an item x that has the disposition of nurturingness, which falls within a
universal and general theory t that has a world to theory direction of fit, which is selected for its
social implications, in accordance with the relevant considerations.

There may be an urge to say more about concepts such as KIND and CRUEL, in the way of

explanations as to why excessive suffering is evil, or why dispositions to nurture are morally good,

but these are the types of explanatory roles that moral theories play.

If duty, in general, is simply what one ought to do, then it seems that DUTY is a thick concept

that is conceptually based upon the character of OUGHT. “Duty” has the character of:

ordinary symbol s, indicates that an argument a, which is an action, satisfies conditions specified by
a universal and general theory t that has a world to theory direction of fit, in accordance with the
relevant considerations.

DUTY, so conceived, is a thick concept that is axiologically underspecified and may be used to

articulate theories that are also aesthetic, epistemic, or moral, as the case may be.

VI.3.4 Thick-Thin Distinction and “Justice”

Finally, there are a group of concepts whose status relative to the thick-thin distinction is

questionable. Samuel Scheffler has argued that JUSTICE is one such concept that is neither obviously

thin nor thick according to the traditional Expressivist manner of making the distinction. The ordinary

understanding of justice regards it as applicable to a narrower range of matters than “ethics” or

289
“morality” though it has the theoretical importance of such thin terms. Scheffler concludes that the

division of our moral concepts into two categories of thick in thin is thus an oversimplification

(Scheffler 1987, 417-418).

In rejecting the traditional two meanings approach to accounting for thick concepts as

parasitic upon Expressivism and for being semantically opportunistic, one loses much of the

motivation for regarding JUSTICE as a thick concept. On the traditional two meanings account, thick

concepts have empirical content, while thin concepts do not. For QI, the differentia between thick and

thin concepts is their role in theory articulation. All philosophical concepts, thick or thin, can be used

to articulate a theory, but only thin concepts articulate philosophical theories at a level of generality

characteristic of its first principles. Thus, while we can use CRUEL to articulate a moral theory in a

work on applied philosophy, we cannot explain the moral theory at its most general level in terms of

“cruelty”. Rather, we must have recourse to concepts whose meaning does not include extra-theoretic

restrictions. On the basis of this manner of distinguishing between thick and thin concepts, JUSTICE is

patently thin, as there is a long tradition, stretching back to Plato, of conceiving justice as carving out

its own theoretical space. My sense is that theories of justice (whether it be a theory of justice in the

soul and the city state from Plato, or in law and politics that we find in Rawls) have to do with

rightness in organization. If this is correct, the axiological differentia of JUSTICE specifies that the

theory that this word articulates is chosen for being a normative theory about organization. Hence, the

character of “justice” is:

ordinary symbol s, refers to an item x that falls within a universal and general theory t that has a
world to theory direction of fit, which is selected for its implications for organization, in accordance
with the relevant considerations.

If this account is correct, JUSTICE is not obviously a moral concept, though philosophers are always

free to account for their theory of justice in terms of a theory they also consider to be sufficient as an

ethical theory, and as the concept underdetermines the issue, it may be that the best theory of justice

290
will also be identical with the best theory of ethics or morality, or there may be some overlap between

moral theories and theories of justice (a very popular recent line of thought since Rawls’ Political

Liberalism).

My final comment on determining whether a concept is thick or thin concerns methodology.

If TTS and QI are correct, we cannot make decisions about the meaning of conventional

philosophical terms by merely looking at ordinary language. Rather, we look to our institution of

philosophy first, and moreover wherever there are strong and settled conventions for translating a

philosophical term from one language into another in philosophical texts, this text-type convention

takes precedence over seeming ordinary language intuitions. It is not always the case that there are

settled conventions. For instance, in the case of “dharma” into English, or “ethics” into Sanskrit, there

are none. However, in the case of “justice,” there is, namely its translation into the Greek

“dikaiosune” or “dikaiosyne”. In deciding upon the character of “justice,” we are thereby also

deciding upon the character of “dikaiosune” or “dikaiosyne” as per its meaning in philosophical texts.

We can authoritatively arrive at decisions on this matter by simply surveying our texts, even in

translation, for ex hypothesi they are semantically equivalent to the ST (so long as they have the

backing of our institution of philosophy, and in the case of translations of the ancient Greeks, we are

on quite solid, uncontroversial ground). Such conventions do not rule out the possibility of

idiosyncratic usage of philosophical terminology in philosophical texts in any language, but

idiosyncrasies are still judged harshly in relation to our settled conventions, when the conventions are

strong. Indeed, if one of the purposes of the philosophical text-type is to articulate and debate

philosophical theories, wilful disregard for text-type terminological conventions will be a programatic

strike against a philosopher. And thus, there will in general be efforts made to understand the

seemingly idiosyncratic text as an unusual but intelligible instance of the conventional, text-type

meaning associated with the relevant philosophical terms. QI is flexible, and its strength is that it

allows translators to exercise an interpretive charity when interpreting texts as a preamble to

291
translating them, in accordance with institutional conventions of philosophical texts. Given the

context and cotextual flexibility of QI, and its institutional grounding, QI provides room for charity

without thereby assuming agreement on substantive philosophical issues.

VI.4. Criticisms and the Distinction of QI

The critic might wonder why I choose the label of “Quasi-Indexicality” for my account of the

text-type features of philosophy, when the account that I have provided places so many restrictions

upon reference. If it were truly indexical, we could expect that speakers are free to determine

reference at their whim, one might think. The reason is that, on my account, philosophical concepts

are best understood as pointing words, and indexicals are pointing words, just like one’s index finger

is a pointing finger. So understood, the meaning of key philosophical concepts can only be

understood, I believe, if we take into account their pointing function, that situates the topic of their

texts in a nexus characterized by a philosophical perspective and intended frame of reference. Key

philosophical concepts are employed in philosophical texts to single out and identify matters of

philosophical importance—and the matters that people wish to identify, and those that they

successfully identify, vary according to a host of factors. The complexity of the considerations that

inform this pointing does not cancel out the indexical function of such concepts, but it does imply that

success in communicating one’s philosophical perspective across contexts (i.e., through translation)
130
does not guarantee that one speaks truthfully or correctly.

The critic might also charge that the criticism I levelled against Dreier’s, Phillip’s, Cohen’s

and DeRose’s accounts equally applies to my QI account. Against the accounts of these four

philosophers, I argued that if we take their accounts seriously as accounts of the meaning of moral

and epistemic concepts, we would have to translate philosophers as though they subscribed to the

views that these four put forward, but this is not true. However, couldn’t we say the same of QI: if it

is the correct account of philosophical concepts wouldn’t we be committed to translating philosophers

as though they believed that QI was the account of the meaning of philosophical concepts?

292
 QI, in recognizing the philosophical freedom of philosophers to choose their theories, could

not be confused with Cohen and DeRose’s account, or Dreier’s societal version of Speaker

Relativism, for QI is explicitly deferential to the theoretical commitments of philosophers, whereas

these theories are not. So QI cannot be guilty of attributing the type of context-sensitive relativism to

philosophers that would have to be the case if we deferred to these three proposals. However, Dreier’s

and Phillip’s personal relativistic accounts are far friendlier to the type of philosophical relativity QI

recognizes. But yet, I argued that Dreier’s account cannot be correct for if we employed it in the task

of translating Plato, we would have to attribute to him the doctrine of Speaker Relativism, which he

clearly rejects. The problem with Dreier’s account, as noted, is that he recognizes a motivational

system as the input of the character of moral vocabulary. If we replace this with a theory-oriented

account, then we arrive at an account that is similar to Phillip’s standards-oriented account of moral

vocabulary, except that QI generalizes to philosophical concepts as such. Is this all it takes to avoid

the trap that I found so objectionable in the indexical accounts from the literature? For, even if QI is

deferential to a philosopher’s choice of philosophical theory, are we not committed to attributing the

doctrine of QI to philosophers if we employ it to in our efforts to translate them? I do not think so.

The main reason that QI avoids attributing to philosophers doctrines that they do not hold is

that it has no explicit stake in how individual philosophers define philosophical terminology. Indeed,

the activity of giving definitions of value terminology is part of the philosophical enterprise and any

determinate account of how to translate value terminology must allow for this type of freedom. TTS’s

account is thus not to give a linguistic definition of value terminology, but to understand such

terminology in terms of their function within philosophical texts. QI attempts to capture this

functioning, and TTS attempts to preserve this functioning across translation. In contrast, Phillips and

Dreier are in fact attempting to give us definitions of value terminology that would explicitly compete

with the definitions of philosophers we wish to translate. Their account must be understood in this

manner, for it is not proffered as the text-type feature of philosophical texts, but as an analysis of

293
what terms like “good” mean in English. If Dreier’s account is correct, thus, Plato’s view on the Good

cannot be correct. However, if QI is correct, Plato’s view may be correct, so long as judgments of

goodness employing Plato’s theory satisfy the relevant considerations, which itself is the subject of

controversy.

It is important to note that while QI is not primarily in the business of substantive philosophy,

it makes clear the cross-platform basis from which philosophical debate and criticism is intelligible.

This points the direction to a certain type of philosophical engagement that is informed by QI.

Against the background of QI, the most promising lines of substantive argument in philosophy are to

demonstrate a practical incoherence between a philosopher’s stated position and their commitment to

articulating the position as a philosophical position, which is intelligible and hopefully persuasive

across contexts. Philosophers have always done this, to the extent that they have chiselled away at

their opponents by making sustained arguments that appeal to some putative objective standard of

relevant considerations. QI, however, clarifies the prospects of this project. It specifies, for instance,

such features as the axiological differentia and cross-theoretical constraint as sticking points against

which philosophical positions can be criticized.

The only substantive issue that QI might be said to have any stake in is semantics, in so far it

is dependent upon TTS, which does have a clear view in semantics: texts are the primary,

determinate, semantic artefacts, while language is indeterminate in semantic significance. If TTS did

not take such a stance, there would be continued obstacles to appreciating how translation can allow

for semantic equivalence, the normative ideal that separates determinate translation from other forms

of semiotic conversion, such as the production of adaptations, or interpretations. But in taking this

substantive stand in the realm of semantics, TTS and QI can still provide accurate translations of

competing semantic positions for in setting themselves out as textual, as opposed to linguistic or

personal accounts of meaning, their concern is not primarily to refute the substantive views of

semanticists, but to preserve the content of their views through translation, which they affect on the

294
basis of the correct account of the axiological space of MEANING, and the vagueness of the objectivity

clause that at once allows authors to decide what is relevant to such topics as “meaning” but leaves

room for such views to be judged on their own merits by objective considerations (whatever the

extent of such considerations). Even when QI has implications that contradict the substantive view of

the authors it helps translate, it can facilitate this because it sets its anchor in the job of the translator.

Couldn’t Dreier and Philips for instance, argue that their account can help us translate moral

discourse even though it contradicts substantive views in ethics? Couldn’t they also claim to have

anchored their position in the job of the translator? I do not believe so. Many philosophers, from
131
R.M. Hare on to Dreier, have fancied that their theories could explain translation and cross-cultural

communication on moral issues. However, because their semantic theorizing was essentially

linguistic, their philosophies do not confront the problems of semantic and translational

indeterminacy that TTS does. The only way to solve these thorny problems of translation and

semantic determinacy, as I have argued in chapters 2 and 3, is first via LET, and its semantic aspect,

TTS. Once the issue of translation is clarified as a textual issue, and not a linguistic issue, then the

question changes from, “what do individuals mean when they use words such as ‘good’” to “what is

the text-type feature of philosophy” for value discourse is nothing if not philosophical? If the question

is thus one of how to translate philosophy, the account must (a) generalize to all philosophical

concepts that are text-type features of philosophy, and (b) capture the unique features of philosophical

discourse across traditions. Prior theorists who have attempted to accommodate variability in the

content of value judgments (such as Expressivists and the Indexical theorists studied in the previous

section) have never taken this route, and have always approached the issue by attempting to explain

what individuals as language users mean by their use of value concepts, such as GOOD. The project

has thus been conceived as a project in the philosophy of mind or linguistic philosophy, not textual

semantics.

295
 TTS, in recognizing two types of meaning—the indeterminate meaning of languages and the

determinate meaning of texts, their components and features—solves the problem that has troubled so

much theorizing in metaethics: the problem of how to accommodate the obvious theoretical diversity

we see in evaluative issues across contexts, without compromising the objective aspirations of the

discourse. Philosophers have to date either opted to affirm objectivity at the expense of diversity

(Non-Analytic Naturalism) or diversity at the expense of objectivity (Expressivism and Speaker

Relativism). To adopt either route is to make philosophy unintelligible.

Finally, the critic might object: if TTS and QI are not explanations of what people mean, how

can they claim to provide an account of philosophical concepts, or the conditions under which

philosophical claims can be judged true? The response to this objection is to note that QI, in

providing an account of how to translate philosophy, provides a way to extract the determinate

content of a philosopher’s claim, and to this extent is an account of the meaning of what people

commit themselves to through their philosophical writing, though they themselves may not be

inclined to put things the QI way. Within these constraints, it provides a means of assessing the truth

or falsity of philosophical claims. Just as TTS and QI employ two standards of meaning (the

indeterminate, philosopher-relative, linguistic meaning and the determinate, a-contextual, text-type

meaning) one might interpret it as outlining two bases for the evaluation of the truth of philosophical

claims: the speaker-relative, and the text-type relative. On the philosopher-relative account, Plato

speaks truthfully when he says that “Good is the Form of the forms” merely because he defines

“good” in this manner. But no philosopher wishes to have the veracity of their claims judged this

way. Rather, philosophers strive to have their positions judged in relation to objective considerations

of relevance, whatever they may be. QI is the condition of such truth assessability.

Part of what might make this account of key philosophical concepts seem incredible if

viewed under the paradigm of indexicals is that, according to the traditional theory of indexicals that

we learn from Kaplan, a distinctive feature of indexicals is that they can easily be employed to

296
express certain necessary truths that do not obviously obtain in the philosophical context. Moreover,

according to Cappelen and Lepore, there are linguistic tests that we can employ that separate true

indexicals off from non-indexical expressions—tests that Cappelen and Lepore think philosophical

terms fail. I shall address these criticisms in the next chapter.

297
VII. Transcending Contexts

In the previous chapter I set out QI and distinguished it from indexical accounts of

philosophically important concepts in the literature. The account I provide is distinguished by being

based upon TTS, while the contrary accounts in the literature are not proposed as accounts of specific

types of texts. Rather, they are provided against the backdrop of the linguistic paradigm. QI had

already made its way into the conversation. In chapter 4 it was shown to be able to facilitate

translation in the face of linguistic diversity. Also, I argued that unlike competing naturalist

approaches to normative discourse translation, QI can preserve the normative and evaluative content

of texts in translation for it mediates translation through normative and evaluative theories, while a

contrasting translation of a text as a species of scientific discourse does not. In chapter 5, I argued that

the failures of Expressivism in normative discourse translation vindicated QI, for Expressivism fails

at exactly the point that QI succeeds. QI treats normative and evaluative discourse as a species of

philosophical discourse, while Expressivism does not.

While the alternate indexical account of philosophical terms in the literature are not so

obviously contemptuous of philosophy as Expressivism is, they too do not capture the philosophical

functioning of key philosophical terms. The reason that they do not capture this functioning is that

philosophy is a type of text, but the standard accounts of philosophical terms as indexicals explain

them as devices of linguistic communication within contexts.

While QI is very different from the options in the literature, it may still seem to fall to

criticisms of indexical accounts of philosophical terms by virtue of its quasi-indexical nature. I think

this is a mistake. QI and TTS are immune from the usual criticisms of indexical accounts of

philosophical terms. Moreover, I shall argue that only TTS can account for how meaning is objective

and transcends contexts.

298
 VII.1. Contextualism vs. Cotextualism
Before responding to concerns of Cappelen and Lepore about Indexical-type accounts of

philosophical concepts, and other potential criticisms that they might have against TTS, on which QI

is based, I wish to clarify the nature of TTS by contrasting it with Contextualism. In Cappelen and

Lepore’s recent and influential book, they take aim at two related perspectives on semantics:

Contextualism (what they call “Radical Contextualism”) and a highbred view they call “Moderate

Contextualism.” If TTS were a version of either of these semantic theories then it might be vulnerable

to Cappelen and Lepore’s criticisms.

Contextualism, or what Cappelen and Lepore call “Radical Contextualism,” is a view with

historical roots in the writings of the later Wittgenstein and J.L. Austin. In short, this view holds that

meaning as such is indeterminate or underdeterminate, except relative to a context of linguistic usage.

Contextualists typically hold that it makes no sense to ask what an expression means independent of a

context. Contextualists express their view either through the doctrine that sentences can,

independently of context, at best form a propositional fragment or radical that requires saturation by

features of context to render a fully meaningful entity that is what is said. Or, Contextualists express

their characteristic doctrine by regarding semantics and pragmatics as intertwined projects.

Contemporary versions of this purist version of Contextualism include the “Truth Conditional

Pragmatics” of François Recanati (2004), The Relevance Theory of Deirdre Wilson and Dan Sperber

(1986), and the philosophy of other authors influenced by Wittgenstein and Austin, including John

Searle (cf. 1978, 1980), and Charles Travis (cf. 1985, 1996).

Contextualism is a species of what I have earlier called the linguistic theory of meaning. The

defining characteristic of linguistic theories of meaning is that they make bits of language, and not

texts, the primary bearers of meaning. (Recall, a text need not be composed of language, according to

TTS. It may be comprised of icons, non-linguistic, ideographical orthographies, or even musical

299
notation, to name a few examples.) Contextualism qualifies the linguistic doctrine of meaning by

insisting that it is only language use in context that has a determinate meaning.

TTS is not a version of Contextualism. The main differentia is that TTS, unlike

Contextualism, regards texts and cotextual features, such as text-type features, to be the primary

bearers of meaning, not languages. However, to understand the distinction between TTS and

Contextualism, it is important to contrast the now conventional meaning of “context” from what I

have been calling “cotext.”

“Context” has come to mean something very specific in the philosophical, linguistic and

anthropological literature that is a departure from its traditional, etymological meaning of “with text”.

According to the Oxford English Dictionary, “context” in the 16th century CE meant, “The whole

structure of a connected passage regarded in its bearing upon any of the parts which constitute it; the

parts which immediately precede or follow any particular passage or ‘text’ and determine its

meaning” (Oxford University 2006). This is very much like what I have been calling “cotext”, after

the fashion of some translation theorists. “Cotext”, as I have been employing the term, concerns

textual matters, of structure, composition, and type. The reason that textuality is so important for TTS

is that, on its lights, translation is sensitive to issues of textual structure, order and composition,

particularly in light of a text-type. “Cotext” refers to such factors involved in determining the

meaning of a text via translation. “Context”, in the philosophical, linguistic and anthropological

literature, in contrast, has come to be used as a synonym of what is ordinarily called a

“circumstance”.132 The idea of text as TTS construes it—structured documents with beginnings, ends,

and authors, and types—is completely foreign to context as it is discussed in the recent literature.

Here, what is thought to be salient under the heading of “context” are factors such as time, place,

speakers and interlocutors themselves, their interests and attitudes (including such controversial

factors as referential intentions), and most importantly, language.

300
 TTS rejects the idea that the circumstance of language utterance renders meaning

determinate. One reason that TTS is so critical of Contextualism is that it makes translation not only

mysterious, but determinate translation impossible. All linguistic theories of meaning make

determinate translation impossible, for translation, it was argued, is only determinate when semantic

equivalence between texts is achieved. But according to linguistic theories of meaning, if there is any

such thing as equivalence in translation, it is equivalence between languages. All languages are

different, both in the narrow sense of vocabulary and syntax, and in the expanded, anthropological

sense that connects language to matters of cultural significance. On the linguistic conception of

semantics, the requisite semantic equivalence is not forthcoming. What makes Contextualism worse

than the mere linguistic theory of meaning is that it practically affirms the point that the creation of

semantically equivalent linguistic expressions across languages is impossible. The Contextualist’s

argument can be formulated thus:

(1) Meaning is rendered determinate by contextual factors.

(2) Changes in contextual factors necessarily result in changes in meaning.

(3) The language of an expression is a major feature of the context of an utterance.

(4) Translation involves, among other things, changing the language of an expression or

substituting an SL expression with a TL expression.133

(5) Therefore, translation cannot consist in semantically equivalent expressions or texts

across languages.

I take it that the first premise is merely a statement of the doctrine of Contextualism. The second

premise I take it is a mere corollary of Contextualism’s main thesis. The third premise ought to be

uncontroversial for the Contextualist. Indeed, it is implied by Contextualism. Particularly if (1) and

(2) are true, it follows that any factor that bears upon the determinacy of an utterance must be

contextual, and as the language of an utterance is one such factor, it too must be recognized as a

contextual factor. The Wittgensteinian tradition and its emphasis on the role of language games as

301
forms of life is of course quite comfortable with this affirmation. Writers in the more conservative,

Austinian tradition usually do not explicitly recognize (3) as true, but the entire discourse of recent

Contextualism that dialectically moves between sentence structure and other environmental factors of

an utterance only makes sense if the language of an utterance (its grammar, and vocabulary) is

contextually relevant. If nothing else, the Contextualist must accept (3) if it is possible for one and the

same sentence (holophrastically conceived) to be a token of two distinct languages. For then, the

linguistic difference would result in the types of contextual variation that the Contextualist is

particularly interested in tracking. I think (4)’s characterization of translation is uncontroversial. The

conclusion I believe follows from all the premises. If language is a contextual factor—(3)—and if

changes in contextual factors result in changes in meaning—(2)—then it follows that translation,

which involves changes in language—(4)—results in changes in meaning. If the process of translation

involves altering the meaning of an expression, semantic equivalence is not possible for translation—

(5).

The most important premise for this argument is (2) I believe. While I think this principle is

demanded by the Contextualist’s outlook, I do note that many Contextualists might think this claim

too strong. The Contextualist, particularly of a moderate stripe, might argue that it is not all

contextual factors that are important, but only the salient, relevant, or major contextual features that

determine meaning. Even with this clarification, the argument goes through with the revised version

of (2)—(2′) Changes in major contextual factors results in changes in meaning—for the language of

an utterance is not a minor feature of the context of an utterance. The language of an utterance is the

widest contextual feature of an utterance for to speak of the language of an utterance is a short-hand,

in Contextualist talk, to refer to the relevant linguistic community of an utterance. Even if the

Contextualist were to avoid identifying a language with a relevant linguistic community, it would be

difficult for the Contextualist to deny that the language of an utterance is a contextual factor to be

recognized in determining the meaning of an utterance. For if the language of an utterance were not a

302
contextual feature, it would seem, on the Contextualist’s own paradigm, to have no obvious standing

in affecting the meaning of an utterance.

Relevance theorists, and others who hold that pragmatics and semantics betray systematic

considerations, such as Gricean rules of conversational implicature, are not immune from this

argument as well. For unless they wish to claim that the pragmatics of language use is completely a-

cultural and independent of linguistic tradition, it seems that whatever systematic considerations they

view as relevant to contextual assessment must be contextual. The relevance of language as a cultural,

and hence contextual, factor is particularly salient for a theory that holds that not everything that is

linguistically communicated is linguistically encoded, as relevant theorists do (Wilson and Sperber

1993). Translation will thus force the Relevance Theorist to decide between linguistic form and

relevance in translation (if it is even possible to untangle the idea of linguistic form from the history

of a particular language) and, either way, the resulting translation will not be semantically equivalent

with the ST on the Relevance Theorist’s own conception of meaning as a combination of linguistic

form and relevant pragmatic implicature.

VII.1.1 Contextualism and Semantic Indeterminacy

There are other ways to appreciate the problems of Contextualism particularly relative to

translation. To appreciate these problems, we start with the recognition that Contextualism does not

recognize texts as a primary semantic phenomenon. Rather, on its lights, the paradigm semantic

phenomenon is the utterance, or speech act. If translation is to work on anything, it is the utterance,

replacing an utterance in one language with one from another. But this is an untenable conception of

translation.

In common parlance, “translation” is sometimes used to refer to the substitution of one

utterance for an utterance in another language in a conversational context. For instance, this type of

“translation” is to be found when we are teaching people a new language, or when one relies upon the

303
bilingual expertise of someone to bridge a linguistic gap. But calling this type of activity “translation”

is not proper. It is really a type of interpretive activity, sometimes called “interpreting” or

“simultaneous interpreting”.134 The reason that professionals who specialize in linguistic conversions

of spoken language are regarded as interpreters and not translators is two-fold. First, translation is

only possible with respect to texts that have a structure that can only be assessed relationally. In other

words, one can only assess the semantic and translational significance of a passage or sentence within

the structure of a text. However, in spoken discourse, one lacks the requisite type of relational co-

textualization. Spoken language is characterized by serial utterances that give no determinate cues as

to the boundaries of the text being uttered, and thus no way to assess where a particular utterance falls

within the structure of the text being uttered. Secondly, the job of the simultaneous interpreter is to

provide enough information to the TL audience for them to be able to conversationally respond or

react to the utterances of a live SL speaker. (In the case of linguistic instruction in a second language,

what one wishes to impart is knowledge of what pragmatic ends a sentence is used for.) The

simultaneous interpreter’s job is thus best done when she is able to explain what an SL speaker is

saying. (The teacher of the conversational version of a language succeeds when she is able to explain

the social significance of an utterance.) Explanation is the proper province of interpretation. What

satisfies us as an explanation does not satisfy the demands of translational equivalence. Translational

equivalence, in contrast, only makes sense against the back drop of the possibility of semantic

equivalence (see chapter 3). Semantic equivalence in turn depends upon the very possibility of

determinacy in semantics.

The Contextualist might argue that these arguments against their utterance-based conception

of translation beg the question for it presumes a textual account of translation. In response, I would

claim that while the first objection relies rather obviously on a textual account of translation, I am not

deferring specifically to my account of translation but to industry standards. Thus, if Contextualism

rejects a textual approach to translation, it must contend with the fact that many who are trained in the

304
art of simultaneous interpretation and translation find the distinction relevant, and Contextualists must

explain to them why this institutional distinction is incorrect. If the arguments I’ve presented in

chapters 2 and 3 are correct, the Contextualist chooses a loosing battle if she chooses to ignore the

textual aspect of translation.

The second objection provides an explanation for the first: what we require in the linguistic

conversion of utterances are standards appropriate to conversational interaction. These are not the

same as standards appropriate to semantic determinacy. Ironically, I think it is the writing of the later

Wittgenstein that demonstrates this point.

In his numerous passages in the Philosophical Investigations that attempt to deal with the

problem of rule-following and linguistic interaction and performance, Wittgenstein chases the

illusiveness of determinacy in the conversational setting. If linguistic interaction in context is the

locus of semantics and determinacy, the only way to make sense of the possibility of semantic

determinacy is in terms of rule-following (whether the rule be one that governs the referential

function of words, or their wider function in discourse). However, it is always possible for us to doubt

whether a linguistic practice accords with any rule that we wish to propose, and moreover many

different rules are capable of explaining the course of past linguistic practice. The conversational data

thus underdetermines the rule, in all cases. This poses a practical problem: how does past practice

determine future practice? The answer to this question will depend upon what account of the practice

participants accept. But as there is no common authority to adjudicate semantic controversies in

linguistic contexts of conversation, participants are left with no decisive or authoritative means of

resolving the disputes about what account of the practice must be followed. (Wittgenstein writes: “no

course of action could be determined by a rule, because every course of action can be made out to

accord with the rule. The answer was: if everything can be made out to accord with the rule, then it

can also be made out to conflict with it. And so there would be neither accord nor conflict here”—

Philosophical Investigations I §201). Wittgenstein appears to recognize that the normative question of

305
what a word means in linguistic interaction, or questions of what the practice demands of us, are

philosophical nightmares that can in no principled way be resolved. Rather, what participants in a

linguistic context must reject is such semantic scruples that yearns for determinacy and instead

continue the practice by attempting to make conversation felicitous. (“…there is a way of grasping a

rule which is not an interpretation, but which is exhibited in what we call ‘obeying the rule’ and

‘going against it’ in actual cases” — Philosophical Investigation § 201). This will ensure that people

actually obey the rules of a practice—whatever set of rules end up describing the course of the

practice in the long run. This aspect of Wittgenstein’s thought appears to me to be very salutary and

insightful if understood as an account of the possibilities for determinacy of sense in linguistic

contexts.

What obscures this Wittgensteinien lesson is that Wittgenstein himself did not completely let

go of the ideal of semantic determinacy in conversation. In the Investigations, he repeatedly

emphasized rules of “grammar” that would point to some type of normative logic of concepts that

could settle philosophical disagreements over the use of such concepts—despite apparently

recognizing that what he calls “grammar” has no normative force.135 By On Certainty, Wittgenstein

seems to have realized that the appeal to grammar is quite inconsistent with his insights about the

pragmatics of conversational settings,136 but at an earlier point he appears to have appealed to it as a

means to respond to the demands of earlier, Russellian semantics.137 However, even in his

formulation of the seeming problems of rule-following in practice, Wittgenstein couches his solution

in normative terms as though giving up on the semantic question of the meaning of expressions or the

rules of a practice was in a very real way to obey rules, such as the rules of logical grammar. But in

reality, what I think Wittgenstein describes as “obeying a rule” is not obeying rules at all (at least

none known to be authoritative to the participants) but something more like creating anthropological

rules that describe our practice via a dialectical process of linguistic performance and interaction. On

the view I am advocating of what we can learn from Wittgenstein, he does not endorse the so-called

306
“sceptical solution” that Saul Kripke finds in Wittgenstein, according to which the scepticism about

rule-following is granted while holding on to the ordinary belief that linguistic usage in context is rule

bound (Kripke 1982, 66-67). Rather, on this reading, the lesson we should learn from Wittgenstein is

that the sceptic is correct about the possibilities of determinacy of sense in linguistic, conversational

settings, and moreover that what we really require in conversational settings is not determinacy of

meaning, but pragmatic felicity. The problem of the determinacy of meaning in context is thus not

solved, but dissolved. 138

A more moderate face of the later Wittgenstein strikes a compromise between the two

extremes, which on the one hand rejects the possibility of semantic determinacy in linguistic context

and on the other tenaciously holds on to the idea of “grammar.” This is the face of Wittgenstein who

allows questions about the meaning of expressions within contexts, but accepts only interpretive

(explanatory) answers as responses—responses that are taken to be constitutive of the ebb and flow of

the practice itself (for instance, when he writes that the “meaning of a word is what is explained by

the explanation of the meaning”—Philosophical Investigations I §560). This is a classic move in the

strategy of Contextualism. In response to the question of semantics, which is literally a request to be

shown the meaning of an expression, what one receives in its place is an interpretation, i.e., an

explanation. However, the admitted interpretations each lack privilege: they may be put to good

rhetorical use to sway the direction of practice, but none are definitive. Indeterminacy is a residual

feature of such interpretive accounts.

In summary, Contextualism in taking the conversational or speech act setting as the paradigm

for semantic analysis is quite unable to account for determinacy in semantics, which is a requisite of

an account of translation. At best, semantic determinacy within a context consists in fidelity to rules,

but it is precisely within a conversational setting that the question of what rule is relevant to the

discourse is impossible to settle determinately. If semantics is indeterminate, there can be no right or

wrong in translation.

307
 Had the later Wittgenstein directed his penetrating insights to the topic of translation, the

history of the philosophy of language may have been quite different. Wittgenstein would not have

been so influential in inspiring the tradition of Contextualism, and it may never have acquired the

attention it receives now. The later Wittgenstein’s reflections on translation are few and quite

superficial. In the Blue and Brown Books he describes the meaning of a word as its role in a practice,

and translation as a means of capturing the functioning of an SL in a TL (Wittgenstein 1960 [1958],

95). This is a view remarkably similar to the one advocated by British Functionalists, which was

roundly criticized in chapter 2 for its unacceptable implications for translation. (It was noted there

that Wittgenstein in his unpublished notes criticized Russell for holding a Functionalist view, but

oddly, not his earlier self.) By the Philosophical Investigations, Wittgenstein had abandoned the

practice-role, functionalist account of meaning because of his insights about the open-ended and

contestable nature of linguistic practice. As for translation, he suggests at one point that translation of

speech can be judged as successful if it allows the onlooker to predict the behaviour of her subjects

(Philosophical Investigations I §243). This was certainly an improvement on the earlier, quasi-

functionalist analysis, but the suggestion stops a few steps short of the problem that Quine uncovers

in his radical translation thought experiment. Quine too worked with this model of translation and

reasoned that linguistic behaviour could not uniquely determine translation and that outright

contradictory “translations” could be licensed by the data. Wittgenstein recognized this type of

problem domestically with language users’ assessment of the significance of the utterances of their

interlocutors. And if my account of Wittgenstein is correct, he dissolved this problem to the extent

that he recognized that semantic determinacy was not necessary for linguistic interaction, but merely

pragmatic felicity, and thus participants in a linguistic practice were free to negotiate among each

other the course of their practice. However, in the case of “translation,” the people whose utterances

the Contextualist wishes to “translate” are not participants in one’s language game: they inhabit a

distinct contextual frame by virtue of their participation in a different language game. Perhaps

308
recognizing this shortcoming of the Wittgensteinian paradigm while not wishing to give it up, Lance

and O’Leary-Hawthorne argue that translation is “ in an important way, attempting to form a single

community where previously there were two” (Lance and O'Leary-Hawthorne 1997, 20). While this

is a possibly admirable regulative ideal in translation, in reality translation does not necessarily

produce one community where there were two. If it did form one community, it may be possible for

the speakers of the different languages to engage in a joint, pragmatic negotiation of the semantics of

their utterances. This would fall considerably short of the goal of semantic determinacy, but it would

still amount to some form of accommodation of what is of foreign pedigree. However, the reality is

that translation is not the literal attempt to forge joint communities, but to make semantic artefacts

composed in one context accessible in another. To forge one linguistic community out of two in

practice is to form a pidgin, not to translate.

At this point, the only option left to the Contextualist is to argue that translation can be

determined according to the convenience of the TL speakers. But this is not that different from the

Functionalist view that was criticized in chapter 2. To constrain translation purely for the convenience
139
of TL speakers is to give up on translation all together. We might as well compose a fictitious

novel about an amusing gas station attendant and call it Aristotle’s De Anima.

VII.1.2 Semantic Determinacy and the role of Institutions in TTS

The Contextualist is likely to object that by criticizing Contextualism for failing to produce

semantic determinacy, i.e., rule-following, I have set the bar so high that no possible semantic

account could explain semantic determinacy. How can any account of meaning avoid the types of

paradoxes and problems of rule-following that plague the Wittgensteinian account of language?

The answer to this legitimate query consists of two parts. The first part consists of an

explanation of how determinacy is understood on TTS’s account. The second is an explanation of

how TTS avoids some of the rule-following paradoxes that Wittgenstein locates in language games.

309
 VII.1.3 How TTS Avoids the Problems of Contextualism Part 1—Meaning
is Textual

First, we must affirm the difficulties that beset Contextualist accounts of meaning as rule-

following within a context and to argue that determinacy in meaning is not to be found as a feature of

activity within a context. Rather, determinacy in meaning is a function of a relationship of texts

across contexts that are made equivalent by the relevant experts in their roles as translators and

teachers. While this activity happens within a context (as does any activity) the normative and text-

type-theoretic implications of the activity of text-type experts transcend the narrow context that they

inhabit because the texts that they render equivalent via the appropriate text-type-theoretic translation

are portable. As genuine, determinate translations of texts, on this account, are semantically

equivalent, a TT and a ST stand to each other as each other’s definite meaning, and moreover these

meanings are not context-bound any more than texts are rooted to the ground. Determinate meaning,

on TTS’s account, literally transcends contexts. In claiming that meaning transcends context, TTS is

not committed to the absurd view that texts float above the world. Every real text has a definite locale

in time and space that is its context. But such contingent contexts have no direct bearing on the

meaning of the text, for the meaning of the text, and that it is a meaning of another text, is a relation it

shares with other texts that are in different contexts. This is naturally mediated by a text-type

institution, whose officers also inhabit definite places in time and space, but it would be a violation of

the idea of context to claim that an abandoned translation of Plato’s Republic in the Antarctic

(determinately meaningful as it is according to TTS) is in the context of philosophy or the text-type

institution of philosophy.

While utterances and speech acts are inextricably context-bound, texts are not. Texts are

portable, and retain a determinate sense (by virtue of their subsumption under a text-type by a text-

type institution) regardless of the context they are in. In other words, texts are in a different category

of semantic phenomena from utterances and speech acts. Texts can be about speech acts and

310
utterances, but they are not the same thing as speech acts and utterances. Thus, for instance, a text

about marriage vows, or one that records the script of a marriage ceremony, marries no one, whereas

the felicitous recitation of the text does.140 The former is textual and semantic because its significance

is not tied to any particular context. The latter, namely the speech act of actually marrying someone,

is not semantic but rather pragmatic, because it has its dignity within certain concrete circumstances.

Austin’s revolutionary How to Do Things with Words thus deserves great respect for pointing out the

oft-neglected aspect of language use in public, namely that it counts as an instance of social action.

However, it is a great strain to countenance his account of locutionary acts (language use understood

in terms of conventional meaning), illocutionary acts (language use understood in terms of intended

effects in context) and perlocutionary acts (language use understood in terms of actual effect in

context) as constituting a theory of semantics. The account takes for granted a background regularity

(and even institutional regulation, as in the case of marriages) that makes such context-bound

phenomena possible. Accounting for this regularity satisfactorily will, in my opinion, take us away

from the contextual and towards the institutional, which cannot be understood under the paradigm of

the context for institutions span contexts.

VII.1.4 How TTS Avoids the Problems of Contextualism Part 2—Social

Practice that Allows for Determinacy is Institutional, not Contextual

Secondly, to understand how TTS avoids some of the problems of Wittgenstein’s rule-

following paradox, we must first begin by drawing a distinction between linguistic interaction and

negotiation in context, and the activity of experts in a text-type institution. The former corresponds to

Wittgenstein’s “language game.” The latter does not. The major difference is that text-type

institutions are institutions, while context-bound, linguistic interactions are not. Both are social

practices, but they have very different normative force and reach.

311
 There is a looser and a stricter use of the notion of “institution.” Sometimes people call any

social practice that goes on for a long time an “institution,” while those that are shorter lived are mere

social practices. I have a more robust distinction in mind that builds upon this. One feature that

distinguishes an institution, like the Catholic Church, or a legal institution, is that qua institution, it is

controlled and regulated by practical authorities, whose decisions result in real consequences. This

contrasts with the idea of expert authority.141 We obey a mechanic of our choosing, for instance, in all

things automotive not because she has power over our cars (she may, but that’s not why we listen to

her) but because she is knowledgeable. We obey a judge, in contrast, not because she is

knowledgeable, but because she has power over us. This is practical authority, whereas the mechanic

only has expert authority. Institutions are characterized by officers with practical authority.

Academic institutions are institutions, even though they are populated by experts, because the

experts have practical power. They have power, for instance, to grant high grades or low grades, to

decide on which texts get published and which do not, and how texts are to be classified. Text-type

institutions, which span and overlap academic institutions in our culture, are similarly comprised of

experts with practical authority. In a mere social practice or a mere language game, there are no

practical authorities. Rather, all the participants must attempt to negotiate their way with their bare

wits.142

Another feature of an institution, lacking in a mere social practice or language game, is that

the practical authority of an institution transcends the boundaries of the context of the institution. An

undergraduate student studying philosophy is not in any clear way part of the text-type institution of

philosophy. Yet, officers of the institution (professors of philosophy, and probationary officers like

graduate students) can yet enter evaluations for the undergraduate student that have effects on the

student’s life that are at no point part of the text-type institution of philosophy. In a language game, or

the paradigms of contexts that Contextualists consider as key to determining the meaning of

utterances, there is no normative reach of negotiations within a context to outside of the context. (If

312
there were, the context of an utterance would not be the determinative factor in its meaning.)

However, the decisions of text-type institutions do have implications for what occurs far beyond the

narrow walls of philosophy conferences, and offices of professors, if nothing else through collective

decisions about how to translate and read certain texts, and which texts are to be promoted as

important and which are not. Thus, for instance, the decisions of the Catholic Church have normative

implications for who counts as a Catholic anywhere in the world (i.e., regardless of context).

Likewise, the question of the semantic content of a text, determinately translatable by the text-type

institution of philosophy, will be determined by the operations of this institution, regardless of where

the text is in the world.

Finally, an extremely important difference between institutions and context-bound social

practices is that institutions can have clearly articulated ideals and norms that govern the institution.

The challenge in the case of institutions is determining when they run afoul of their ideals, and when

they are acting in conformity to their ideals. However, this is a narrower problem from the one that

Wittgenstein recognizes in context-bound, social practices. There the problem is the futility of coming

up with an “interpretation” that gives us an account of the very mission of the social practice. The

reasons that institutions can manage the clear articulation of normative ideals while context-bound

social practices have problems are two-fold. First, institutions have practical authorities that are able

to decisively resolve controversies as to practical objectives on a case by case basis, while context-

bound social practices do not (if such language games did have such practical authorities, they too

would be institutions). Ideally, such authorities resolve controversies by laying down new rules to

settle the controversy in a manner keeping with past practice and often the explicitly stated ideals of

the institution. Not every institution maximizes on this potential power. For instance, text-type

institutions typically run on a consensus model, with informal hierarchies of expertise exerting some,

but slow, influence on the institution as a whole. Thus, a critical amount of institution-wide

agreement is necessary before strong changes are felt outside of the institution. Yet, text-type

313
institutions have achieved such critical consensus time and again. This happened, for instance, when

the natural and mathematical sciences were shed from philosophy, and when, later, psychology was

largely removed from philosophy. These episodes occasioned a decisive refinement of the text-type

of philosophy. Secondly, because institutions can divide labour via specialized officers, a sense of

order and purpose can be concretely felt in the institution in a way that context-bound, social practices

typically cannot realize.

QI, as I understand it, articulates the normative ideals of the text-type of philosophy, today.

And while the label of “QI” is idiosyncratic, the doctrine that it attempts to systematically elaborate

can be heard in most any introductory philosophy class, almost anywhere in the world. If introductory

students learn anything about philosophy, they learn that there are terms like “real”, “ethical,”

“knowledge” and “reason” that philosophers have theories about (under fancier titles like

metaphysics, moral philosophy, epistemology and logic), that they attempt to argue for these theories

by trying to make cases for what they think is important to buttress their views. They learn, moreover,

that this argument has been going on for a long time, and it doesn’t look like it is going to end any

time soon, and yet philosophers think that philosophy isn’t a waste of time—rather, they seem to

think that there are right and wrong answers to these theoretical questions. Later students might come

to learn of philosophers who paradoxically criticize the whole project of philosophy, such as

N g rjuna, Wittgenstein and Heidegger, but that these philosophers too seem to have theories

articulated with words like “meaning”, “understanding,” “truth” or “being” and these philosophers

seem to think that there are right answers concerning these theories, despite (or perhaps, because of)

disagreement.

VII.1.5 Objections from Contextualism

The Contextualist might still object and claim that there is no difference between TTS and

Contextualism, for the decisions of practical authorities can still be disputed on TTS’s account.

314
Moreover, some of the Wittgensteinienesque worries about rule-following can be recreated for

institutions. We can worry, for instance, whether current institutional practice accords with the past.

The response begins by noting that this worry is considerably weaker than the

Wittgensteinian worry about social practices as uninterpretable. In the case of institutions, reaffirmed

practical actions in support of a precedent have a special normative force in disambiguating the

institutional direction and practice for which there is no analogue in the case of a mere social practice

or language game. This is simply a function of the fact that in an institution there are practical

authorities whose opinion matters more than others on a practical level and moreover this importance

has reverberations outside the practice whereas in a language game there is neither such authority nor

any such determinate, extra-contextual effect. In the case of the language game, history has no

determinate normative force, whereas in an institution it does.

Moreover, it is important to note that in the case of institutions, disputes as to whether recent

practice conforms to institutional ideals seems to have a determinate answer, whereas in the case of

the language game there is no analogue for there is no way to determine what the ideals of the mere

social practice are. In the case of institutions, past decisions of practical authorities count, and the

more they are reaffirmed in history, the more important they are. Nothing of the sort can be said of a

mere language game for there is no distinctly disambiguating action in a language game like practical

authority.

Thus, a paradoxical feature of the investment of power in the hands of practical authorities in

institutions is that it opens the door to objective criticism of institutional decisions. As institutional

decisions that disambiguate the course of the institution occur over time, it is always possible to

criticize recent and past decisions in light of a dominant trend in institutional self-understanding made

apparent by the decisions of its authorities. Given the historical nature of institutions, it is even

possible, at any given time, that practical authorities may be acting in widespread departure from the

dominant trend in the institution, and that these departures may be criticized as institutional failings.

315
In other words, objective criticism of institutional actions is possible in the case of the social practice

of institutions. However, in the case of the context-bound language game, where every action can be

seen as both in conformity and against the course of the practice, there seems to be little room for

such objective criticism. (Consider the difference between a controversy over what, independently of

recent practice, Canadian etiquette demands, and a controversy over what, independently of recent

court decisions, Canadian law demands. The former seems to have no objective answer, whereas

objectivity seems wholly possible in the latter case and not foolish either. Normative arguments

concerning Canadian law that track the trajectory of Canadian legal institutional decisions in the long

run will be vindicated. Arguments concerning Canadian etiquette, in contrast, seem to never be

judgable or successful, for it will always be possible for participants to argue that a long-standing

practice is a departure from the true ideals of the practice. The difference, once again, comes down to

the fact that no action in the case of a mere language game or context-bound social interaction has a

definitive force, whereas the decisions of authorities in institutions do.) 143

Many Contextualists might find that the Wittgensteinian focus of my account of

Contextualism is misleading. They might argue that some Contextualists, such as Relevance

Theorists, regard context as providing important details of semantic assessment, but apart from

context something like basic rules of a language, provide us with an a-contextual framework to assess

the significance of context. On this account, Contextualism is something more like Stanley’s Subject

Sensitive Invariantism. My response to this defence is two-fold. First, if context plays the role of

rendering meaning as such determinate, all the problems that I’ve identified for contexts via

Wittgensteinien arguments stick, for his considerations show, I believe, that context has no obvious or

decisive normative force. The Wittgensteinien arguments about rule-following do not obviously

criticize a view of semantics that holds that context is important by virtue of noncontextual factors. I

think this view, call it Context Relevance, is uncontroversial. Even Cappelen and Lepore who argue

for a so-called “Insensitive Semantics” recognize the relevance of context to understanding the

316
meaning of some expressions. Secondly, if a theory, such as Relevance Theory, holds that there is a

connection between linguistic form and pragmatic implicature that is systematic, the theory must hold

that such a connection is culturally and linguistically contingent. In other words, if Relevance Theory

is to be plausible, it must recognize that what speakers regard as pragmatically relevant to linguistic

utterances will vary according to linguistic community. However, if Relevance Theory is plausible in

this way and not implausible (in holding that there is no cultural or specifically contextual factors

affecting the connection between linguistic encoding and pragmatic relevance) then all the worries

that Wittgenstein brings up about rule-following in context attach to Relevance Theory. For it is

characteristic of such a non-institutional, social practice that participants in this practice can contest

specific relationships between linguistic encoding and pragmatic implication, and there appears to be

no decisive manner to settle the controversy. The question of what is the right linguistic-pragmatic

dyad is unanswerable. Speech acts can be judged for their felicity, but to acquiesce into pragmatic

felicity is to give up on the idea of semantic determinacy.

VII.1.6 What we can learn from Contextualism

In closing, I wish to state only that Contextualists teach us a lot about language use, and the

problems associated with it. Their observations about language are often quite insightful and not

worthy of outright rejection. What TTS does reject is the notion that such observations tell us

anything determinate about meaning. There is of course some connection between linguistic encoding

and pragmatic implication, and there may even be a long-term regularity in a culture as to

expectations about pragmatic implication. These may or may not be processed systematically. If they

are processed without systematic considerations, as Recanati suggests (2004, 38-50), context is

central to helping people navigate their linguistic world. But TTS resists identifying any of this with

meaning, or at least, determinate meaning. If meaning is simply context-bound, it is literally nothing

at all to speak of from outside, and even from inside the context there is no determinate account about

317
meaning to be given. However, if meaning transcends context, then there is an objective of

determinacy that we can aim for in an account of semantics. Aiming for determinacy in our account

of semantics is not simply a theoretical luxury. It is rather demanded by an acceptable account of

translation. And translation is no mere theoretical luxury. It is demanded by the fact that we live in a

small and crowded world, and must understand the semantic artefacts of those whom we do not share

contexts with. If meaning is context-bound, then there is no hope of us ever understanding anyone

outside of our context. Moreover, if it is context-bound, it seems that our “understanding” of each

other is nothing more than our felicitous social interaction, which falls considerably short of our

epistemic ideals of understanding—particularly where values are at stake.

VII.2. Insensitive Semantics

Recently, Herman Cappelen and Ernest Lepore have argued that Contextualism makes

communication across contexts impossible for meaning on its lights would change as contexts

change. Moreover, they charge that Contextualism is guilty of a type of inconsistency when it

speculates that meanings of sentences change according to contexts for such a generalization only

makes sense if the Contextualist discounts their own context (Cappelen and Lepore 2005, 10). In

contrast, Cappelen and Lepore advocate what they call Semantic Minimalism, according to which the

semantic content of a sentence is the minimal content that all utterances of the sentence share

(Cappelen and Lepore 2005, 143-144). Thus, on this view, “Rudolph has a red nose” has the minimal

content of Rudolph has a red nose.

Any sentence, on their account, can be used in diverse manners to imply a plurality of

propositions through speech acts (a position they call “Speech Act Pluralism”) but the purpose of

semantics is to provide the bare minimum knowledge that a speaker needs in order to employ

sentences to their varied pragmatic ends. Meaning, on this account, is abstractly or theoretically

linguistic, and context-insensitive. In short, the bulk of what the Contextualist regards as the meaning

of a sentence, Semantic Minimalism places in the category of pragmatics.

318
 Cappelen and Lepore distinguish a family of semantic views from what they call Radical

Contextualism. These views are species of “Moderate Contextualism.” Moderate Contextualism,

according to Cappelen and Lepore, regards meaning as generally not context-sensitive, but attempts

to expand a basic list of context-sensitive expressions identified by David Kaplan, to include many

concepts not recognized by Kaplan as context-sensitive.144 Moderate Contextualists contribute to

context-sensitivity bloat either through semantic opportunism or through being misguided. The

misguided semanticist is thoroughly interested in the semantic project, but is unable to find a way to

accommodate the meaning of a certain term except by expanding the basic list. This, Cappelen and

Lepore hold, is misguided.

VII.2.1 TTS and Insensitive Semantics

TTS and Insensitive Semantics, as Cappelen and Lepore advance it, have much in common.

For instance, Cappelen and Lepore are critical of a view they call the Mistaken Assumption: “A

theory of semantic content is adequate just in case it accounts for all or most of the intuitions speakers

have about speech act content, i.e., intuitions about what speakers say, assert, claim, and state by

uttering sentences” (Cappelen and Lepore 2005, 53). TTS too is critical of this assumption, for on

TTS’s account, texts, and not speech acts, are the proper focus of the semantic enterprise, for speech

acts are semantically indeterminate, which is to say that an account of a speech act’s success or

failure tells us nothing about how the semantic content of the speech act is to be preserved in

translation. Questions of translation, according to TTS, must take into consideration matters of

textuality, which are of a whole other category from speech acts.

TTS can also partially agree with Cappelen and Lepore’s statement of the task of semantics:

Semanticists disagree on what the central semantic features are (truth conditions, intentions,
extensions, propositions, functions from worlds to truth values, prototypes, stereotypes, situations, or
whatever), but they do tend to agree that semantics is a discipline that aims to characterize

319
 systematically certain features of linguistic expressions and to do so in a way that captures general
truths about languages, and not just truths about particular speakers in specific contexts. (Cappelen
and Lepore 2005, 58)

Our previous excursions into the semantics of the philosophical text-type should make us sensitive to

the fact that “semantics” and “meaning” are terms that are up for grabs, and the scope of these terms

are not a priori obvious to all concerned. However, to the extent that theorists are concerned with

providing a theory of meaning for languages, TTS is committed to the view that such a theory ought

to capture the most systematic features of language because such features are the most general in

nature. TTS, as I’ve noted, is whole-heartedly for distinguishing determinate meaning as context

transcendent, and thus it has great sympathy with the notion that semantics cannot be concerned with

what speakers mean (in the loose sense of mean) in particular contexts. But TTS disappoints those

who think that meaning is essentially linguistic, for on TTS’s account, very little can be said about the

systemiticity of language and meaning. Cappelen and Lepore write:

There are languages. Languages have words. Words combine into complex expressions and
sentences. The semantic values of words contribute to the semantic values of the complex
expressions and sentences of which they are a part. Semantics is about how best to specify the
semantic value of the lexical items and their contribution to the semantic values of complex
expressions and sentences in which they occur.(Cappelen and Lepore 2005, 57)

Call the view articulated by Cappelen and Lepore in this passage the “Bottom-Up” view of meaning,

according to which smaller linguistic units comprise larger linguistic units, such as texts. According

to TTS, the Bottom-Up view is incorrect for text-types also exert a top-down dynamic in meaning,

whenever text moulding is brought to bear in rendering a text determinate, or frequently when a text

is subsumed under an institutional text-type. The institutionalisation of texts introduces a novel

dynamic in semantics that both creates the possibility of determinacy in semantics, but also ruins the

dream usually associated with the doctrine of “compositionality.” As text-types are essential

ingredients to be factored into semantics, very little can be said about the general semantic features of

320
a language. We do not miss this lack of systemiticity, because linguistic, conversational

communication is primarily a pragmatic phenomenon, not a determinate, semantic phenomenon. In

passing, it is important to note that TTS is not averse to compositionality, but on its account, one

cannot understand how a text is composed independently of a text-type, and thus the Bottom-Up

approach is only one version of compositionality. The subsumption of a text under a text-type does

not thereby impose a meaning on the text, but it constrains the functioning of the language in the text

in such a way that one cannot understand its meaning purely in terms of the grammar or vocabulary

of a language. This is confirmed by the fact that so-called “linguistic competence” is insufficient to

understand and read texts. Linguistic competence will help us buy ice at the corner store: it is not

enough to help us read philosophy, literature or science.

Given that the argument I am putting forward is not wholly in step with Cappelen and

Lepore’s thinking on semantics, it should not be surprising that TTS and their Insensitive Semantics

are at odds, particularly over the issue of context-sensitivity. Yet, there ought to be some agreement

between our views as the motivations for my TTS and Cappelen and Lepore’s Insensitive Semantics

overlap. And there are. For instance, Cappelen and Lepore wish to criticize indexical views of

philosophical terms put forward by authors such as Dreier and Cohen because such accounts imply

speaker relativism, or the view that judgments using concepts such as RIGHT and KNOW are assessed

as true or false relative to the standards indexed to the speaker. It is no interest of TTS to defend

speaker relativism as a feature of the semantics of philosophical concepts, for the semantics of

philosophical concepts must help us translate all philosophical texts, not only those that are

committed to speaker relativism. Thus, to this extent, I agree with Cappelen and Lepore’s concerns

with Dreier’s and Cohen’s accounts. However, Cappelen and Lepore believe RIGHT and KNOW are not

context-sensitive. Why? Is rejecting Speaker Relativism sufficient to reject the context-sensitivity of

these concepts? I have argued that it is not. The reason that concepts such as RIGHT and KNOW are

context-sensitive is that the text-type of philosophy demands that we be able to pry apart the question

321
of what these concepts designate when employed, from the question of whether a philosopher is

correct to designate as they do via these concepts.

One reason that the question of reference can be pried apart from the question of the propriety

of judgments expressing such reference is that reference can always be direct. For reference to be

direct is for it to be unmediated by the semantics of the judgment that expresses the reference. I can

use the word “cat”, without wishing to alter its feline meaning, in a literal manner, to refer to my

computer, while mistaking the computer for a cat. My judgment “this is a cat” (referring to my

computer) would thus be false. The reason that we take this to be false is that it is possible for us to

assess the propriety of the judgment on the basis of the constituents of the judgment, in comparison to

the fact of what I was referring to with my use of the word “cat”. By analogy, I could mistakenly

refer to something I recognize as evil and is (for the sake of argument) evil with the term “good” and

be analogously mistaken. I could, for instance, judge incorrectly that “butchery is good”, not for any

deep philosophical reason, but because I was nervous and thought I was using the concept EVIL but

instead got jumbled and employed the concept GOOD. This is a rather trivial way in which we can pry

apart the question of the reference of terms, from the propriety of judgments that express such

reference. It rests upon mental and linguistic error, and not anything deep about philosophical

concepts.

In the case of philosophical texts, and their translation, there is a more profound way in which

we can and must pry apart the context-sensitivity of philosophical concepts from the propriety of the

judgments that they express. In this case, it is possible to pry these questions apart because what I

have identified as the objectivity clause that is part of the character of a key philosophical term can be

read in more than one way. This is the clause that specifies that items are designated, informed by

theories chosen by philosophers for axiological reasons, according to the relevant considerations that

are thought to be objectively persuasive From one perspective, the clause can be read as referring to

the contextual factors that philosophers bring to philosophical argumentation, such as their choice of

322
theory. Call this the narrow or internal reading of the objectivity clause. The narrow reading is one

part of what is necessary to produce a faithful translation of a philosophical text, for a faithful

translation of a philosophical text must preserve the theoretical commitments of the author of the

philosophical ST. However, if this were the only constraint on translating philosophy, we would be

committed to translating all philosophers as making true claims simply because they defer to their

own theoretical commitments by virtue of it being theirs. If this philosopher relativistic, narrow

reading of relevance were the only way of reading the text-type feature of philosophy, there would be

no room for serious debate or argument in philosophy. Entire sections of philosophical texts would

become nothing but semantic noise that could be left out of translation for it would not be tracked by

the semantic aspects of the text-type feature of philosophy as relevant. However, this would do

violence to the project of translating philosophy. Thus, it is also important to appreciate that the

objectivity clause can be read, and must be read, as also deferring to objective standards of

relevance—whatever they may be. Call this the wide or external reading of the objectivity clause.

Even the subjectivist, who wishes to make a case for ethical subjectivism, or a more general

metaphilosophical subjectivism, or the relativist who wishes to argue that certain judgments are true

relative to certain contexts, wants to convince other philosophers of the propriety of such views on the

basis of objective considerations, and not simply the ones they choose because they choose them. To

recognize that the objectivity clause has objectivist implications is simply to respect the character of

the philosophical text-type as devoted to the debate and argument of philosophical issues, which can

only be possible if the text-type presumes that there is something to be objectively right or wrong

about in the realm of argument. (How we can determine what these theory independent objective

considerations are is a matter that I shall deal with in the final chapter.)

There is thus a seemingly paradoxical quality to philosophical arguments and texts that we

ignore at our peril. If we ignore the relativistic, indexical aspects of philosophical argumentation, we

close ourselves off from the real presence of theoretical diversity and pluralism in the history of

323
philosophy. If we ignore the objectivist aspects of philosophical argumentation, we close ourselves

off to the very possibility of genuine philosophical debate. This aspect of philosophical argumentation

was rather imperfectly grasped by Jean Paul Sartre in the realm of ethics, where he observed that

while moral decisions must be made by deferring to existentially felt considerations, one is all the

same choosing for everyone when one chooses a position in ethics (Sartre 1975 [1964]). Neither the

indexical nor the objectivist aspects of philosophical texts can be ignored. Rather, we must find a way

of accommodating both in our account of philosophical semantics, in light of TTS, for the mandate of

such an account must be to help us translate philosophical texts. Traditional linguistic theorists of

meaning can ignore this duality of philosophical concepts for they are not motivated, nor do they have

the theoretical resources, to explain how we could determinately translate philosophical texts. The

traditional semanticist’s project is to either vindicate one substantive philosophical position at the

expense of others (explaining all intelligible divergence from the ideal philosophical view by means

of direct referential error)—e.g., New Non-Analytic Naturalism—or to affirm theoretical or

normative diversity in philosophy at the expense of profound philosophical debate—e.g.,

Expressivism or Indexical, Speaker Relativism. Our project cannot be so simplistic, but neither can

the traditional semanticist’s project, if translation is a test of meaning. Hence, the importance of the

narrow and wide readings of the objectivity clause to QI.

To the extent that QI accepts the objectivist, context-independent nature of philosophical

texts, philosophical concepts on its account will fail tests that are designed to track context-

sensitivity. To the extent that it represents the indexical, context and cotext variant aspect of

philosophical texts, philosophical concepts on its account will pass some tests.

VII.2.2 TTS and Cappelen and Lepore’s Tests

A test that philosophical concepts pass, as well as uncontroversial indexicals, is the alleged

test of necessary truths. According to David Kaplan, a feature of indexicals is that they can be

324
employed to make necessarily true statements, such as “I am here.” These, Cappelen and Lepore

assert, are the types of sentence of which no utterance can be false (Cappelen and Lepore 2005, 78-9).

Cappelen and Lepore do not put philosophical concepts to this test, but clearly analogous “necessary

truths” can be formed, such as “evil is worse than good,” “good is better than evil,” “the moral is the

ethical,” “an illusory object is not real,” etc. The explanation for these supposed necessary truths

comes down to the coincidence of contextual factors that the characters of these concepts pick out.

This is the supposed epistemological aspect of indexicals. The syntactic correlate of this epistemic

test is the test of anaphora that presumes that indexicals are able to enter into anaphoric relationships,

such that subsequent context-sensitive sentences can piggy-back on the reference of the earlier

context-sensitive expression. Thus, one can utter, “That’s a table but it is not a book” truthfully, just

as one can utter truthfully, “the Good is the Form of the forms, and it is not a mere material object.”

In both cases, the “it” can refer back to the first context-sensitive expression (“that” and “good”)

without absurdity. There are sentences that appear to fail this test. Consider the sentence “Tigers are

mammals, and it is a big domain.” If “tiger” or “mammal” were context-sensitive, we should expect

that “it” could obviously refer to them. However, Cappelen and Lepore believe that it fails to refer to

either “tiger” or “mammal” in this sentence thus showing that neither “tiger” nor “mammal” are

context-sensitive (Cappelen and Lepore 2005, 77). In general, I do not believe that we should get

exited about these tests. In the case of the former test of a priori truths for indexicals, there are

obvious and accepted counterexamples to the alleged necessity of “I am here”, such as “I’m not here

now,” commonly heard on answering machines.145 Yet, the test of anaphora seems to have no

counterexamples that I know of, except the possible counter intuition that “it” can anaphorically refer

to tigers and mammals.

A test for context-sensitivity that philosophical concepts appear to patently fail is what

Cappelen and Lepore identify as the blocking of “intercontextual disquotational indirect reports”.

325
True context-sensitive expressions, according to Cappelen and Lepore, block the straightforward

equivalence between sentences that are quoted in indirect speech, and those that are disquoted:

Take an obviously context-sensitive expression, e.g., take ‘tomorrow.’ Consider an utterance by

Rupert on Tuesday of ‘John will go to Paris tomorrow.’ If someone tries to report on Wednesday
what Rupert said with his utterance on Tuesday with an utterance of ‘Rupert said that John will go to
Paris tomorrow,’ his report is false because the expression ‘tomorrow’ fails to pick out what it picked
out in Rupert’s original utterance. The presence of ‘tomorrow’ in the disquotational report figures
prominently in an explanation of why the report is false (Cappelen and Lepore 2005, 89).

A related test that philosophical concepts appear to fail is the blocking of “collective

descriptions”— the integrated collection of two descriptions that employ context-sensitive verb

phrases:

As an illustration consider the context-sensitive ‘yesterday’: Suppose we know there are two contexts
in which ‘Yesterday John left’ and ‘Yesterday Bill left’ are true respectively (though we don’t know
the times of these contexts). It doesn’t follow that there is a context in which ‘Yesterday John
and Bill left’ is true. (Cappelen and Lepore 2005, 100)

As Cappelen and Lepore show, philosophical concepts do not obviously pass either the test of

intercontextual disquotational blockage, or collective description blockage:

(Know) Any utterance of ‘A knows that he has a hand’ can be reported by ‘She said that A knows
that he has a hand’ and any two such utterances can be reported by ‘They both said that A knows that
he has a hand.’

(Believes) Any utterance of ‘A believes that B is shady’ can be reported by ‘She said that A believes
that B is shady’ and any two such utterances can be reported by ‘They both said that A believes that B
is shady.’

(Moral terms, e.g., bad) Any utterance of ‘Killing penguins is bad’ can be reported by ‘She said that
killing penguins is bad’ and any two such utterances can be reported by ‘They both said that
killing penguins is bad.’ (Cappelen and Lepore 2005, 95)

326
 VII.2.3 Troubles of Cappelen and Lepore’s Project

Cappelen and Lepore’s argument is not formulated with QI in mind, and thus it is difficult to

anticipate what their objection to it might be. QI concedes that philosophical concepts will not

behave as context-sensitive expressions through and through. It also explicitly distances itself from

the speaker relativism that Cappelen and Lepore explicitly criticise, placing it in an odd category from

the perspective of Cappelen and Lepore’s position. However, QI, informed by TTS, can form the

basis for a series of objections to Cappelen and Lepore’s view.

First, we may ask what the relevance is of the apparent tests that Cappelen and Lepore

formulate. Certainly they serve to separate off a narrow list of expressions from others in English. But

why should this have any clear implication for the context-sensitivity of philosophical concepts,

unless philosophical concepts must be forced into the narrow two categories that Cappelen and

Lepore seem to recognize (that is, context-sensitive expressions like indexicals, and everything else)?

Let us assume, for a moment, that Cappelen and Lepore are correct about the behaviour of

expressions in English, and moreover that the tests track features of the English language that

separate out some context-sensitive expressions from other types of expressions. Clearly, the realm of

the philosophical is wider than English, for philosophy constitutes a type of text found in a variety of

languages. In fact, the English language is a late arrival on the scene, compared to philosophy. Why

shouldn’t English (if Cappelen and Lepore do indeed hit upon some deep normative structure of the

language) bend to philosophy and thus accommodate the context-sensitivity of philosophical concepts

as a third category that is quasi-indexical and not simply indexical or absolutely context-insensitive?

Why must philosophy bend to English? Cappelen and Lepore would likely find the question

preposterous. They would likely argue that the semantics of a language set the conditions for what is

possible in the realm of semantics, and thus it is philosophy that must be deferential to the rules of a

language. This, I believe, would be an unfortunate argument.

327
 The idea that the rules of a language dictate its semantics might make sense if translation was

unimportant to semantics. However, particularly in light of the rule-following paradoxes that

characterize social practices, TTS presents us with the only means of objectively fixing the meaning

of semantic phenomena. Quine was correct to note that if we cannot determinately translate, we have

no right to believe that there is anything objective about semantics. What I have argued is that, though

translation cannot produce equivalences of rules or meanings of languages, it can highlight how the

apparent rules of a language play out in a text by translating it. The translation provides us with a kind

of snap-shot of the operations of a language, constrained within the institutional perameters of a text-

type. Failing determinacy in translation, there is no tangible sense to the idea of a rule of language for

meaning would be irreducibly context-sensitive, open-ended, and indeterminate. Translation, at least,

provides us with a manner of testing the “rules” of a language at a given juncture in its operations,

under certain conditions.

Cappelen and Lepore also want to resist the conclusion that meaning is irreducibly context-

bound. And there is good reason to resist this conclusion: for Wittgensteinien reflections on rule-

following and normativity in contexts show that contexts are normatively indeterminate. If meaning is

irreducibly context-bound, it is irreducibly indeterminate.

Cappelen and Lepore can be credited for recognizing, in their own way, the Wittgensteinian

lesson --- though they do not credit him with the intuition. On their account, there is no algorithm to

determine the pragmatic implicature of sentential utterences (of English sentences) in a context. This

is their doctrine of Speech Act Pluralism. On Cappelen and Lepore’s account, the best we can do is

specify the “proposition semantically expressed” by all utterances of a sentence, which is specified by

a T sentence: e.g., “Rudolph has a red nose” is true if and only if Rudolph has a red nose. The latter is

the “minimal content” of the sentence (Cappelen and Lepore 2005, 152). However, given that

Cappelen and Lepore’s theory of meaning is only a theory of meaning of English, it does not specify

the minimal content of an English sentence in all contexts. It merely specifies its meaning in one

328
massive context—the context of the English language—or only those contexts where all speakers

speak English as understood by Cappelen and Lepore.

The Anglophonic-centred nature of Cappelen and Lepore’s account warrants stress. They

explicitly formulate “Radical Contextualism” as the thesis that “No English sentence S ever

semantically expresses a proposition” (Cappelen and Lepore 2005, 6). They also recognize that their

theory is based upon Kaplan’s insight that there are no “monsters” in English (Cappelen and Lepore

2005, 117), where a monster is “an operator “which when prefixed to a sentence yields a truth if and

only if in some contexts the contained sentence (not the content expressed by it) expresses a content

that is true in the circumstance of that context” (Kaplan 1989, 510). Thus, for instance, if there

were monsters in English, we could be Contextualists, according to Cappelen and Lepore.

However, following Kaplan, they rule out this possibility. Cappelen and Lepore quote Kaplan

approvingly (I include their explanatory footnote):

(9) In some contexts it is true that I am not tired now.

For (9) to be true in the present context it suffices that some agent of some context not be tired at the
time of that context. (9), so interpreted, has nothing to do with me or the present moment. But this
violates Principle 2! [“i.e., the thesis that indexicals pick out their referents directly from the context
of utterance, without mediation. This means that the value of an indexical is fixed by the context of
its utterance, and cannot be changed by the logical operators in whose scope it may occur.”
(Cappelen and Lepore 2005, 117 fn. 20)] Principle 2 can also be expressed in a more theory laden
way by saying that indexicals always take primary scope. If this is true-and it is-then no operator can
control the character of the indexicals within its scope, because they will simply leap out of its scope
to the front of the operator. I am not saying we could not construct a language with such operators,
just that English is not one.” And such operators could not be added to it. (Kaplan 1989, 510)

If Cappelen and Lepore are merely articulating a defence of Kaplan’s theory about monsters, it

follows that Contextualism is wrong, on their account, for purely contingent linguistic reasons. There

are two problems with this. First, it would seem that the Contextualist simply needs to speak a

different language than the one that Cappelen and Lepore are speaking in order for their thesis to be

329
correct. And I wonder what makes Kaplan, Cappelen and Lepore so sure that English is not a

language that contains monsters—indeed, Cappelen and Lepore’s argument suggests that

Contextualists think it is.146 Moreover, the very fact that Contextualists insist on speaking as though

there are monsters in English, and moreover since they appear to be speaking the same language that

Cappelen and Lepore are speaking (publishers, after all, would consider their texts as equally written

in English), there is strong empirical evidence that Cappelen and Lepore’s language actually contains

monsters. If Principle 2 were the only way to make sense of indexicals, perhaps an argument could be

made that no language could contain monsters. However, Principle 2 is a particular conceptualization

of the phenomenon of indexicality, and Kaplan notices this. Thus, he allows that there could be

languages that accommodate monsters. This brings us back to the earlier point, that either

Contextualists simply need to speak a different language to be correct, or Cappelen and Lepore may

be wrong about English. (Indeed, if rule paradoxes apply to a practice such as “English” which is not

an institution, then it seems that it is an open, rhetorical question as to whether English contains

monsters or not.) Secondly, and more importantly, it would seem that a philosophical text espousing

Contextualism would be untranslatable into English, if one requires such an operator to articulate the

view, and if English is truly a language that lacks such an operator. But if Contextualism is

untranslatable into English, it would seem that translation is indeterminate, for all attempts to

translate the Contextualist’s texts into English would be equally acceptable with no objective manner

of deciding between alternative translations. This is grounds enough for us to reject Cappelen and

Lepore’s theory. But it also ought to trouble them, for it suggests that their theory of semantics

renders meaning equally indeterminate.

Given what has already been argued in chapters 2 and 3, TTS leads us to conclude that the

question of whether English has or does not have monsters is indeterminate and uninteresting. It is a

matter of indeterminate significance for we can always introduce, in translation, novel expressions in

a text to allow translation to be determinate, and thus, even if it should be the case that English

330
speakers at large do not recognize monsters, we can all the same introduce such creatures, with the

proper framing devices, in a TT. I realize that Cappelen and Lepore have no stock in TTS, but I think

that the criticisms and concerns that I have brought up here suggest a rather deep criticism of their

project: namely, that it is philosophically uninteresting for it reduces a matter of philosophical interest

(namely the question of whether meaning is context-bound) to a matter of linguistics. But linguistics

as such cannot solve the philosophical problem, for languages are changeable. Of course, Cappelen

and Lepore, quoting Kaplan, do not understand languages this way. But their emphasis on the

peculiarities of English (as though there are determinate facts about the semantics of the language)

renders their view vulnerable to a major criticism.

The Anglocentric nature of Cappelen and Lepore’s thinking, and the fact that they believe

that a theory of meaning should be for one language, ought to trouble us, for these support the view

that their so-called Minimal, Insensitive Semantics is really an unassuming version of Contextualism

itself.

If Cappelen and Lepore’s account of sentence meaning were truly context-insensitive, it

ought to be able to specify what the minimal content of “Rudolph has a red nose” is in Swahili. What

is the Swahili minimal content of this sentence in English? Can we even make sense of the request?

Certainly. Either the minimal content for the Swahili speaker is the inscription “Rudolph has a red

nose”, or it is what the English sentence speaks about. If it is the inscription “Rudolph has a red

nose”, the theory has problems, for our Swahili speaker will not be able to determine what counts as

this minimal content from a background of English utterances. Even if our speaker isolates this

inscription, she may still not know whether it is really a proposition, and not an interrogative or

imperative, or not merely one long word. Cappelen and Lepore believe that the major justification for

minimal content is that it provides a way to check against confusion and errors in communication. If

nothing else, it is a starting point (Cappelen and Lepore 2005, 185). However, our Swahili speaker

will not be able to employ the inscription as a check against error for in isolation its utterance would

331
be consistent with most any Swahili sentence. With language lessons, our Swahili speaker could

come to appreciate the minimal content of genuine English sentences, but that would be a different

context. If the minimal content of a sentence is what is spoken about, then we are back to all of the

problems of radical translation for the Swahili speaker may, with time, come to a fairly systematic

understanding of what this proposition means, but there may be many alternative interpretations that

are not all compatible with each other but equally plausible. Nothing like a cross-contextual minimal

content will emerge from this route.

Cappelen and Lepore might object that it is a violation of their theory to even pose the

question of what the minimal content of a sentence in one language is, from the perspective of

another language. Languages are not contexts but systems, and thus what one is attempting to do is

provide an account of the content of a sentence from one system in another system, but this is not the

way in which the content can be specified. The trouble is that Cappelen and Lepore have high hopes

for their patently linguistic and Anglocentric theory:

The common thread that runs throughout our criticism of Contextualism is that it fails to account for
how we communicate across contexts. People with different background beliefs, goals, audiences,
perceptual inputs, etc. can understand each other. They can agree or disagree. They can say, assert,
claim, state, investigate, or make fun of the very same claim. No theory of communication is
adequate unless it explains how this is possible. Contextualists cannot provide such an explanation.
The solution proposed in this book is a context-insensitive semantics (i.e. the view we call Semantic
Minimalism), combined with Speech Act Pluralism. (Cappelen and Lepore 2005, x)

Let us be clear: Cappelen and Lepore set the success of Semantic Minimalism to explain how

communication across contexts is possible as the criterion according to which their theory is to be

judged. They do not attempt to argue for their view on the basis of the more modest aim of being able

to explain how communication across English language contexts is possible. Of course, they do not,

for this would be to concede that their view is really not context-insensitive, but a very wide and

modest form of Contextualism. Thus, in motivating their view, they set themselves high goals of

332
context insensitivity. In the fine print, it becomes apparent that theirs is really just a view about

communication within English language contexts. In reality, their view fails on the ambitious

objective. Their implicit view that language is really an abstract system of rules and that English does

not contain monsters is really only effective to explaining how communication across contexts is

possible if one has narrowed the range of relevant contexts that one is interested in.

There are two ways in which Cappelen and Lepore’s account is really not an account of

communication across contexts. First, Cappelen and Lepore are not attempting the interesting task of

attempting to understand how a person in context A can communicate with someone in another

context B. This would constitute a genuine instance of explaining communication across contexts.

Rather, what Cappelen and Lepore are attempting to answer is the relatively uninteresting question of

how one person can move through several contexts and communicate with the people in those

contexts.

Secondly, as we have seen, Cappelen and Lepore cannot really explain how one and the same

person can travel through several context and communicate in each one, because of the explicitly

linguistic conception of meaning that Cappelen and Lepore present—an account of meaning that

restricts meaning to a particular language, namely English. Real life contexts are characterized by

linguistic diversity, interactions in multicultural environments, as well as communication across

species and across levels of intellectual acuity within our own species. A pragmatic conception of

intracontextual communication can take this challenge in stride for it construes the problems of

intracontextual communication wholly in terms of pragmatics, and not (determinate) semantics. Body

language, gestures, emotional expression and the extension of the principle of charity are perhaps the

most basic means that animals (human to human, and human to non-human) manage to communicate

across the various contexts they negotiate. Such means of communication are possible for the ultimate

criteria of success in communication are pragmatic and not semantic determinacy. It is important to

note that Cappelen and Lepore can have no principled objection to a pragmatic conception of

333
communication that takes communication to be contextually negotiated, for they themselves appear to

affirm such a view to the extent that they endorse the doctrine of Speech Act Pluralism. Their

insistence that there be a minimal content that underlies this pragmatic process is thus quite odd.

Cappelen and Lepore might argue that they are not being dogmatic in tying Speech Act

Pluralism to a minimal content. Rather, they are demanding simply that there should be some order

and no surprises in communication. If language is to be a means of communication, the ground rules

must be simple and apparent to all:

…Conversations happen fast. Someone speaks; sounds hit the audience’s eardrums; they must be
processed; often a reply is expected immediately. There’s little time for reflection and exploration. It
is because our linguistic devices are so effective for communication that conversation is able to be as
fluid as it is. They are easy to use and it is not surprising that they are easy to use.

To use the dubious metaphor of language as a tool for a moment: if words are tools, then they had
better be pretty easy to use because they don’t come to us with instruction manuals and even if they
did, there would be no time for us to consult these instruction manuals when we’re steeped in the
middle of a fast and furious conversation. (Cappelen and Lepore 2005, 112)

If this is a defence of the necessity of there being minimal content, it is fairly weak for words

unconstrained by any type of content are even easier to use than ones that are constrained by content.

Certainly, instruction manuals play no part in the art of conversation, but this only strengthens the

view that in conversation one is creating a type of meaning through linguistic activity, rather than

obeying determinate rules or being restricted by content. Cappelen and Lepore however expand on

their desire to maintain minimal content as a type of model for meaning, and context-sensitivity as a

rare exception:

We’re highlighting these obvious features of communication in order to register a very simple point:
If an expression e has its semantic value fixed in a context of utterance, that had better be obvious to
all of us. Context sensitivity can’t be some obscure phenomenon that you need to read scholarly
books and articles about in order to recognize and master. Context sensitivity is a surface
phenomenon. Every speaker knows it when he’s confronted with it; and he knows that every other

334
 competent speaker of his language knows it as well, and all speakers know how to exploit context-
sensitivity in the heat of a conversation. (Cappelen and Lepore 2005, 112)

We might ask why context-sensitivity cannot be obscure. There are several good reasons to believe

that context-sensitivity only becomes apparent after reflection. First, an interest in philosophy is often

sparked by the apparently obscure nature of the context-sensitivity of philosophical terms.

Philosophically inclined people often notice the diversity of uses of “good,” “right,” “real,”

“knowledge,” to name a few such terms. Such uses appear as context-sensitive because (a) the items

that they are used to refer to change according to circumstances of use, speakers, times and places,

but also, (b) there is such a great diversity of views surrounding these terms that, prima facie, one

must understand the different existential factors of speakers and authors to make sense of their

idiosyncratic uses of these terms. It is only the unreflective who is untroubled and blind to such

context-sensitivity. For the unreflective, there is no real problem for moral philosophy, or

metaphysics. But this ignores the very real social problem of coordination requiring some consensus

on moral and metaphysical issues, and the very real diversity of such views in the world.

Secondly, there is good anthropological evidence that speakers are generally unable to

explain the meaning of context-sensitive expressions and that it takes a great deal of effort for them to

overcome this inability.

The distinguished linguistic anthropologist Michael Silverstein observed, in his now classic

article “The Limits of Awareness,” that native speakers are often able to explain the reference of

context-sensitive expressions if they are anaphoric, but if they create a context anew they often have

difficulty explaining the sensitivity of the expression to context and will merely provide the reference

of their expression as the meaning, even though the criteria for context-sensitivity is obvious to the

observer who is not a native speaker (Silverstein 2001 [1981]). Silverstein thus presents us with

empirical confirmation that accounts like QI may be plausible, even though native speakers are

unlikely to explain the meaning of such expressions in terms of context-sensitivity.

335
 The main reason that context-sensitivity likely remains obscure is because it matters little

within a context of communication, for here the exchange is almost completely characterizable in

terms of a shared context. What implications people draw from each other’s utterances will be a

matter settled by contextual factors that are common to both speakers, and hence it is easy to imagine

that this shared context generalizes over all contexts. We know however that it is the particular

context of speakers that plays a great role in helping them communicate because people within

contexts can communicate despite malapropisms, and sometimes even despite linguistic divides.

Context sensitivity only becomes pressing when we attempt to extract semantic content from a

context. Indirect speech reports are the simplest form of such an exercise, but it is really a relatively

uninteresting attempt at such an exercise. To elevate indirect speech reports to the paradigm of

semantic content extraction from a context (as Cappelen and Lepore do) is like elevating a visit to

one’s neighbour’s home to the paradigm of world travel. Visiting one’s neighbour and indirect speech

reports are alike in so far as they both share a great very many contextual factors rendering them part

of a one larger, culturally negotiable context.

Cappelen and Lepore will likely defend themselves by insisting that their account was never

meant to cover cases where monoglot English and Swahili speakers attempt to radically understand

each other. Indeed, it seems that such a scenario is far from the concerns of semantic minimalism. I

personally have no sympathy with this defence in light of the big claims that both Cappelen and

Lepore make for the prospects of their Insensitive Semantics. If they are truly concerned with

communication across contexts, they must consider such multicultural and multilinguistic scenarios.

Perhaps Cappelen and Lepore do not think that such scenarios are typical. But this is a myth that can

only be paid for by those who enjoy an exceeding amount of social privilege. Magazines, TV ads, and

people in high places (i.e., faculty in universities) will tend to look and sound like one’s self—they

will even speak the same language if one has a certain privilege that not all philosophers have. To the

336
privileged, diversity disappears as a strange phenomenon that only occurs at the boarders of

civilization.

Communication and translation ought not to be difficult topics to come to terms with, as they

are ubiquitous throughout human history, yet the intellectual tradition that we in the West inherit from

Europe is informed by two of its worst vices: ethnocentrism, and anthropocentrism. Most human

cultures are ethnocentric. For instance, much recent scholarship has noted the xenophobic nature of

Indian, Vedic culture (cf. Halbfass 1991). However, such a culture was never anthropocentric to an

extent common in today’s, Western-dominated world. The idea that animals were moral patients, for

instance, or that they could communicate, was never disputed. But Europe, particularly in its Judeo-

Christian stage, has made an art of these human weaknesses by strengthening and combining them.

The perfection of these evils has resulted in benefits for humans, particularly of Western descent:

political and cultural domination of the entire globe, along with a blueprint for industrial progress that

views the entire world as ours to manipulate and destroy in our bid to maintain this advantage (even

humans who do not look or sound like us). It is our grave tragedy that we are only now, at this

incredibly late date, starting to appreciate as a species that this ethnocentrism combined with

anthropocentrism is a moral, political and environmental disaster of unprecedented proportions.

With respect to semantics, these two vices combine to form a conception of meaning and

communication that is at once culturally specific (and Anglocentric in the case of Cappelen and

Lepore), to the extent that individual languages are considered to be the bearers of meaning, and

specific to humans, to the extent that language as a distinctly human practice (defined in terms of a

syntax, vocabulary, and some underdetermined connection to truth) becomes the paradigm for

communication. These two vices contribute to linguistic bloat: language’s importance is conceived in

such great terms that theorists who claim to be explaining how communication across contexts is

possible cannot even contemplate the possibility of real-life contexts of communication where a

337
language is not shared. Indeed, the same bloat results in our inability to explain how meanings

expressed in one language could be reproduced in another via translation.

Cappelen and Lepore might regard my moralizing as a digression. Communication, they

might claim, involves all sorts of semantic phenomena that cannot be explained by pragmatics. In

communication, people “assert, claim, state, investigate, or make fun of the very same claim”

(Cappelen and Lepore 2005, x). This requires a fair bit of semantic structure that mere pragmatic

felicity cannot account for.

The trouble with this response is that it assumes that there is a clear line between pragmatic

interaction and the more sophisticated interaction characterized by assertion and investigation. But

there is not. The specialized interaction that is “fast and furious” is simply a special case of the

general pragmatics of communication. What seems to make all personal, intra-contextual

communication possible between animals of any kind is a basic grasp of body language and

emotional expression. Indeed, the very act of agreeing and disagreeing is strongly tied to this. And in

practice we rely upon these cues all the time in conversational settings to determine whether we are

having a felicitous interaction with our interlocutor or not. Much nonlinguistic interaction and

linguistic interaction can go off without a hitch, relying upon this basic animal aspect of our being,

and thus agreement and disagreement can occur without their being anything like a determinate

content for interlocutors to disagree or agree about. Disagreement and agreement of the sort that

Cappelen and Lepore concern themselves with is really a very specialized version of this general

pragmatics of interaction. It involves the prospect of objectivity in discourse made possible by the

textual promise of determinacy in meaning. Objectivity of the sort that relies upon determinacy of

meaning appears as a visceral possibility as linguistic interaction begins to approach institutional

standards of text-types. On Cappelen and Lepore’s account, it is rather mysterious what and how

people draw pragmatic implications from speech interaction and that they could converge on such

interpretations. But against TTS, it is not mysterious at all. The more trained linguistic actors are in a

338
text-type institution, the more speakers will tend to converge upon interpretations of their

interlocutors. They will moreover have the resources of the text-type features to set the foundations

for discourse that aspires to objectivity and reflexivity. They will have institutional means of raising

issues and resolving disagreements. In other words, the type of fast and furious interaction that

Cappelen and Lepore have in mind only becomes possible as participants begin to treat each other’s

utterances as though they were parts of a text translatable by a specific text-type. Once the

conversation comes to an end, they will have a very good candidate text for translation under the text-

type that the interlocutors are trained in. In between this highly specialized discourse characterized

by a high degree of precision and the radical experience of attempting to deal with people who do not

speak one’s language, or animals, for that matter, who don’t understand your language,147 are the

workaday linguistic interactions we have with people who share our language and sometimes even

with those who do not. Here we trade in the general indeterminacy of language. Sometimes this is a

matter for vexation; often we pay no heed as long as our day runs smoothly. The type of interaction

that Cappelen and Lepore want to explain (and ironically do not, except via recourse to “speech act

pluralism” which appears to be no explanation at all) is important, but it is not exhaustive of

communication and only intelligible as the culmination of a gradual social complexity of institutions

that rests upon our more basic biological aspect as communicative beings that we share with all

animals.

In short, the notion of a minimal content plays no role in explaining communication as such.

And it certainly plays no role in explaining specialized communication by highly trained experts in

text-type institutions. Here, it is the virtual application of a text-type to the conversational setting that

does the work. Our nature as linguistic beings makes possible discourse characterized by the

possibility of objectivity. However, our language on its own does not ensure the objectivity of our

interaction or the determinacy of sense. Unless we are in a position to translate our linguistic artefacts

by subsuming them under a text-type institution, they remain semantically indeterminate. Our typical

339
linguistic interactions with our fellows, outside of the constraints that a text-type institution allows us

are not that different in kind from the wag of a tail or a chirp of a bird. It allows us to get through the

day pleasantly, but it’s far from clear what we determinately mean.

VII.2.4 TTS, Derrida and the Fulfilment of the Promise of

Insensitive Semantics

The first philosopher in the Western tradition after the linguistic turn to truly put thinking

about language and meaning on the path of context insensitivity in response to the rise of
148
Contextualism was Jacques Derrida. In his classic criticism of Austin, he noted that while Austin

had attempted to put semantics on a new footing, he was really perpetuating the dominant trend of the

Western tradition, to conceive of meaning as speech-oriented, conversational and, hence, context-

bound (Derrida 1982 [1971]). This Western tradition, which in the analytic tradition has come to be

dubbed “Contextualism,” is roughly what Derrida has called logocentrism. Derrida rightly notes that

the very possibility of genuine meaning—what I have been calling determinate meaning—is based on

the ability of inscriptions to survive our absence.149 Derrida also noted that the dominant trend’s

anthropocentricism was of a piece with its logocentrism (Derrida 2000, 406). The idea that meaning is

a particularly human affair is simply Contextualism on a species scale. Cappelen and Lepore are to be

praised for their suspicion that context can play no role in determining meaning, but they, like Austin,

have failed to break the shackles of Contextualism by conceiving of communication in terms of a

shared language.

Derrida, unfortunately, made little headway with his insight. Derrida correctly noted that a

distinction has to be drawn between speech and writing (Derrida 1974). Derrida’s insight thus

resonates with the argument that I have been presenting. Unfortunately, no sooner does Derrida draw

the distinction than he collapses it. On his analysis, all utterances are really instances of writing, text

or marks. If text is coextensive with speech, we can no longer explain how texts survive context, for

340
some, namely utterances, do not. This has allowed him to make the rather strange argument, reviewed

in chapter 2, that translation is a type of forced assimilation, or a paradoxical effort to destroy and

conserve meaning at the same time. Indeed, the rather tortuous and paradoxical nature of Derrida’s

philosophy at many points serves as a reductio ad absurdum of what happens when one recognizes

the speech-writing distinction, and yet fails to properly respect it by identifying language with text

and not with speech. Over all, Derrida is to be commended for truly grappling with the issue of

context insensitivity in semantics, while Cappelen and Lepore fail to even understand the possibility

of a solution to the problem. However, that Derrida comes so close to jumping across the canyon that

is the problem of insensitivity in semantics and yet fails makes for a very messy and tragic accident.

TTS does not fall into the trap that Derrida does. Texts are different in kind from languages,

or speech acts. Texts come in types, and they retain an identity despite translation into distinct

languages. Languages, in contrast, are the semi-rule-bound, indeterminate semantic resources that

humans typically make texts out of, and it is the texts that survive context. Texts of a specific type

dictate to us how they are to be read. The context that a text finds itself in has no bearing on its

meaning. Rather, it is the text-type that sets the conditions of what a text means determinately, across

contexts. And, when appropriate, the text-type allows the text to be read in a manner that preserves

features of the context of the author. This contrasts sharply with speech acts and utterances, which

must always live in a certain context in order for them to have their particular semantic effect.

TTS, not Cappelen and Lepore’s Minimal Semantics, is the only account of meaning that can

explain how we communicate across contexts. No linguistic conception of meaning can truly explain

how a person in one context can communicate to a person in another context, for wherever language

can be traded between speakers, we seem to have one language game (whether the context be

spatially definable as a local interaction, or a telephone conversation spatially mediated by

technology). Texts, in contrast, are determinately meaningful, portable semantic phenomena that are

able to transcend contexts and thus facilitate communication across contextual boundaries. Unlike

341
language, texts can survive their authors and thus cross contexts defined by space and time. They can

even survive changes in their linguistic form. To explain how they determinately transcend contexts,

we need an account of text-types. Minimal Semantics, in contrast, cannot explain the semantics of

texts, for texts come in a variety of types, and their translation is a matter of determinate equivalence.

The closest that Cappelen and Lepore come to the pluralism of TTS is in their doctrine of Speech Act

Pluralism; however this is explicitly an account of the pragmatics of utterances, which are

indeterminate on their own account.

TTS is also part of an explanation of how people communicate within a context, and through

all the contexts that they travel to the extent that TTS regards language to be merely one of many

semiotic devices that, apart from their textual environs, are semantically indeterminate.

Communication intra context, on this view, can only be judged according to pragmatic criteria. And

thus language poses no obstacle to communication, between humans or between species. If we relied

upon the determinacy of language to facilitate communication, and if this was cashed out in terms of

Cappelen and Lepore’s minimal content, only speakers of the same language (actually, only speakers

of English) would be able to communicate with each other. But life in a big, cosmopolitan, multi-

ethnic city will put the linguistic account of communication to rest. And reflections on the demands of

translation as a test of meaning suggest that Cappelen and Lepore’s minimal content is woefully

inadequate to explain determinacy in meaning. At best it provides schemas for intralinguistic

translation, and no direction for interlinguistic translation, let alone texts of different varieties. (This is

why their theory can only be a theory of the meaning of English: Semantic Minimalism has no

resources to translate into other languages the biconditional sentences that shows minimal content.) If

Cappelen and Lepore were to defend their account by insisting that the cosmopolitan scenario is the

exception and not the rule, and thus not what semantics should focus upon, they would be simply

reviving the ideal and myth of the Old World, that uniformity and cultural homogeneity is the norm.

This is almost never true, and those who put great effort into convincing their fellows of it show

342
themselves to be both sensitive and troubled by the very real and subversive presence of diversity in

their midst. Those who ignore this reality altogether show that they have done nothing to self-

critically disavow the privilege that they are the arbitrary heirs of.

VII.2.5 Moderate Contextualism and QI

As this section comes to a close, it should be apparent that though QI and TTS hold that there

is nothing sacred about a basic list of context-sensitive expressions that Kaplan identifies and

Cappelen and Lepore seek to defend, it is not a form of Moderate Contextualism. The reason is that

TTS is not a form of Contextualism at all. If it is an ism, it is a form of Cotextualism. If any semantic

theory has claim to being insensitive to context, it is only TTS.

QI, thus, as a text-type feature of the philosophical text, is context insensitive in an important

sense: namely, it structures the text-type of philosophy, and thus every individual philosophical text

bears it, regardless of the context the text is in. It is QI that largely sets the conditions according to

which we properly read philosophical texts. This is the sense in which all determinate meaning is

context insensitive, according to TTS. QI will also be context-insensitive in another sense that is

particular to the philosophical text-type: it will defer to objective standards of relevance as the

grounds of the argumentative or debate-oriented nature of philosophy. It is, however, context-

sensitive in an important way: to fully understand its operations in a text, one must take into account

the existential input of an author, in addition to the fictional authors specified in the text, which it

defers to.

VII.3. QI, Belief and Knowledge

I wish to close this chapter with some observations about QI that will hopefully tie up some

loose ends.

As noted, QI is an account of the text-type feature of philosophy. It is not primarily an

account of what people in general mean when they use philosophical terms. This is not to say that QI

343
is utterly disconnected from ordinary language use. In cultures where the institution of philosophy has

deep roots, ordinary use of key philosophical terms will be strongly influenced by QI due to the

institutional reverberation of the text-type of philosophy. We could thus recognize QI as a kind of

second order account of the meaning of philosophical concepts. Ordinary speakers will likely

associate philosophical words with rather elaborate theories, or if they do not individually have such

an association or conscious awareness of such diversity, such theoretical diversity will all the same be

a type of sociological fact of key philosophical term usage. Reflective members of such a linguistic

community—who are usually few in number—will likely be sensitive to the variety of theoretical

commitments that are associated with key philosophical terms, and apparent tensions and

inconsistencies in such usage will invite them to think more deeply about the resolution of such

conflict. But we shouldn’t expect this to be the norm. As noted, there is strong empirical proof that

native speakers are often poor at understanding and explaining the context-sensitivity of the terms

that they use (even though this may be obvious to the outsider), and thus they may, for all intents and

purposes, deny that there is anything context-sensitive to the philosophical concepts they employ.

However, such denials are the luxury of the language user and not the translator.

We could imagine a society in which philosophy as a text-type institution has no great roots.

For such peoples, no term would bear any obvious sociological association with thin or thick concepts

of QI. The range of theoretical options associated with most of their vocabulary would be narrow, and

the idea of significant and profound disagreement on potentially philosophical matters might seem

odd. Thus, they may have terms that are ordinarily understood as referring to the proper organization

of their society, or the differentia of non-illusions, but they will not be given to reflection on the

context-sensitivity of such terms. None of these terms will be value terms, in our text-type sense,

thick or thin.

Such a society may all the same appear to trade in subjective-description concepts. They may

express likes and dislikes, but in so doing they neither express anything controversial nor do they

344
disabuse their neighbours for expressing their likes and dislikes. If we were not careful, we could

mistake such a society’s views as constituting the raw content for texts of value discourse. But this

would be a mistake. Values are inherently controversial. This is why the problem of translating value

discourse is prima facie puzzling, for it seems to belie the putative ordinary convergence on linguistic

usage that linguistic theories of meaning put forward as a condition of translation. “I like vanilla” is

not a value claim. “Vanilla is good” is a value judgment. Psychologically, the two may be linked, but

semantically they have a very different function in texts. The former could form the substance of an

autobiography, or descriptive psychology, the translation of which would never have to be formed by

recourse to value terminology. We could even translate a text that employs value terminology as

though it were a mere device for subjective description in a manner that stripped it of its value

content. In such a text, “vanilla is good” could be converted, determinately, into “I like vanilla.” Such

a conversion might be called for in a text of sociology or psychology, for instance, or perhaps in a

literary text or poem. But none of these translational conversions would instance an effort to retain

value discourse through translation.

Part of the problem in understanding the distinction between subjective-descriptive concepts

and value concepts is that in ordinary talk, aesthetic judgments seem to be closely tied to subjective

descriptions, for both seem to be subjective. But semantically, they are distinct. This was noted early

by Kant, when he noted that judgments of like and dislike, or pleasure and pain, are not aesthetic

judgments at all for they directly relate to the psychological desire for continued experience or

aversion to continued experience. Judgments of taste, in contrast, attempt to evaluate whether an item

is beautiful or ugly independently of a presupposition of the continued existence of the object (Kant

1974 [1790], 39 §2). Beauty, even if subjective in the sense of not existing independently of the

judger, can still be controversial in so far as we can philosophically argue over whether one judges

properly in deference to the proper criteria of beauty. Even in the case of relative judgments of

beauty, a similar criticism is possible. We could, for instance, identify relative principles of beauty

345
that a judger of taste fails to respect. In the case of likes and dislikes, and our description of them, we

can perhaps criticize people for failing to properly describe their psychological states, but we cannot

identify the error by way of a principle of beauty or taste.

Another factor that confuses subjective description with value talk is that some philosophers

have argued that value terminology is simply a means by which people describe their psychological

states. Hobbes, for instance, held this position with respect to value talk outside of civil society.150

This, however, is a patent confusion. It is an attempt to accommodate the diversity of evaluative

views without any consideration to translation, for a translation of autobiographical, descriptive

psychology and one of philosophy operate according to very different principles. If QI and TTS are

correct, Hobbes cannot be right in so far as he attempts to provide a second order account of the

meaning of value terminology. Of course, his move may have some payoff for political philosophy,

as the Contractarian tradition has shown. But it is one that is often bought at the unfortunate expense

of semantic clarity. We could still translate Hobbes and his confusion, in so far as we identify his

psychologistic-cum-constructivist theory as the theory that he defers to in his use of moral

terminology. Moreover, since he actually rarely uses terms such as “good” and “right” but only

mentions them, we could understand him—as we do all philosophers engaged in a dialectical process

of arriving at the right theory of the good—as relying upon some explicit sense of what is relevant to

theory and referent selection in value talk while leaving the theory variable blank.

Gauging whether a society has or does not have value concepts, and what terms instantiate

such concepts, has seemed difficult in the past. In my own research in Indian philosophy, the

challenge in the case of demonstrating that “dharma” is indeed a thin moral concept was that what we

have on the whole left to us by history are texts, and texts do not emote in a manner that we believe is

psychologically and sociologically consistent with the controversial nature of values. Philosophical

texts articulate views employing philosophical terms, but philosophical texts typically do not glower

at those who they disagree with. Even when discussing issues of the most ethical seriousness, they

346
can come across as exuding an air of dispassion that makes such value terminology blend seamlessly

with seemingly nonevaluative terminology of an ontological or epistemological theory. I was able to

find clear textual cases recording the controversial, anger-provoking nature of “dharma” that we

would expect from controversies over moral issues in India’s major philosophical traditions.

However, this was clearly not the dominant connotation in philosophical texts, and the expectation

that this emotional function would be instanced throughout the philosophical corpus was part of the

problem to appreciating the thinness of “dharma.” Its meaning, if it had a moral meaning, could not

be its expressive or emotive function.

I was able to solve the problem of the meaning of “dharma”, and successfully depict it as a

thin concept, because the hypothesis in fact explained a lot of sociological and textual data that was

otherwise baffling. Baffling was the obvious connection “dharma” has with concerns of obviously

moral significance, such as social obligation, along with its use for a variety of cosmological and

metaphysical items. More puzzling still was the fact that every Indian philosophical school insisted

upon using the term, and having some position on the topic of dharma, but all disagreed as to what

dharma amounted to. The hypothesis that “dharma” is in fact a thin moral term explains all of these

strange textual factors, provided that we are able to understand the thinness of moral terms not in

terms of the substantive moral theories that are so popular in the recent western tradition—theories

that are explicitly anthropocentric and positivistic. Moreover, the numerous seemingly nonevaluative

uses of “dharma” could be understood in this light too. When, for instance, a philosopher used

“dharma” to set off a realm of social action, but one that they were not particularly passing judgment

on, we could understand this use of “dharma” after the manner of uses of “ethics” or “moral” that

have this categorizing function, as in “ethical judgment,” or “moral controversy.” Such a categorizing

function is wholly consistent with the thinness of “ethics” and “moral,” provided that the thinness is

not understood in an expressive or emotive manner.

347
 The solution I offered involved understanding the very occurrence of “dharma” in a text as

signalling to the reader that the issue under consideration was of moral importance to the author, and

thus the challenge for the historian and student of Indian philosophy was to understand the systematic

connections between the instances of moral terminology in a text, as articulating a moral theory that,

in many cases, was coextensive with the systematic concerns of a philosopher. The seemingly chaotic

uses of “dharma” in Indian philosophy fell into line relative to the moral theories of the philosophers

in question. And any time we seemed to think that it was not possible to understand the usage of

“dharma” in accordance with a moral theory of the author, the obstacle, I argued, was our prior

commitment to substantive moral theories that were inconsistent with those of the philosophers we

were studying. We thus were to take uses of “dharma” as the textual data according to which we

could construct a moral theory of the author, provided that we were willing to interpret “dharma”

from the start as a thin moral term.

I had made this argument about interpreting “dharma” and translating Indian moral

philosophy in a manner that deferred to general considerations of translating ethics and philosophy,

but I must admit that my approach to the problem was not as systematic as it could have been. I had

thought that I could give an account of the meaning of moral terms and judgments that was explicitly

second order (i.e., in a manner that did not attempt to presume answers to substantive questions) and

in some respects the approach was more successful, I believe, than the indexical accounts examined

here in section 2. I had argued that what we require is an account of moral judgments and terms that

identified their semantics without attempting to comment on substantive questions. The goal was to

deliver the peculiar flavour of moral discourse in an account of moral semantics, without thereby

presuming to compete with the views that one wishes to translate. However, I had not properly given

consideration to the problem of translation and semantics, and thus could not proffer a textual account

that would truly offer an anchoring for the second order account I was seeking. There is a trivial and

translationally indeterminate sense in which every philosopher means by their use of philosophical

348
terminology what their philosophies say. The trick is providing an account of moral semantics that

will allow one to preserve this more ordinary sense of meaning in translation, and in so doing,

translate the philosophy of those we wish to translate. Only a textual approach to translation and

meaning can truly offer this.

QI is in some important sense a maturation of my effort to understand the meaning of

“dharma.” It fully captures the thinness of “dharma” for, on the QI interpretation, Indian philosophers

articulated theories selected for their social implications with this term. The theories in question were

systematic and also functioned as the metaphysical, epistemological and sometimes even logical

theories of the philosophers, thus “dharma” would be found to refer to items that were not obviously

moral to the Westerner, but say, cosmological or metaphysical on their account. However, QI retains

the communicative insight that guided my argument in Ethics and the History of Indian Philosophy.

There I argued that the very presence of a term like “dharma” in a text is proof that the philosopher

considers the items under consideration as moral in significance. QI recognizes this by way of the

narrow reading of the objectivity clause, so necessary in the translator’s task to fix the content of a

philosophical text (cf. 323).

So much for societies with philosophical terminology. What of societies that are pre-

philosophical? Bernard Williams speculated about such a society, which he calls the “hyper-

traditional society” (Williams 1985, 142). He speculated that such societies would have thick terms,

but no thin terms. Williams introduces the classic, two-components account of thick terms, according

to which they combine an empirical, descriptive content, along with a thinner evaluative content. Thin

terms, on this account, have only the latter. The introduction of reflection and theoretical controversy

by way of the thin term would undercut the possibility of moral knowledge (Williams 1985, 167).

Thick terms have their primary use in the hyper-traditional society, according to Williams. Thick

terms play a role in moral knowledge in this context because their empirical content provides them

with grounding in the world. This not only permits the possibility of them being inferentially

349
supported by non-evaluative, empirical judgments (Williams 1985, 154), but also allows for the very

possibility of convergence on moral claims. Social convergence on judgments, whether they be the

judgments of science or ethics, is a condition of knowledge (Williams 1985, 171). With the

introduction of thin terms characteristic of reflection on moral issues, one can no longer form beliefs

of the sort formed with thick concepts (Williams 1985, 167). Thin terms highlight the radical

contingency of the combination of expressive and descriptive meanings in thick terms and thus

undercut the empirical foundation that allowed for ethical knowledge in the first place. The

sophisticated ethical deliberator would thus not be able to make the types of judgments they did

before (Williams 1985, 167). They would see their earlier ethical judgments as knowledge only in a

non-objectivist sense that ties knowledge to bare social convergence. They could not bring

themselves to make those judgments now that reflection has shown them the possibility of objective

knowledge (cf. Williams 1985, 147). Barring thick concept judgments, there seems to be no

prospects for ethical knowledge, on William’s account. For there to be objective ethical knowledge,

reflection itself would have to give rise to ethical knowledge, and for this to be possible, there must

be some sense given to the idea of what it is to track the truth of such judgments. We should expect,

on Williams’ account, that rational deliberators should arrive at some convergence on ethical

judgments in this objective sense, but none seems to be forthcoming (Williams 1985, 151-2).

Williams’ argument relies upon substantive views in epistemology, concerning the

importance of convergence to ethical or philosophical knowledge and the importance of tests that

“track truth” that have no direct relevance to the argument presented here. I have presented an

account of philosophical meaning that relies upon convergence among experts, over time in an

institution, to render translations determinate. But in rendering translations of philosophical texts

determinate we are not primarily pronouncing on the truth or falsity of substantive issues. If

successful, our institution would not interfere with the plausibility of the major going options in

philosophy. Thus, convergence is not explicitly built into an account of philosophical concepts on

350
QI’s account. Objectivity is certainly one aspiration of philosophical concepts, according to QI, but

this is not obviously the same as convergence. Bracketing such controversies over convergence, the

rest of William’s account of the threat of reflection to ethical knowledge is untenable.

Thick concepts, for QI, are not a combination of an empirical content and a valuing

component. Thin concepts are theoretical. Their characters present criteria for usage that could in

theory form the foundation for convergence in philosophical judgments if this is desirable. Thick

concepts, in contrast, simply have a narrow range of theoretical employment. If we give up on our

thick concepts we would be able to make logically equivalent judgments using thin concepts and

would feel none the more sheepish for it, for reflection on the semantic foundation of philosophical

concepts that permits translation shows us that our thick judgments, to the extent that they had any

determinate, philosophical content at all, were made with some theoretical assumptions to begin with

that do not hang on the thickness of concepts.

The move thus, for QI, is not from a society that is traditional to one that is reflective, but one

that is unphilosophical to one that is philosophical. The unphilosophical society has no value

concepts, whereas the philosophical society does. The philosophical society is thus the society in

which diversity in philosophical perspective is a sociological fact that explains the diversity of its

philosophical term usage. It is not necessarily reflected in the society’s ability to present to the world

a corpus of philosophical texts (though usually, they will).

If philosophical knowledge is possible on QI’s account (and I have no reason to believe it is

not), it is possible because we can at times make judgments employing philosophical concepts

according to the relevant considerations. Having philosophical knowledge is thus greatly aided by

being a participant in a philosophical tradition for the tradition will help one sort through

considerations of relevance. But we might get things right for purely external reasons some times, and

if being right is more important than being justified, we may wish to count such successful

philosophical judgments as instancing knowledge. Sorting these issues out takes us well beyond the

351
scope of this dissertation. However, suffice to say that if QI and TTS are correct, Williams’ story

about reflection destroying ethical knowledge is not supported by the best account of the thick-thin

concept distinction. This is because a fair portion of his argument against objective ethical

knowledge depends upon the view that thick terms are an amalgam of empirical and evaluative

content. According to QI, this is the wrong analysis of thick concepts.

Finally, we might wonder whether it even makes sense to contemplate the existence of a

society that is prephilosophical. Can there be a society that has no philosophical concepts but only

subjective-description concepts or purely descriptive concepts? Is there a society, for instance, that

has a purely descriptive term for organization, but no inclination to talk of good organization or just

organization but yet will be seen to make claims like “I like this organization”?

The question being posed here is similar to the question that scholars of African philosophy

often wrestle with: is philosophy universal? Or, put another way, what is it for a society to have

philosophy? A bad way of determining an answer to this question is to look to the language of a

culture to determine whether it has a word that corresponds to “philosophy.” This is a bad way of

settling the issue after the manner of attempting to understand what “ethics” means by asking people

what they mean by “ethics.” This is bad for it presumes that a determinate answer to these questions

will be linguistic in nature. As we have seen, linguistic accounts of meaning are quite incapable of

explaining how translation is possible.

The answer that TTS offers to this question begins with the clarification that whether a

society has philosophical concepts is not something that can be determined through linguistic

research. The question can only be settled textually. If we attempt to determine this question while

looking at language, considered abstractly, we are in for grief. This is exactly what scholars of Indian

philosophy have done for generations. If one looks to dictionary entries on “dharma,” one will find a

dutifully compiled list of designata that the term “dharma” has been used for in Sanskrit. What does

this settle? Indeed, if foreign scholars were to attempt to swoop down and determine whether we in

352
English have value terminology, they may end up listing, in their lexicon, a list of items called “good”

with no criterial indication of its cross-contextual, cotextual functioning. Such definitions are not

wrong so much as they are translationally indeterminate. Theorists such as Hare and Dreier have

speculated that we could bridge the gap by attempting to look for terms in the foreign language with

the same commending function or character as “good” in English. However, it was never clear on

these proposals how we would know that we have hit upon a term that had the same character as

“good.” Philosophical concepts, contra Dreier, are not context-sensitive in the manner that ordinary

indexicals are (cf. VII.2.2 TTS and Cappelen and Lepore’s Tests 324-327). Thus, while we may have

putative linguistic tests that separate out basic lists of pronouns and pointing words in a language

(which is doubtful if languages can contain monsters), there seems to be no such test for value

terminology. This is because speakers are likely not to concede the context-sensitivity of such terms

in their society: far too much hangs on the good, the right, the moral, the real, and knowledge for

people to make such claims about the meaning of their shared linguistic resources. And indeed, at the

level of substantive philosophical debate, our goal is to make the case that our views meet objective

criteria of relevance. Philosophy mandates this appeal to objectivity that opens up context and

cotextual sensitivity at the level of texts, and shuts it at the level of substantive philosophical

argument.

Thus, the question of whether a society has philosophical concepts is nothing more than the

question of whether a society has texts that require the philosophical text-type for translation. An

answer to this question, I believe, has to do with our willingness to philosophically engage with the

semantic artefacts of a society. We can, after all, translate the same text according to multiple text-

types, according to TTS, and in each case we will be preserving some distinct semantic modality from

the ST in our TT. Our success in being able to translate a text from a given society as philosophy

would thus be the platform from which we would be able to read back, into the language of a society,

philosophical terms as instantiating the text-type features of philosophy according to QI.

353
354
VIII. Conclusion

VIII.1. Summary of Argument for QI

Evaluative/Normative discourse as defined at the outset of this dissertation is a discourse

characterized by the central role of nominal markers such as “right” or “good and verbal markers such

as “ought.” My thesis that I set out to defend is that we require what I call the “Quasi-Indexical

account of thick and thin” concepts to translate evaluative discourse.

Translation seems like an unworkable challenge under the prevailing linguistic paradigm.

According to this paradigm, determinate meaning is a feature of languages (whether such meanings

just is the use of symbols in a practice, its metaphysical referents or associated concepts expressed by

such terms in usage) and translation is thought of as an attempt to match symbols and expressions

across linguistic practices on the basis of cross-linguistic synonymy. However, the historical reality of

the distinct histories and cultural conditions for the development of languages and the semantic

profiles of expressions in a language make a complete semantic equivalence across languages

unintelligible. Rather, if translation is possible at all, it is possible because the aim of translation is to

preserve some, and not all, semantic features of a text in translation. The deciding factors in what

features of a text are preserved in translation are text-types. Text-types identify specific features of a

text (or text-type features) to be preserved in translation. Translation is successful when those features

of a ST identified by a text-type are preserved in translation.

Normative or evaluative discourse is discourse characterized by normative and evaluative

concepts. A successful translation of normative discourse must thus preserve normative and

evaluative concepts, such as GOOD and MORAL, in translation, not contingently, but necessarily. In

other words, if we want to know how to translate normative discourse, we need to identify the text-

type of which concepts such as GOOD and ETHICAL are the text-type features. To recognize these as

355
text-type features of philosophy and simultaneously as evaluative and normative concepts is to

recognize these concepts as the means by which universal and general theories with a world-to-theory

direction of fit presented in light of relevant considerations that are thought to be objectively

persuasive structure the semantic unity of a text. This theory-articulating function is quasi-

indexicality. To privilege such concepts in the translation of a text is to understand the meaning of a

text in light of its normative and evaluative theory-articulating function, and thus we understand the

text to be preserved in translation in light of this philosophical function.

We can contrast the process of philosophical translation with the translation of a text of

science. As noted in chapter 5, we could translate a text of ethical interest as a text of science, but that

would be to treat it as a type of descriptive psychology or sociology, and translate it not in light of the

context-independent recommendations we could understand the text as making, but rather as an

empirical study of the application of culturally specific concepts in order to discern whether it tracks

any properties of local interest in the context of the SL culture The notion that what we have in

translation makes any recommendations to the reader on how the contingencies of their world and

cultural context is to be judged is left out. Rather, we understand the text as making claims about a

theory that is to be judged in light of the contingencies of the world as delineated by the cultural

context of the SL author. What we are left with is simply an empirical study of term use in a society.

The content is spectacularly unnormative and unevaluative for the discourse does not provide theories

and reasons that the reader is supposed to take as applicable to their context. The theories and reasons,

to the extent that there are any in such a translation, apply only to the context of the SL persons as

descriptions of properties tracked by their linguistic practice.

Science can track matters that are not culturally peculiar. Indeed it can, and does, in the case

of physics, chemistry and biology. The scientists’ theory here plays a role in conceptualizing the

theoretical entities and these theories are not culturally bound in the way that eudaimonia in Greek is.

If we decide to treat such terms of ethical interest as scientific terms, then we are left with looking to

356
the contingencies of a linguistic practice (ancient Greek, for instance) to help us understand what

properties they may be tracking, and we have no a priori reason to think that any other culture would

track such properties with their cultural practices or that it could even be of interest to them. This

contrasts sharply with the scientific interest in genuine natural kind terms of a language. In this case,

the natural kind terms may be originally present in only some languages (say because the species it

refers to is local), but the property that it tracks is not simply constituted by a cultural practice, but

rather is a part of the common world that we all live in.

Philosophical translation, in contrast, preserves theories and reasons that are universal and

general in nature, and thus the content to be preserved in the translation of a text of philosophy

includes theories and reasons that are of relevance to readers in any context, and not simply the

context of the SL speakers. Thus, in saving the normative and evaluative theories in philosophical

translation or their putatively objective relevant considerations, philosophical translation saves

considerations that are relevant to any context, and not simply the context of the SL speakers. Thus,

in a philosophical translation of Aristotle’s works, we treat “eudaimonia” not primarily as a term that

tracks a property, but as a quasi-indexical, world-to-theory, theory and reason articulator, and we can

understand Aristotle as making normative claims about what anyone should think is happiness, if

“happiness” can also play such a theory-articulating role in the translation. To read Aristotle’s texts

thus is to read him as making claims to be taken seriously in any cultural context.

QI can show us how to translate normative and evaluative discourse because it provides the

anatomy of the very structure of the philosophical text, and in its various incarnations, it provides the

structure of the philosophical text articulating theories of a world-to-theory direction of fit, for various

axiological differentia.

Expressivism, as we saw, might seem like a contender to account for normative and

evaluative discourse translation. But it fails by virtue of its contempt for philosophy. Expressivism

regards the philosophical reasons agents have for their axiological convictions as inessential to their

357
semantics, and thus when pressed into the role of saving content for normative and evaluative

discourse translation, it leaves out the reasons that persons have for their convictions. The result is a

garble that does not accord with any realistic expectation of an accurate translation of normative and

evaluative discourse.

QI is an account of the text-type of philosophy, and the theory that explains the semantic

significance of text-types is TTS. TTS avoids traditional problems that have plagued translation

theory by recognizing translation to be a process that concerns, primarily, the semantic equivalence of

texts relative to types. Traditional theories of translation in the linguistic paradigm typically assume

that in order for translation to be successful, one requires cross-linguistic synonymy. Translation is

conceived on this account as a word for word, sentence for sentence exchange across language. The

linguistic account of translation is like a theory of medicine that has nothing to offer the unwell. It is

a serious failure as a theory of translation in exactly the same way that a theory of medicine that has

nothing to offer the unwell is a serious failure as a theory of medicine.

TTS is not a failure as a theory of translation because it does not assume that successful

translation depends upon cross-linguistic synonymy. To expect that successful translation requires

cross-linguistic synonymy is to expect that translation must not have to contend with radical cultural

diversity. But there is no point of a theory of translation except to mediate cultural diversity. The

linguistic paradigm is flatly not up to the challenge, for it attempts to ground the objectivity of

meaning in what is culturally contingent, namely language. TTS, in contrast, recognizes languages for

what they are—culturally evolving phenomena that are semantically indeterminate and it recognizes

the objectivity of our best epistemic claims as grounded not on language as such, but on text-types. In

deprivileging the role that cultural contingencies play in meaning, we can understand how

determinate translation is possible. It is possible through texts of definite types, and in the process of

facilitating translation as a textual process, translation itself contributes to cultural change. But like all

factors in a culture, its effect is not single-handedly determinate.

358
 QI is able to facilitate the translation of normative discourse because it does not expect texts

to be semantically simple or flat. Rather, it provides the resources for the translator to save the

relevant features of a ST in a TT that would allow us to recognize the TT as bearing normative

discourse and semantically equivalent to the ST. Because of the selective role of all text-types in

translation, they can at once play the role as guarantees of cross-cultural communication, but also,

they can play the role as standards against which competing claims relative to a text-type can be

observed. In chapter 3 I argued that the text-type of science can help us understand how Darwin’s

Origin of the Species is a superior account of biological life to the Book of Genesis in light of

institutional expectations that are reducible to the very text-type of science itself. I think that the text-

type of philosophy provides us with the very same ability, to wade through competing ethical and

philosophical claims. It forms the foundation for what I shall call a philosophical realism. Before I

explain how this is, I shall respond to two concerns. The first concern is that the account I have

provided cannot claim to determine translation in a culturally and linguistically neutral fashion. The

second objection is that the liberalization of knowledge that comes about on TTS’s account is

unacceptable.

VIII.2. History, Culture and Objectivity

I have provided an account of translation and the meaning of texts that claims to be a-cultural

and a-linguistic. Substantively this is a desideratum of an account of meaning for it can provide us a

way to think about philosophical problems that is not culture bound or relativistic. The way to

understanding how determinate meaning transcends culture and language on my account is made

possible by two factors: first the indeterminacy and polysemy of semiotic systems, especially

languages, and correlatively the text-type-theoretic nature of translation as a selective process of

semantic content preservation. It is the text-type that is our Archimedean point from which we can

understand the objectivity of meaning, not language, and the text-type allows us this because of its

relative lack of ambition. It isn’t concerned with everything that is meaningful, just somethings. We

359
can make a mess of things still if we are not self-conscious about the semantic norms of a text-type

and if we do not consistently apply them in translation, but this is an easily correctable error, defined

by its deviance from the fundamental semantic norms of a text-type.

I do not know of any philosopher who has proposed this solution to the problem of the

cultural transcendence of meaning, but there are many who will balk at the claim that translation and

meaning can transcend culture and context.

Hans-Georg Gadamer, for instance, argues at length against the movement he identified as

“Historicism” prevalent in the 19th century in Europe. According to Historicism, previous ages were

not aware of the role that historical context played in their thinking, but given an awareness of such

presuppositions, and the fact of cognitive change over time, the historically aware age can overcome

such prejudices and have an objective appreciation of meaning (Gadamer 1996 [1960], 197-242). A

major problem with historicism, according to Gadamer, is that it naively does not factor its own

historicity into its self-conception: it assumes a position outside of history, when it too is a product of

historical forces (Gadamer 1996 [1960], 299-300). Another problem with historicism is that it

assumes that “interpretation is related to its object as an accident is related to an underlying

substance” (Odenstedt 2003; cf. Gadamer 1996 [1960], 484). Gadamer’s view is that textual

interpretation is not different in kind from the experience of art (Gadamer 1996 [1960], 163). In both

cases there is a dissolution of the boundaries between the work and the experiencer.

In view of the finitude of our historical existence, it would seem that there is something absurd about
the whole idea of a unique, correct interpretation. … Here the obvious fact that every interpretation
tries to be correct serves only to confirm that the non-differentiation of the mediation (Vermittlung)
from the work itself is the actual experience of the work. This accords with the fact that aesthetic
consciousness is generally able to make the aesthetic distinction between the work and its mediation
only in a critical way-i.e., where the interpretation breaks down. The mediation that communicates
the work is, in principle, total.(Gadamer 1996 [1960], 120)

360
 Gadamer takes from Heidegger the notion that interpretation is always situated. He writes,

“The standpoint that is beyond any standpoint, a standpoint from which we could conceive its true

identity, is a pure illusion” (Gadamer 1996 [1960], 376 ) According to Heidegger, all interpretation is

a type of projection (Gadamer 1996 [1960], 266-67; Heidegger 1962). What we do when we interpret

a text, following this Heideggerian thread, is we mediate our interpretation of a text in light of our

tradition (Gadamer 1996 [1960], 293). Interpretation, on Gadamer’s account, is a “fusion of horizons”

where a text and a speaker meld in a unified experience (Gadamer 1996 [1960], 378), just as the

distinction between the viewer and a work of art meld into a single experience. Translation, on his

account, is just a type of interpretation (Gadamer 1996 [1960], 384-6). Oddly, Gadamer holds that all

understanding involves prejudice. This is not simply a negative notion. Rather, prejudice is “a

judgment that is rendered before all the elements that determine a situation have been finally

examined” (Gadamer 1996 [1960], 270). But this emphasis on prejudice may simply be in keeping

with his emphasis on the importance of projection in understanding. The goal of this prejudice

however is a type of interpretive charity: “When we try to understand a text, we do not try to

transpose ourselves into the author's mind but, if one wants to use this terminology, we try to

transpose ourselves into the perspective within which he has formed his views. But this simply means

that we try to understand how what he is saying could be right” (Gadamer 1996 [1960], 292).

“Texts are ‘enduringly fixed expressions of life’ that are to be understood; and that means

that one partner in the hermeneutical conversation, the text, speaks only through the other partner, the

interpreter. Only through him are the written marks changed back into meaning”(Gadamer 1996

[1960], 387). But this is only possible within a tradition.

While Gadamer holds that all of humanity is involved in one conversation (Gadamer 1990,

13), he paradoxically also holds that the type of cross-cultural translation and research that I urge a

textual approach to meaning can help us accomplish is a mistake:

361
 If I were compelled to classify the ways in which other cultures were transmitted, and to say
something about Lao Tsu, the great Chinese sage, then I would need to ask: Is that religious
literature? Or, is it philosophic literature? Or is it poetry? But I would not know the answer, and not
simply because I do not know the answer, but because it is in the history of our Western civilization
and culture that these three forms of discourse and of language have emerged out of one another, and
have done so in the mode of a continual exchange with one another. (Gadamer 1990, 20)

Gadamer here articulates an unease with cross-cultural scholarship that appears to have been
151
influential among some in the Heideggerian tradition.

Given these Gadamerian insights, could it be said that view that the I am advocating is simply

a new form of historicism, which naively claims to transcend my own historical and cultural context?

Am I incorrectly holding that interpretations (in which Gadamer also includes translation—Gadamer

1996 [1960], 384) is a timeless accident of a text that is the substance? Am I attempting to deny the

role of prejudice in interpretation and translation?

TTS is not a form of historicism as it recognizes the role of history in sharpening text-types,

which are indispensable to translation. But text-types by virtue of their narrow and selective nature

are not concerned with all the semantic modalities of a text. Thus, the fact that the institution of

philosophy as we know it today may have its roots in Greece does not imply that the type of text that

we now understand under the heading of philosophy must be solely of a Western provenance. We can

recognize fully that other cultures will have had their own institutions and social processes, but it is

consistent with this that such local institutions can give rise to texts that are profitably treated as texts

of philosophy.

Text-type institutions are a type of tradition, but they are not coextensive with a culture as

such. Their concern with a narrow range of phenomena renders them too narrow. As argued in

chapter 3, a text-type cannot be thought to reduce to the literal meanings of terms in a language for its

concern is far more specialized and non-systematic than the nature of meanings of terms in a

language. Languages may be evolving systems, but text-types and text-type institutions are not. Text-

362
types are semantic portals for the conversion of semiotic phenomena according to definite criteria and

accuracy is assured when we diligently apply text-types in translation in a manner that comports with

our expectations. If we were to translate a text of poetry as though it were science, and yet call the

result a poem, we would probably be incorrect.

Thus, while text-types have a historical and traditional basis, they are not culture- or

language-bound. Because of their selectivity and formality, it is possible that one text could be both

poetry and philosophy, for instance. I suspect that Lao Tzu’s Tao Te Ching is one such text.

Thus, Gadamer’s analysis fails in part because he appears to collapse tradition with culture at

large. If a tradition were just a culture, we might think that we would not be able to understand what

type of text Lao Tzu wrote because we would not have convenient cultural slots for his text. But when

the question is converted into one of scholarly text-type institutions, the question is converted from

one of whether Lao Tzu’s text corresponds to anything we are familiar with in our cultural environs,

to one of whether his works can be translated according to specific text-types.

But the very fact that we can take a more specialized approach to tracking the question of

what type of text Lao Tzu’s Tao Te Ching is, shows that whatever traditional slots we thought a

culture furnishes us with are in and of themselves quite indeterminate. Even if it were true that in the

Western tradition philosophy, poetry and religion constituted distinct discourses, the fact that our text-

type-theoretic criteria can recognize one text as instantiating more than one type shows that these

cultural distinctions are foggy, or of indeterminate significance.

What then can we make of prejudice? Prejudice as Gadamer understands it has no obvious

role in interpretation or translation. It may have a role in bad interpretation and bad translation, but

the notion that we make a judgment in advance of all the details about a text in order to translate or

interpret it is wrong as a generalization. We cannot come to know whether a text is amenable to being

translated as a text of philosophy without first sizing it up. In cross-cultural research, what we

typically do is bring multiple text-types to bear and try them on for size. Some will work, others will

363
not. One cannot get much science out of Waiting for Godot. We can, as we saw in chapter 3, treat a

text like the Book of Genesis as a text of science, but the results will be very unflattering for the Book

of Genesis as an account of the origin of life.

Of course, in order to get cross-cultural research off the ground, scholars may have to get

dirty in the task of radically interpreting cultural practices in order to come to a working knowledge

of a language and cultural practices. But according to the account that I have pressed, on which a text-

type-theoretic account of translation rests, linguistic meaning is indeterminate to begin with. The

meanings of words and sentences that we might identify never fully account for their full range of

possibilities. Moreover, the more contradictory and conflicting our accounts of the meaning of

expressions in a language are, the better our translations will likely be, for in translating we are

always attempting to save some features of a text for preservation in translation, and thus it is only

against the backdrop of a rich appreciation for the polysemy of a semiotic system that we can come to

have good translations.

This initial stage of radical interpretation might involve a type of prejudice. For instance, at

this point, we may have to employ a Quine-Davidson style principle of charity in order to make sense

of the linguistic behaviour we are viewing, and to apply such a principle is to choose the

interpretation of the linguistic behaviour that makes the most amount of our subjects’ sentences

true—on our own lights (Wilson 1959, 532; Quine 1960, 58-9; Davidson 1996 [1974], 463, 2001

[1973], 137). However, if the arguments I have presented in chapters 2 and 3 are correct, this is not

what real translation is. It may be a form of interpretation, but it is not semantically determinate—

even on Quine and Davidson’s lights. Rather, semantic determinacy comes about once we have

completed the initial stage understanding the very rich indeterminacy/polysemy of a language or

semiotic system. Semantic determinacy thus occurs when we correctly employ a text-type in

translation in a manner that comports with our expectations. Coming to a rich appreciation of a

language’s polysemy takes time, and it will likely take several efforts at translation, over decades if

364
not centuries, to begin to get a handle on such polysemy in the case of very old, classical languages

with a large literary body. But the real translations, in such cases, would not involve prejudice as such

for we would not have to employ a principle of charity in translating a text. Rather, we employ the

relevant text-type and this significantly constrains how we can render a text in translation. Once it is

apparent that a text has enough relevant features that it can be translated as a text of moral

philosophy, the decision to render such a translation obliges the translator to understand the text in

light of the theories and reasons articulated by key philosophical terms constituting the text, and the

results can be very surprising. Here, the text-type provides us with a constraint in translation and thus

liberates us from having to rely upon our intuitions to determine meaning. Prejudice has very little

role to play.

Thus, while we might start out our quest for understanding with prejudices, the culmination

of research in excellent translations overcomes prejudice, and thus the notion that understanding

always involves prejudice is wrong.

The heavy emphasis on prejudice in Gadamer’s account, I think, is a result of his conflation

of the importance of a tradition with a culture. We can have a tradition of studying and translating

certain types of texts, but such a tradition can be international, as it is: philosophy and science are

international pursuits undertaken in various regional languages, by a loose network of scholars who

comprise textual institutions. These are not identical with the cultural origin of any of these scholars,

or with the historical birthplace of such institutions. Indeed, in the case of many textual institutions, it

is probably difficult to name just one cultural origin. Even in the case of philosophy, which we

understand as stretching back to Greece, the fact that we can with such success translate and study

non-Western texts under the rubric of philosophy as we understand it today suggests that what we are

tracking is not something that is culturally peculiar.

Recognizing textual institutions and text-types allows us thus to affirm the importance of

traditions in translating texts of importance, but it helps us criticize the notion that cultures set the

365
conditions of semantic determinacy of their artefacts. Culture is a confluence of factors and is

typically itself a dynamic process and the notion that it sets the conditions of semantic determinacy

for anything is hard to make sense of. Texts and their types do not change very quickly and are inert,

and the international character of text-type institutions provides a type of buoyed anchor for cultures

which are in themselves speeding processes of change.

To recognize text-types is thus to recognize meaning that is textual, and not linguistic. Had

Gadamer recognized the text-type as an important semantic factor in translation, he would likely not

have concluded that translation is simply a form of interpretation and he would likely have not placed

such an emphasis on the identity of a tradition with culture. And, he likely would not have put so

much stock on prejudice either. The notion that cultures are unified traditions that are not themselves

constituted by constant influences from the outside is a result of prejudice. The reality is that there has

almost never been the type of unified cultural traditions or languages that philosophers in the

linguistic paradigm theorize about, and this is because the shared history of humanity has been a

history in large part mediated by translation.

In closing my response to Gadamer, I stress that TTS shows us how we can have culturally

neutral and linguistically neutral perches to determine translation because it deprivileges the role of

culture in meaning. If one thought that culture was the foundation of semantic determinacy, then

indeed it would be very difficult to understand how any translation or platform for assessing

epistemic claims could transcend culture and language for the totalizing and systematizing nature of

language and culture would provide one no escape. To make a clear and intelligible claim would be to

make one that is clear and intelligible to a culture at large. But if we criticize culture and language as

an indeterminate semantic process of change and flux, then our objective is not to necessarily please

language and culture at large. Indeed, in many cases, accurate translation will be quite inaccessible to

members of a culture a large, and it will in all likelihood flout local convictions. Indeed, this is the

whole value of philosophy as an international discipline, concerned with texts from disparate cultural

366
sources. Through studying philosophy the world over, we recognize that the various conventions and

convictions of philosophers themselves are often products of cultural contingencies and we come to

learn this because we subject their texts to translation. Thus, translating Lao Tzu teaches us that the

division of poetry and philosophy, if there is really such a division in the West, is pure cultural

peculiarity, and that the Tao Te Ching is philosophy with a strong poetic quality. (Gadamer thus did

not follow through the modus tollens, so to speak.) If the local was the seat of semantic determinacy,

this would not be possible. But if the local and cultural is a process of change and evolution, this is no

problem at all.

What allows us to facilitate such dynamism are texts of definite types, and trained experts

whose institutional identity transcends their native cultural and linguistic identities. Indeed, we may

not always have the healthiest of institutions in these respects. I’m sure that philosophy, the text-type

institution I am most fond of, could do with a good deal of shaking up. We too often relegate the

translation of our texts to linguists and philologists who insufficiently understand philosophy as a

type of text. We also have many philosophers in our midst who live strange divided lives: when

educating their students, they emphasize the text-type of philosophy as the major institutional fact

they are to learn, but in their own research, the show no consciousness of it. There is room for

improvement. But there almost always is.

VIII.3. Objectivity and Epistemic Pluralism

One might attempt to argue that TTS is simply another form of relativism for it makes

determinate meaning relative to text-type institutions. Indeed, text-types are relative to institutions

only because we individuate text-type institutions by reference to their text-type-theoretic concern,

but so understood there is no conflict between various text-type institutions, for each seizes upon a

distinct constellation of semantic modalities in a text. There is thus no conflict to arise between the

translations that issue from these concerns. Indeed, these institutions are in fact at many points

complementary, as noted: texts will often be comprised of differing orders of types and ruling types

367
might mandate and trigger alternative text-type-theoretic readings of subsidiary sections of a text in

light of the polysemy of the text being translated.

One worry that I have not addressed at all is that the view of knowledge that TTS supports is

too liberal. If knowledge, determinate knowledge, is really textual, grounded in an institution, then it

seems that there can be any number of different textual institutions all claiming to yield knowledge.

Yet, this would result in absurdity. What we need is really one master, policing institution to decide

what institutions count as yielding knowledge. This should be science, or perhaps scientistic

philosophy. Or else, all manner of crazy institutions could claim to have knowledge. For instance,

astrology might just as well be a fountain of knowledge. We can’t have that!

My short answer is that this worry is unfounded, for each text-type provides the conditions

according to which its claims to knowledge must be judged. So understood, we do not need a master

police epistemic institution to give out licences to text-type institutions (we might do so for policy

reasons, but not for epistemic reasons).

Let us clarify what types of institutions can form the foundation for objective knowledge.

They must be defined by a text-type, but a text-type abstracts from authorship, and content. Thus, for

instance, the Catholic Church, I think, couldn’t claim to be a text-type institution even though it

claims to uniquely be able to interpret the Bible because it is concerned with one book (the Bible)

with a specific author (God): there is no way to determinately translate any text that we cannot treat

as simply one among many by way of its subsumption under a type. (The Christian claim to the

specialness of the Bible, if true, renders the text quite semantically indeterminate, which is to say, one

can get anything one wants out of it.) Rather, an institution that claims to have knowledge must have

a type, which is to say that it abstracts from content and authorship. The notion of a religious text, I

think, is largely left in the lurch by TTS as a nonsensical notion, and rightly so. Texts that are

recognized as important for religious reasons have nothing obviously in common from a semantic and

translational perspective: the texts that are ladled as “religious” in Asia are, with few exceptions, texts

368
of philosophy, concerned with providing reasons and arguments for universal and general theories on

all manner of philosophical interest (many of which are atheistic), while the sacred texts of the

Abrahamic traditions are a mixture of histories, prophecies and commands (all concerned with a

God)—the notion that we have one type here, is absurd. Not surprising: what makes a text religious is

not how it must be translated, but the importance it is given by people: this is a pragmatic criterion

not a semantic criterion. If religious institutions are not text-type institutions (and they can’t be,

because they are not concerned with genuine types of texts, but specific texts of specific providences)

then it is quite clear that they cannot be fountains of knowledge. We can study them as sociological

phenomena, but they themselves are not the types of institutions that have any unique epistemic

validity.

Astrology, in contrast, qualifies as a text-type institution. First of all, it is clearly institutional:

it is international, and there are professional organizations that meet, train and provide certification

for astrologers. Interestingly, there are also differing schools of thought in Astrology (Western

astrology that does not take into account astronomical shifts, but is based upon an account of the

nature of the heavens that is now out of date due to the procession of the equinoxes; Indic astrology

that claims to track the procession of the equinoxes and understand the relationship of the stars in

light of astronomical shifts in the zodiac; then there are innovations in astrological theory about how

to read charts). Most importantly, astrology has a type of text that can be instantiated in multiple

semiotic systems that abstracts from authorship and provenance. The text is the astrological chart.

Here’s my chart, as per the Indic system:

369
 The chart on the right is a translation of the chart on the left (or vice versa): same chart, but

distinct semiotic systems. We could even exchange the abbreviations of the planet names for symbols

(such as “ ” “ ”“ ”), or their names in Sanskrit, represented with Devanagari (the script of

Sanskrit). The text-type that dictates the norms for content preservation in this text-type also, for

instance, provides us with the bases to evaluate that the moon occupies the sixth house in my chart,

and is in Scorpio, and is with Mars who rules the house that the moon tenants (Scorpio). More

astrological knowledge:

370
This is Einstein’s chart. Einstein and I have the same ascendant (Gemini) and both have our moon in

Scorpio as per the Indic system. (There are other interesting similarities: we were both late to talk as

children, we’re both interested in the relationship between multiple perspectives and objectivity, we

are both interested in things hidden and not obvious, we both have frizzy hair, we both play the violin,

we are both vegetarians, but he’s much smarter and has a better chart on traditional modes of

reading.) Should we recoil from thinking knowledge of astrological charts, and how to calculate them,

is real knowledge? I’m not sure why. It’s not as though these matters are the subject of whims. If a

scholar wanted to study astrology, for say, its sociological or religious importance (one cannot know

much about traditional South Asia without knowing something about Indic astrology), knowing how

to calculate such matters as a person’s chart is indispensable. Does recognizing that there is such a

thing as an astrological text-type imply that we must live our lives around the predictions of an

astrologer?

This is where the role of the text-type in knowledge of an institution is crucial. If the

astrologer were to have some special knowledge about our future, this must be closely related to the

text-type features that her text-type identifies. From what I can see, the text-type for charts tracks the

371
placement of heavenly bodies relative to putative zodiacs in the sky, which is largely an empirical

matter. But it doesn’t say anything obvious about the signification of such bodies for our lives. Of

course, translation is always a combination of text-types and the polysemy of semiotic systems and

the notion that the details of the significations of various planets and their placement in a chart is a

matter to be gleaned from a broader appreciation of the polysemy of the various astrological bodies is

not unreasonable. However, the text-type of astrology is so apparently slim that it does not appear to

provide criteria according to which we could wade through the cacophony of significations that one

can find associated with any planet or heavenly body in astrological lore, particularly in light of the

various multiple relational significances that planets in astrological charts take on in combination.

Thus, translating an astrological text into alternative systems of representation, and understanding the

basic norms of such a translation, is quite insufficient to provide any type of criterion that could

license a prediction for a person in light of their particular chart. From what I have seen, the

significations of planets, such as the “moon” in relationship to various houses, zodiac signs and

asterisms is so great that astrologers really just start picking and choosing significations as it is

convenient for them in light of what they already know about a client who is seeking a reading. The

astrological reading thus ends up being, often, nothing but cherry-picking significances out of the

polysemy of the client’s chart based upon intuitions that are pragmatically felicitous in the eyes of the

astrologer and client. But we shouldn’t be so harsh on astrologers. Deferring to intuitions is an

epidemic in many areas of knowledge, including philosophy. In each case, we have seen that linguists

and philosophers simply defer to their intuitions in light of the polysemy of terms in languages to

yield an account of the meaning or significance of some concept and they often take the pragmatic

felicity of such readings to be proof of success.

If objective knowledge is possible, it must be possible in light of the basic norms of

translation that govern the institution’s concern for such norms would provide the objective,

institution-wide criteria according to which knowledge can be had. In the absence of such norms, one

372
has no clear direction in epistemic matters and it is a free for all, chase of intuitions. Such a pursuit is

not meaningless, for intuitions often track the polysemy of semiotic systems and contribute to them,

but there is nothing determinate about such an exercise. If astrology were to yield astrological

knowledge that was predictive (and not simply a matter of nomenclature) then at the very minimum

the text-type concerned would have to provide a type of algorithm to calculate predictions. The very

text-type would thus be such an algorithm, and instead of astrological readings one would receive

something like an astrological translation that is the prediction. Interestingly, work is underway in

this direction (an internet search for astrological readings will turn up many computer generated

readings based upon birth time and place details)—of course, it remains to be seen if the results match

the expectations set up by astrologers.

VIII.4. Relativism: Castles in the Sky

In this section I will address the main argument presented by Steven Hales in his excellent

book Relativism and the Foundations of Philosophy. Hales in my view deserves special notice for a

few reasons. First, his book is extremely lucid. Secondly, his is not an argument that attempts to make

its case by appeals to intuitions. Third, Hales’ attention to the diversity of philosophical convictions

in the world at large is a pleasant corrective to the un-reflective chauvinism that is quite ubiquitous in

the recent literature. It is a must read, in my view.

Hales begins with the notion of a philosophical proposition. He admits his account of these

creatures is vague (Hales 2006, 123) but he notes that “Part of what is traditionally taken to separate

philosophical propositions from empirical ones is that the former and not the latter are either

metaphysically necessary or impossible” (Hales 2006, 21). The trouble is that philosophers are not the

only ones who are interested in philosophical propositions. Christians, shamans and philosophers all

stake their claim to differing philosophical propositions. The challenge is thus to find some way to

decide between the various claims of the Christian, the shaman and the philosopher, for it seems that

they prima facie come in conflict on a number of fronts. The tradition of “analytic rationalism” that

373
Hales brings to light favours reason as a means of adjudicating between conflicting philosophical

propositions. Reason, so employed, endeavours to achieve a wide reflective equilibrium among the

prima facie philosophical propositions that someone might hold.152

However, it so happens that the traditional Christian, the proponents of shamanic religious

practice in addition to the philosopher all use reason in exactly this same way. The trouble is that

reason clearly has a role to play in dialectically mediating a person’s philosophical convictions. But:

There is no metaperspective, or absolute cross-platform epistemic standard, to which we might

appeal in order to determine whether Christians are right that there are non-physical immortal souls,
or whether the majority of rationalist philosophers are right and there are no such things. Each belief
seems justified and supported by the basic belief-acquiring methods and subsequent ratiocinating
intrinsic to their respective perspective. (Hales 2006, 123)

The only difference between philosophers in the rationalist tradition, Christians and shamans is the

manner in which they acquire their beliefs—philosophers defer to intuitions, Christians to revelation,

and shamans to drug-induced experiences. But reason has the same role of systematizing the derived

convictions in all three cases.

Hales’s argument, while lucid and updated for the contemporary reader, represents arguments

of existentialists of yore such as Kierkegaard, who argue that reason at best can systematize brute

convictions (this, in contemporary parlance, is what is involved in bringing convictions into wide

reflective equilibrium), but it cannot adjudicate between them. This leaves three options on Hales’

reckoning:

(1) Scepticism: since we don't know which is the right way to gain beliefs about philosophical
propositions, we have no philosophical knowledge. It would be mere luck if we picked the method
that actually led to truth. (2) Nihilism: there are no philosophical propositions. Putatively
philosophical propositions are either non-cognitive, or, more likely, covertly empirical. (3)
Relativism: the truth of philosophical propositions is relative in some way to doxastic perspectives.
(Hales 2006, 5)

374
Hales deftly shows scepticism to be self-refuting, for he rightly argues that scepticism is itself a

philosophical proposition. Relativism, which is usually thought to be self-refuting, is reconfigured

into a relatively modest absolutist thesis on Hales’ account. The type of relativism that Hales wants to

defend is not conceptual, or semantic. Nor is it the bold claim that all truth is relative. He argues that

the relativist must accept some absolute claims, especially what he calls the axioms of relativistic

logic (Hales 2006, 118), which shows how it is that every philosophical proposition is true from some

perspective. The only serious contender for Nihilism that Hales considers is what he identifies as

Naturalism. According to this variety of Naturalism, the only truths are those of science: there are no

philosophical propositions. Hales devotes a quarter of his book to responding to Naturalism. He

concludes that the arguments for Naturalism pose no real threat to the existence of philosophical

propositions. He closes his book thus:

A philosophical edifice built on rational intuition appears so comprehensive, systematic, and closed
that it is hard to imagine how a comparable structure could be erected on different foundations. It is
high time that philosophers come to recognize that their own familiar foundations have no privilege
and their ways of knowing have no monopoly on the truth. Indeed, the reason for this is that there is
no single truth to be had, but many perspective-dependent truths-relativism arises from the very
foundations of philosophy. (Hales 2006, 185)

Hales is correct that philosophy is intimately concerned with perspectives and that there are

many such. QI explicitly identifies perspectives as theories that philosophers hold to buttress their

philosophical term usage. Indeed, we could even agree with Hales that perspectives, or philosophical

theories in our nomenclature, are “abstract intensional objects” (Hales 2006, 113) as they putatively

determine the extension of philosophical concepts. Where Hales’ argument falters is his claim that

there is no perspective-independent “absolute cross-platform epistemic standard” from which

perspectives can be evaluated (Hales 2006, 123).

How indeed can we make such a bold claim? Are we claiming that there is an Archimedean

point that is not a perspective that can adjudicate between contrasting and competing philosophical

375
perspectives? Indeed, that is the claim that is being made. How can such a bold claim be made?

Because we have done the work of systematically addressing questions of semantics and translation

that would allow us to recognize and translate philosophical claims regardless of their cultural or

linguistic origins, a feat sadly that has never been undertaken.

Hales explicitly does not want to present an argument for relativism from semantic

considerations. He says that his epistemic argument is orthogonal to semantic arguments for

relativism (cf. Hales 2006, 2). However, he needs some semantics to underwrite his argument: just

enough to tell us what these philosophical propositions are. He says that they make claims of

metaphysical necessity and impossibility, and his reason for this characterization is that philosophers

often attempt to find counterexamples as means of refuting a philosophical proposition. But he

recognizes that his account is vague (Hales 2006, 123). The problem is deeper: his account suffers

from what Cappelen and Lepore call semantic opportunism. Hales’ account of philosophical

propositions is not based upon any systematic considerations of meaning. Rather, Hales’

characterization of philosophical propositions is purely phenomenological and based upon observable

features of philosophical claims and philosophers’ reactions to them. We’ve seen what trouble such

phenomenological accounts of meaning get us into when we attempt to move to translation.

Linguistic accounts of the meaning of expressions are typically accounts of how they seem to us, and

these are quite semantically indeterminate.

But there is a deeper problem yet for the notion that the philosophical proposition can

underwrite an argument for relativism. It is not clear, based upon the description that Hales gives, that

we would know when we ran into a philosophical proposition in translation. This is the trouble with

thinking about semantic matters without giving due attention to challenges of translation. We tend to

think that the meaning of claims presents themselves unambiguously and all that is left for us to do is

to group such meanings under a collective name, such as “philosophical proposition”—as though

sentences in foreign languages come labelled with the tag “philosophical proposition.” We tend to

376
think, given our own facility in using language, that this semantic determinacy comes to light by pure

linguistic competence or that we could translate such propositions by pairing up words across

languages based upon their native semantics. But languages and texts are polysemous and this

polysemy comes to light when we attempt to account for the meaning of expressions in one language

in light of the terms and sentences in another. The inherent polysemy of semiotic phenomena thus

makes the question of whether and what counts as a philosophical proposition as something that

cannot be determined in the absence of text-type considerations that we bring to bear in translation. If

we can identify texts, or portions of texts, that are translatable as philosophical discourse, then we

would have set the ground work to abstract from such texts philosophical propositions.

If the argument for the translation of philosophy that I present is correct, then QI sets out the

nature of philosophical concepts in translation, and thus paradigm cases of philosophical propositions

would be those formed with key philosophical terms, whose semantics operates according to QI. In

other words, a full philosophical proposition would consist in the employment of a term such as

“real” as a quasi-indexical device of theory articulation, and the full proposition would have all the

relevant variables that QI specifies saturated (variables such as the axiological differentia, the theory

variable and the objectivity clause).

Philosophical propositions are thus not met with floating in the air, but rather they are beings

whose meaning consists in their role in philosophical texts as defined by text-type considerations.

Perspectives, on this understanding, are the theories that underwrite the employment of key

philosophical terms in a philosophical text. Philosophical propositions thus are intimately tied to

perspectives.

So far, so good. We have shown earlier that QI, the only systematic and viable account of

translating philosophy on the go (which satisfies the textual constraints of translation and the

particular requirements of translating philosophy, as defended in this dissertation and based upon my

own research in cross-cultural translation of philosophy), can account for philosophical propositions

377
and it recognizes the perspectival nature of the philosophical proposition that is essential to the

relativists’ argument. The theory variable identified by QI is the variable for the philosophical

perspective.

But the theory variable specified by QI is only part of what makes for a philosophical text.

The other very important, indispensable feature of QI is the objectivity clause, or the considerations

that are brought to bear in philosophical texts. These reasons serve two semantic functions. They have

a narrow or internal function, and a wide or external function. The narrow function serves to fill out

the particular philosophical orientation of an author, and moreover serves to highlight the reason-

giving aspects of philosophical texts as marked for preservation in translation. This is its narrow or

internal function and the translator of philosophy must certainly keep track of such reasons in a text

from this internal perspective. But such reasons in philosophical texts aspire to objectivity. Failing to

note that the reasons that authors have for their philosophical convictions are meant to be objectively

persuasive—i.e., independent of the perspective—will result in very strange translations that would

not even qualify as philosophical. We could, failing to recognize this wide or external aspect of the

objectivity clause, translate value terminology as though it were merely subjective description

concepts. The semantics of philosophy thus underwrites the expectation that philosophical

propositions are objectively true or false. Their truth and falsity cannot be understood as guaranteed

by the perspective, but for all. Failing to appreciate this will result in major failures of translation.

Perhaps there are minimalist uses of truth such that philosophical propositions asserted from a

perspective are true, but this type of truth falls considerably short of what the semantics of a

philosophical proposition claims for itself.

So we know that the translator must understand the reason-giving aspect of philosophical

texts as aiming at objectivity independent of perspective. But where is this cross-platform standard by

which we can evaluate the truth and falsity of such reasons? QI! It is the very semantics of the

philosophical text that comprises the external environment of a philosophical perspective and it is

378
from here that we can criticize perspectives—without appeals to intuitions. Philosophical propositions

that are vindicated according to such perspective-external criteria can be said to be objectively true in

the best way possible: independently of cultural or linguistic considerations. QI thus forms the

foundations for what I shall call philosophical realism. Moral realism, metaphysical realism,

epistemological realism, and so on, are contained within the possibility of philosophical realism. QI

unlike any other proposed manner of grounding philosophical claims can uniquely claim to be

objective for it has shown itself to be employable in the determinate translation of philosophical texts,

and a prerequisite of this accomplishment is that it have a semantic foundation that is not reducible to

any particular culture or language. If my argument for QI and TTS is sound, all of this has been

established.

It may seem rather odd that I am advancing QI as a type of cross-platform standard from

which we can evaluate philosophical perspectives for in its ability to translate philosophical texts

regardless of philosophical orientation, it seems that it has absolutely no substantive bite. This is a

mistake. QI is able to translate philosophical texts because it adopts an institutional approach to

semantics. But it is these same institutional standards that can be called upon in assessing the

adequacy of philosophical perspectives. In this respect, knowledge we gain from philosophy is no

different from knowledge we gain from science. In both cases, possible knowledge is closely

associated with and underwritten by the semantics of the particular text-type. The text-type of science

is concerned with the articulation of descriptive theories in light of broad considerations of empirical

confirmation. In the case of philosophical theories, we must look to the anatomy of QI to provide

criteria according to which we can assess philosophical perspectives.

Here it may seem that we have a problem: many of the features of normative concepts à la QI

are too thin to do any work. The axiological differentia of philosophical concepts is determined by

textual means, which entails that almost any theory that wants in gets in, and is represented by the

axiological differentia. I suggested that on institutional grounds, we should recognize the axiological

379
differentia of ethical theories as the criterion of a concern for social implications. But even the ethical

egoist gets into ethics under this account, for the ethical egoist too has a moral theory that she chooses

for its social implications. It just so happens that, on her lights, the best configuration is one where

everyone is out for themselves.

On its own, the axiological differentia may not be able to do much, but when we substitute in

another aspect of philosophical theories à la QI, it becomes relatively robust.

All philosophical theories are like pictures. Normative theories are held up as standards

against which the world is to be judged, while non-normative philosophical theories seek the approval

of the world. No philosophical theory is descriptive of the world, for philosophical theories are

universal and general in scope and thus abstract from contingencies, such as the empirical world. Yet,

they must at the very least be consistent with such empirical contingencies to the extent that their

axiological differentia is characterizable in some descriptive fashion. QI in specifying that

philosophical theories are characterized by the criteria of universality and generality provides us a

way we can assess whether any given philosophical theory lives up to its philosophical aspirations

relative to its axiological differentia. Thus, not only must it in some sense be consistent with certain

empirical facts, but it must go beyond these facts and universalize and generalize in a manner that is

distinctly philosophical.

Here’s an example of what I am talking about. Recall that I suggested that we understand the

axiological differentia of ethical theories in terms of a concern for social implications. If this were the

right descriptive generalization of the common denominator of theories articulated with “ethics” in

philosophy, we would rightly expect that ethical theories be universal and general in their grasp of

social phenomenon. If they arbitrarily left out, from their picture of the ideal state, beings that are

perfectly capable of being social, then such theories would fail in their universality and generality in

respect of the axiological differentia of ethical theories. Or, if such a theory truncates its depiction of

the social in a manner that artificially leaves out motives and considerations that are constitutive of

380
the social (say, for instance, impartial considerations or motives of self-sacrifice and benevolence),

then the picture will also fail in universality and generality, with respect to the axiological differentia.

In otherwords, ethical theories that are anthropocentric, or egoistic, would fail the test of the proposed

axiological differentia of ethical theories of social concern in light of the universality and generality

criterion contained in QI. Indeed, I suspect that most moral theories in the literature would fail such a

test: ants, bees, dogs, pigs and bears can be and are social, and the restricted focus of so much

contemporary moral theorizing to anthropocentric issues is rather arbitrary.

Certainly, much rides on the account of the axiological differentia that we use to evaluate

philosophical theories. I’m not sure how many plausible options there are. With respect to ethics, it

was parenthetically noted, that there was a wide literature on the topic of axiological differentia, but

the peculiar characteristic of this literature was that it did not address the question in an institutional

manner. Rather, it conspicuously appealed to intuitions, linguistic and cultural factors, such as

linguistic practice, or the ever-present “we” of linguistic philosophy, that defers to one’s greater

cultural background (cf. Frankena 1963, 1980a, 1980b, 1980c; Hare 1955). Very few authors in this

literature paid much attention to the breath of theories actually articulated in the history of philosophy

that are as a matter of institutional fact treated within the province of moral philosophy or ethics (cf.

Falk 1963; Quinn 1986). When we think about the range between deontic theories, virtue theories,

theories that emphasize the place of the deliberator as central to ethics (Egoism, Libertarianism, and

Kantianism for instance) and those that stress reasons that come from outside (Divine Command

theory, explicitly altruistic theories or metaphysically heavy theories such as Utilitarianism), the

range of options that can describe the common point of departure is very small. No account that

conspicuously made use of philosophical terminology qua QI (such as “goodness” and the like)

would help us in our task, for we are not only looking for a characterization that uniquely groups such

concerns, but also one that we could determinately use in translation. To make use of philosophical

terminology in this characterization would merely defer the question of axiological differentiation,

381
and if we insist upon understanding such terms circularly, it will make translation indeterminate, for

foreign language texts do not have our terms. We are ideally looking to define cross-corpus

synonymy of key philosophical terms in part on the basis of shared axiological differentia (of course,

terms must also have the same valence, and thickness) and thus conspicuously circular definitions

would only result in translational indeterminacy. At any rate, there is no need to make use of circular

definitions for we can just as easily provide a descriptive generalization of the commonality of

theories in philosophy articulated with key philosophical terms without employing such terms. The

result in all cases will be a very broad characterization, but such breath is quite in keeping with the

spirit of philosophy and we shouldn’t be too disheartened by it.

Given the right account of the axiological differentia of philosophical theories, QI is well on

its way to being able to provide the cross-theoretic basis from which we can criticize philosophical

perspectives. Once we cut away the vast number of moral theories that fail the initial test of QI that

pits the theory against the expectation that the theory be a truly universal and general theory of the

sort it axiologically aspires to be, we may be left with a batch of theories, which are neither egoistic

nor anthropocentric, for instance, but are yet perhaps incompatible. Does this mean that we arrive at

a narrower range of philosophical perspectives, each of which supports its own truth but for which

there is no common arbitrator? No. Each theory will have every other theory as an outside from

which its philosophical propositions can be assessed. An overlapping consensus of a pool of common

propositions can be identified—(cf. Rawls 1996, 133-172). These philosophical propositions will

perhaps be counted as propositional schemes that are alike in all respects except for the theory that

drives their functioning in a philosophical text. This refined stock of philosophical propositional

schemes will in turn be the foundations, or relevant considerations, that the objectively correct moral

theory will systematize. The theories that made it up to this point couldn’t be reified as the final

theory: we would be seeking a theory at this point that did not obviously affirm any of the

propositions that fell outside of this overlapping consensus. The entire process is guided by the

382
objectivist aspirations of the philosophical text-type, which makes the objective, perspective-

independent vindication of perspectives and their reasons the driving force behind its unique

semantics. In this final case, each theory that passes the initial test of the axiological differentia,

universalized and generalized, still appeals to considerations independent of its own perspective, and

the overlapping consensus between such theories marks out what perspective-independence can

amount to at this point.

I suspect this approach to philosophical knowledge will rankle some. First, it might be

claimed that the universality and generality criteria of philosophical theories, particularly in the case

of normative theories, is supposed to characterize the application of normative theories, thus a theory

like egoism is universal and general because it is proposed as applicable in all cases that a moral

decision is called for. The universality and generality criteria are not supposed to characterize the

actual perspective.

Secondly, one might object that the manner in which I am recommending the axiological

differentia be used to evaluate moral theories or philosophical theories is unfair. Almost no

philosopher or theory sets out to be just or primarily a theory about social matters, as the putative

axiological differentia holds. Consider the libertarian moralist who articulates an ethical theory that is

explicitly concerned with the rational, self-imposed constraints of the individual. Social implications

are somewhat downstream from the main concern of such a moral theory.

Both of these complaints stem from the notion that the account of philosophical realism I am

forwarding fails to take at face value how philosophical theories themselves aim to be evaluated, and

rather replaces such accounts with some putative institutional considerations that most philosophers

never sign up for. This complaint is natural, but unfounded. QI, unlike most accounts of philosophical

semantics, takes into full consideration the theories that philosophers themselves develop and

implement in their texts. It actually defers to their theories in an effort to understand their textual

artefacts and it defers to the full range of philosophical theories to determine the axiological criteria

383
of axiological differentia. Of course, QI as I have formulated it does not explicitly direct us to this

task, but it does specify that each philosophical perspective aims to be vindicated objectively, apart

from its narrow range of concerns. Thus, every contrary theory in the history of philosophy is the

“outside” of every given theory articulated relative to a specific philosophical term and thus the very

objectivity of the philosophical perspective as understood by QI leads us to assess the axiological

differentia of a key philosophical term by taking stock of the full range of theories articulated with

such a term. There is no other non-perspectival way to assess the axiological differentia of a

philosophical term. But this is not difficult for us to assess: we are after all concerned with something

institutional that has made it into the literature, so to speak: historians of philosophy are conversant

with such breath. Now recall that each philosophical perspective in effect signs up for such a

challenge, to be judged as objectively true. But there is a price to be borne for making philosophical

claims, namely that one does, in effect, sign up for having one’s perspective judged independently of

it. This is in effect what the objectivity clause aspires to and it is by virtue of the fact that the

philosophical text-type concerns the dialectical interplay between theories and reasons that aim to be

objectively persuasive, that we can have anything like a philosophical claim, be it ethical,

epistemological, metaphysical or logical. The only perspective-independent platform from which

philosophical claims can be vindicated is the institutional platform that guarantees their semantic

determinacy in translation. But such a cross-theoretic platform is not in the business of respecting the

wishes of each theory. Its only constraint is whether it can allow us to translate any philosophical text

and theory. QI is the only plausible semantics to underwrite the translation of philosophy. Viewed in

this manner, QI is ideally suited to be employed in the evaluation of philosophical theories and it is

thus rather natural that such criteria might be found to disqualify some theories as philosophical

failures, but that is simply the chance we take when we make philosophical claims. Thus, for

instance, that a theory such as Egoism does not wish to be judged on its very constrained perspective

but rather on its applicability is for Egoism to paradoxically wish at once to be justified independent

384
of itself but also on its own terms. A philosophical perspective can’t have it both ways, and if it can, it

is still always possible for us to assess it independently of the terms it wishes to be assessed by.

Hence objectivity in philosophy.

Another objection that might be lobbed at my response to relativism is to argue that QI and

TTS are just another philosophical perspective and thus it can no more claim to provide a perspective-

independent manner of assessing philosophical claims than any other. I think this is the wrong way to

understand QI and TTS. TTS explains how philosophy as a text-type is meaningful, and QI explains

how it is that we need to understand a philosophical perspective’s contribution to a philosophical text.

It is thus explicitly deferential to philosophical perspectives and does not obviously take positions on

substantive matters. I have argued that we can use QI to criticize philosophical perspectives, but it

only works as a criticism because philosophical perspectives sign up for perspective-independent

validation by being philosophical. If they did not, they would simply be a type of psychological text.

We miss this aspect of philosophical perspectives by thinking about meaning in terms of language,

words and propositions, but it is thrust upon us in the textual context for here we are forced to explain

why we should treat a text as a species of philosophy and not something else. Thus, the critical ability

of QI is generated from the very aspirations of independent philosophical perspectives and the

projects of philosophers. If there were more than one way to understand and translate the

philosophical text, then perhaps the perspectival nature of my proposal might count against it,

however, there are no other viable accounts of the semantics of philosophy that can explain how we

can translate philosophical claims. Moreover, it is important to recall that relativism, on Hales’

account, requires some absolutist claims in order to be coherent, and thus relativism itself is not a

doctrine that is made completely from nowhere. It just so happens that TTS and QI are superior for

they are the means by which we can identify and translate philosophical propositions and

perspectives, while relativism in all of its forms has no viable explanation of this task.

385
 QI might be unpopular and lead to counterintuitive results, but that, I think, should be taken

as evidence in part that it is on the right path. If all of our intuitions were vindicated in philosophy, it

would be hard to understand how we were getting at anything objective.

Here’s the protocol for philosophical knowledge based on QI. Test each theory against the

particular axiological differentia of a concept in light of the universality and generality criteria. Those

that fail to meet such standards (i.e., fair very poorly relative to the universalized and generalized

descriptive content of the axiological differentia) are to be discarded. If there is more than one theory

left standing, find the overlapping consensus among such theories to mark out a new stock of

philosophical propositional schemes, and attempt to construct a theory that systematizes them. Given

the universal and general push of philosophy, we ought to begin this task with the most universal and

general philosophical concepts that we have: thin concepts with axiological differentias. Thicker

concepts, with extra-theoretic constraints, will have to wait, and it may turn out that many thicker

concepts fail to be vindicated by this process.

The difficult concepts to adjudicate will no doubt be those with no axiological differentia:

right, wrong, good, etc. Perhaps once we have arrived at the objectively correct theory for all

normative concepts with axiological differentia, we could understand goodness as articulating a

theory that is synoptic of value as such. But this seems to be an incomplete project: what normative

concepts there are is in part a function of institutional artifice.

Could moral knowledge change? If the axiological differentia of moral and ethical concepts

changes, moral knowledge would change, but it is unclear what would prompt this. If it does occur,

we can only understand this as part of the very dialectical structure of philosophy that seeks to foster

debate by ever widening the discussion.

It may seem that the approach I am advocating might lead to a stock of prima facie

philosophical principles with no hope of systematization. This is a possibility, but one that reduces

when we recognize that there may be some wisdom in attempting to systematize the winning

386
propositional schemes for all thin philosophical concepts. If we are not getting anywhere with just

ethics, for instance, revising the project as one that is concerned with being systematic in philosophy,

institutionally defined, may be wise. The objectivity clause may push us in this direction.

There may be room for a residual type of moral particularism. We might find that having

arrived at our stock of moral philosophical propositional schemes, we are not making head-way in

systematizing them, and thus we may have to attempt to apply them on the fly in actual

circumstances. I suspect, even after we have properly systematized philosophy, we will be left with

this residual existential challenge. In part, the pressure to move to a type of flexible and

individualized application of such principles may arise from the fact that there are many normative

concepts: ethical considerations, though important, are not the only ones. If moral propositions are to

be overriding, we’ll need some external reason to think this: overridingness plays no part in the actual

semantics of moral discourse translation from what I can tell for the overridingness thesis is a feature

of some but not all moral theories. I suspect that the only pressure that could vindicate the notion that

moral convictions are overriding will be from metaphysics, if it could be shown that the social is

more foundational than other matters of normative interest.

After all of this, it may appear that more than one theory can systematize the final stock of

philosophical propositional schemes. If we are having trouble choosing just one theory to systematize

our stock of vindicated moral propositions, we would certainly do well to reconceive the effort as

having to systematize all normative propositions. If this should not yield a unique result, we can

revise the project as one of having to systematize all vindicated philosophical propositions. If this

should not yield a unique result, it may be that we are left with a type of philosophical particularism,

where philosophy yields a stock of vindicated philosophical propositions but no unique way to

systematize the propositions. This may not at all be such a bad finishing line for philosophy, for we

would still have philosophical knowledge that is objective. Our philosophical knowledge would be of

the form of propositional schemes, vindicated by the institution of philosophy. The list of vindicated

387
philosophical propositions would serve as a minimal standard by which all philosophical theories are

to be judged.

Could every philosophical question be resolved by this method? For instance, could we

decide whether it is the case that God exists by this approach? Indeed, I think so.

Typically, “God” is treated as a proper noun, not as a quasi-indexical theory articulator. It

could be understood as a key philosophical term though, and then we could attempt to arrive at the

philosophically objective or correct theory or desiderata of a theory of God in light of first securing an

account of the axiological differentia of theories about God, and then testing the various theories in

light of the breath of their conception as per the universality and generality criteria. I suspect that

most conceptions of God, particularly those that favour some persons over another, or some species

over others, would be chucked out at an early stage for lack of universality and generality in their

conception, but none of this would establish whether God exists. We would simply have arrived at the

best conception of God in light of philosophical considerations—a type of conceptual analysis of the

notion of God that did not appeal to fuzzy notions of intuition and the like. Whether God exists would

be a question settled by our effort to pin down the correct metaphysical theory, or at the very

minimum the final stock of objectively correct metaphysical propositional schemes. “God,”

understood as a proper name, stands for many items and it may be that God on some account exists

(for instance, Spinoza’s God probably does exist) but this will not necessarily please everyone and the

results might not match up with our philosophically refined conception of God.

Could we make mistakes in our attempt to answer such questions? Indeed we can, and thus

there is an additional dose of realism injected into philosophy for it is always possible for us as

officers of a text-type institution to stray from its fundamental norms. Thus, claims of philosophical

knowledge are in practice subject to revision, but this is the type of revision we should expect if

philosophical knowledge is objective and not simply a matter of stipulation.

388
 In response to Hales’ excellent book, I conclude that the very fact that we can determinately

recognize, for instance, that the shaman, the follower of religious revelation and the rationalist

philosopher have distinct philosophical perspectives is the cross-theoretic platform from which

competing views can be adjudicated. I have called this platform QI and I have set out its anatomy and

wider semantic and translational grounding in TTS. That Hales did not recognize this common

underpinning of philosophical perspectives is simply a function of the ever-present linguistic

paradigm of philosophy that treats meaning as a type of obvious, semantically determinate

phenomenon relative to words and sentences, without need of identification by text-type-theoretic

concerns. Hales is correct for thinking that there is such a thing as a philosophical perspective, but

these intensions are not linguistic, but rather part of a type of discourse that aspires to perspective-

independent objectivity.

In recognizing that the follower of Christian revelation and shamanic, drug-induced visions,

are, among other things, authoring texts that are susceptible to a philosophical reading and translation,

we undo a latent and spurious distinction in much philosophy, that there is some type of clear

conceptual distinction between religion and philosophy. For this to be possible, the notion of

RELIGION has to be clear. The notion of RELIGION is perhaps the single most pernicious concept at

work in the world today—pernicious because it is incoherent and is used in an unprincipled fashion to

group many unrelated concerns under a common heading, and because it operates in dialectical

opposition to a spurious conception of philosophy. RELIGION served initially to ironically insulate

Western “religions” from rational criticism and is now used by philosophers to ghettoize non-western

philosophies. Religion may be an activity—perhaps we can understand it as a set of rituals and

practices based upon a claim of philosophical infallibility of some perspective, person or tradition.

But philosophy, contra Wittgenstein, is not an activity, but a type of discourse or text. To oppose

philosophy and religion as mutually exclusive categories is like opposing oxygen and the air as

mutually exclusive categories. It makes no sense on systematic reflection, and only furthers the

389
nonsensical view that “faith” can be the source of values. We cannot determinately understand and

translate value discourse except as a species of philosophical discourse and we cannot determinately

understand and translate philosophical discourse except as a discourse about universal and general

theories that aspire to be objectively true, independently of their perspective. Faith has no work to do

here: it’s all theories and reasons. If so-called “religion” has something to offer in the way of values,

it is because it makes philosophical claims, which by virtue of being philosophical, can be adjudicated

according to text-type-theoretic considerations.

Indeed, the same is true for the value and metaphysical claims produced by drug-induced

visions, yogic trances or by meditations on pop culture. Anywhere there is a claim to objectivity in

such matters, and not a mere articulation of preference, we have philosophy as a type of culture- and

language-transcendent discourse.

VIII.5. Conclusion
The aim of this dissertation was to argue that we require what I call the Quasi-Indexical

account of thick and thin concepts in order to translate normative discourse. The argument for this

account of the thick and thin concept distinction was based largely on general considerations of

translation (that translation is not a word for word exchange across languages based upon synonymy

but a process mediated by text-types) and the specific considerations of translating philosophy. It is

quite ironic that for such an old discipline as philosophy (perhaps the oldest in the Western world) no

systematic tract has been written on the unique challenges to translating philosophical texts.

Normative discourse, for its part, it was argued, can only be accurately translated if understood as a

species of the philosophical text. The considerations that people bring to bear in normative discourse

cannot be sharply disassociated from wider philosophical theories and convictions that the translator

must keep in mind in order to properly translate normative discourse. This is a lesson that was learned

not from the armchair, but in overcoming a major obstacle in cross-cultural research, namely the

problem of accounting for the place of ethics in the Indian tradition. Perhaps the topic of translating

390
philosophy could not be systematically addressed until we had come to the point that we have in our

intellectual history. This is the point where we have wrestled with, and attempted to explain, meaning

in terms of language, while simultaneously attempting to deal with the challenges of intellectual

diversity that is a brute fact of human history and cultural diversity. My success in defending this

thesis is a function of these important events in our recent intellectual history. It is clear, however,

that the old way of doing philosophy and moral semantics, by means of deferring to language as the

final arbitrator, is simply a non-starter for understanding translation and normative discourse across

cultures. A new, principled way of thinking about translation and semantics is necessary, and the best

way to access this is to defer to our best insights gleaned from actual cases of translational practice

and cross-cultural scholarship. QI is the result of these investigations and cannot be easily gainsaid.

The argument for philosophical realism that I have presented at the end is not part of the

thesis that I set out to defend. All I set out to defend was the indispensability of QI to normative

discourse translation. The problem of translating normative discourse that QI has proven itself in

overcoming is no small matter: the Indian philosophical tradition is one of the three major

philosophical traditions of humanity, and the fact that it could solve a problem that has stumped

scholars for so long speaks highly in its favour. Given that it fits with a sound, textual account of

translation, that it eschews the naïve linguistic approach to meaning and translation, that it captures

the structure of the philosophical text, and that it has actually proved itself in actual, major, cases of

cross-cultural scholarship, there can be no rational cause to doubt its indispensability in translating

normative discourse.

391
Endnotes

1
Apart from occasional comments, the topic of translating philosophy is almost never discussed in the literature.
Certainly, nothing systematic has been written on the topic (cf. Schmidt 1984; Ashton 1971). Lawrence Venuti in his
Scandals of Translation has a chapter devoted to issues of translating philosophy, but his own training and expertise is
so far from philosophy that his comments are at most informative about how someone without much expertise in
philosophy can become distracted by matters that are of no real consequence in translating philosophy. Venuti
presents a detailed analysis of G.E.M. Anscombe’s stylistics in her translation of Wittgenstein’s Philosophical
Investigations, but Venuti seizes upon issues of connotation. He writes, “For the translator, a more literary approach
turns the philosophical translation into a minor literature within the literature of philosophy” (Venuti 1998, 123).
His analysis is not so much wrong but irrelevant to what is at stake in translating philosophy. Like most translation
theorists, Venuti talks about translation from the perspective of what is relevant to his own primary area of training
and interest, which in his case is literature. This is a major problem with the translation studies literature: no one to
date has presented a systematic account of translation that would allow translators to objectively determine what
types of issues are relevant to their translation projects. What is important for the translation of science is not what
is important for the translation of literature, and neither is obviously or necessarily relevant to the translation of
philosophy. What is relevant is not a matter for negotiation or caprice, but is determined by norms of translation
that underwrite the type of text being translated. Miles Groth has written a book titled Translating Heidegger (2004)
which deals with problems in translations of Heidegger that have resulted in the poor reception of his ideas in
English-speaking countries, as well as Heidegger’s own views on translation. But here too, we find nothing
systematic about translation as such, and translating philosophy in particular. The entire discussion is undertaken as
though to translate Heidegger is a sui generis matter that systematic considerations of translating philosophy (or more
general considerations of translation in light of the diversity of translational projects) do not bear upon. One
collection of papers that comes close to dealing with the particular problems of translating philosophy is Buddhist
Translation: Problems and Perspectives (Tulku 1995). Nothing systematic here either, but the collection is something like
a compilation of the pre-Socratics of translating philosophy. My Ethics and the History of Indian Philosophy
(2007a)contains the kernel for the argument I shall present for translating philosophy.
2
I am not sure if “localism” is a term in usage. If not, I am coining it to distinguish a similar phenomenon, namely
relativism. Relativism is a localism, generalized. The ethical localist, in contrast, really does not care about anyone
else’s localities. The ethical localist who is not a relativist usually has theories of morality and moral knowledge so
that no one else but a few in their privileged locality can have moral knowledge. Aristotle with his emphasis on
ethics as the knowledge of how to live well in the Greek polis, the Brahminical notion that ethics is derived from
the sacred tradition, and Confucius with his emphasis on li as defined by his culture, put forward theories of moral
semantics and epistemology that make it difficult to understand how anyone else can have moral knowledge. The
relativist takes all localities to be privileged for their inhabitants.
3
The hierarchical nature of moral worth is rather naked in almost any Dharma !stra, and the Brahmin, or priestly
caste, is at the top of this heap—no wonder, for these texts were composed by Brahmins themselves. For instance,
Manu writes, “No greater crime is known on earth than slaying a BrBhmana [i.e., Brahmin]; a king, therefore, must
not even conceive in his mind the thought of killing a BrBhmana” (Manu VIII. 381), though it should be noted in
fairness that the dharma !stra’s also betray an exceptional sensitivity to the moral importance of most all sentient
beings that is often quite at odds with its own ritualistic outlook (see for instance the contradictory stance taken on
killing animals in book five of Manu) (cf. Manu 1886). The contradictory nature of these codes certainly supports
the hypothesis that they were a collection of moral decisions of certain schools of thought, each putatively based
upon the Vedas.
4
Confucius lumps women in with servants and says of them that they are difficult to deal with: “if you are familiar
with them, they cease to be humble; if you keep a distance from them, they resent you” (Analects 17:25)—the
possibility that the trouble was with Confucius, and not women and servants apparently did not occur to him. His

392
stance on women is in some respects at odds with his strong emphasis on the importance of respecting and obeying
parents, which runs through the Analects (cf. Analects 4:18-19) (cf. Confucius 1963).
5
For a very wide collection of such technical responses to Quine’s technical arguments that brings together
responses that are not easily found, see Dagfinn Føllesdal’s edited volume, Indeterminacy of Translation (2001).
6
It is a common misconception that Quine’s argument about the indeterminacy of translation is an argument about
the lack of evidence that could decide controversies of translation. This reading of Quine’s argument has lead some
eminent authors to conclude that Quine is simply bringing to light the fact that translation is underdetermined by the
empirical data, in exactly the same way that all empirical theory is (Chomsky 1975, 182; 1980, 258 n.26; 1969).
However, Quine resists this reading of his thesis and claims that his thesis is far stronger (Quine 1969b). Quine’s
argument is rather that there may be cases in translation where no possible evidence could solve problems of
translation. The only way to make sense of the poor prospects for translation, on Quine’s account, is in terms of his
sceptical views about cross-linguistic synonymy that would underwrite translation. Quine’s holism explains his
pessimism on that ground. The radical translation thought experiment is thus, all things considered, a weak
explication of his thesis about translation—and the reason that he spends so much time making a relatively weak
but expansive argument for his thesis is that his real target is not translation but meaning (Quine 1960, 73; cf.
Gaudet 2006). He makes a stronger argument for the indeterminacy of translation in passing: “…the conflict is
precisely a conflict of parts seen without the wholes. The principle of indeterminacy of translation requires notice
just because translation proceeds little by little and sentences are thought of as conveying meanings severally”
(Quine 1960, 78-9). If the parts in question are sentences of languages, and the wholes are the languages themselves,
then the problem of the indeterminacy of translation is the problem of finding linguistic equivalents for items that
have their full range of significance within their native language. The indeterminacy of translation thus turns upon a
very simple and undeniable fact about real languages, namely that sentences are a function of languages (for
grammatical reasons if nothing else) and moreover that languages are all different. Translation, pressed to find
equivalents across languages, must try to match up sentences from one language with sentences in another language,
but none of these are in any strict sense semantically equivalent. Thus, there is no principled way to assure success
in translation. Jerry Fodor and Ernest Lepore seem to come close to conceptualizing this argument in their
diagnosis of what they call “translation holism.” They write, “The argument for translation holism seems to be one
that assumes that meanings supervene on intersentential relations—that they are something like inferential roles—
and hence that translation preserves meaning only if the inferential relations among many of the sentences in the
home language preserve the inferential relations among many of the sentences in the target language” (Fodor and
Lepore 1992, 6). They distinguish this from “anthropological holism”, the view that “There is an internal
connection between being a symbol and playing a role in a system of nonlinguistic conventions, practices, rituals,
and performances—an internal connection, as one says, between symbols and Forms of Life” (Fodor and Lepore
1992, 257). I am really not sure if there is a very sharp distinction to be drawn here. I think Fodor and Lepore are
correct for noting that what is desirable is if this translation holism is incorrect, but meaning is yet not atomic
(Fodor and Lepore 1992, 28). I do not find that Fodor and Lepore show how this can be.
7
For an explication of Quine’s scientific realism, see J.J.C. Smart’s essay, “Quine’s Philosophy of Science” (Smart
1969). Quine writes, “Science reveals hidden mysteries, predicts successfully, and works technological wonders”
(Quine 1969a, 133). Certainly, it would be implausible to deny the latter two accomplishments of science, but what
on Quine’s view allows him to think that science is actually revealing anything? I do not think that Quine is on
strong ground here, and Kuhn’s appropriation of Quine’s views on translation shows this, I think. Quine thinks that
the problem of natural kinds can be resolved into a mature science as simply part of its working order, but that does
not in itself solve the problem of objectivity for pragmatic felicity is not a guarantee of objectivity. See also an early
criticism of Quine’s attempt at reducing away the problem of natural kinds into the workings of a mature science
(Wilder 1972).
8
The unpopularity of Quine’s thesis have widely noted in the literature (Boghossian 1999, 331-2, 1996; Gibson
1998, 28; Lycan 1994; cf. Miller 1998, 128-51). Robert Kirk’s response to Quine warrants a note because his book,
Translation Determined, is the most systematic account of the problem as Quine sets it out, and Kirk argues that

393
translation is perfectly determinate on Quinian grounds. Presuming a certain common starting point for the
acquisition of language, and a parallel linguistic development, translation remains determinate, which implies that
the problem that Quine identifies is really a result of computational uncertainty, and not the very nature of meaning
(Kirk 1986, 215-36). The trouble however is that translation must deal with such computational difficulties that
arise from the reality of cultural and linguistic differences across languages. Kirk’s solution to Quine’s problem thus
operates at the same level of abstraction from real cases of translation as Quine’s thought experiment does. To
embrace such difficulties of computational uncertainty as a fact that will not go away is to simply accept that there is
such a thing as cultural diversity and this is not a static phenomenon but shifting and evolving. Translation must be
successful in the face of this reality, not in the face of unrealistic abstractions from culture.
9
Hilary Putnam is famous for changing his mind. Internal realism is no longer a view that Putnam argues for. What
is not obvious is whether he rejects whole sale the view he argues for in his now classic “Meaning of Meaning”
article (Putnam 1975). The cultural foundations of meaning that he argues for in that paper may be consistent with
his most recent Ethics Without Ontology (2004), which seems to me to be even more pointed in its relativism in
affirming a Wittgensteinien “pragmatic pluralism” (Putnam 2004, 22). One of the purposes of this work appears to
be to establish that there can be objectivity in ethics without appeal to heavy metaphysics. It also attempts to
forward Putnam’s now characteristic view that the world is a mereological continuum and concepts and languages
organize this continuum into distinct sums. I am personally unsure how any of this is supposed to translate into
objectivity in ethics. Putnam begins this work by stipulating what he means by “ethics” (Putnam 2004, 15-32).
However, controversies about what we mean by moral concepts cannot be easily detached from substantive
controversies themselves, and thus the very starting point is partial and perspectival, and hardly a grounds for
objectivity. Putnam also claims that people around the world converge on certain moral truths (Putnam 2004, 75) ,
but I am quite unclear on what makes Putnam so clear on his ability to identify such agreement across languages,
particularly if concepts are relative. Putnam also defends an Aristotelian conception of moral knowledge as had
from being raised in the right way (Putnam 2004, 4, 23, 26, 100), but this can hardly be said to be a ground for
objective moral knowledge, as it is incurably provincial. He quotes approvingly Rawls’ view that the history of
moral philosophy is a family of theories, but all of the theories that Rawls (and by extension, Putnam) identifies are
European (Putnam 2004, 104). Putnam’s entire discourse is quite chauvinistic in my view. Perhaps chauvinism is
not the same as cultural relativism, but it is unclear how any of this is supposed to protect objectivity. Maybe the
shortcomings of this work can be attributed to the fact that it is a collection of a series of public lectures. However,
the disregard for the possibility of cross-cultural controversy on ethics and a cosy provincialism runs deep in
Putnam’s current thought. He distances his thoughts on meaning from translation, and instead prefers to think of it
in terms of use within a linguistic practice (Putnam 2004, 40-1). His collection of papers The Collapse of the Fact/Value
Dichotomy and Other Essays (2002) attempts to secure the objectivity of values, but does not discuss the issue of
translation at all from what I can see. I find myself sympathetic to many of Putnam’s substantive views in ethics: he
comes across as a reasonably compassionate person. But I find his entire approach quite lamentable and
symptomatic of that American anthem, “We are the world.”
10
Another way to conceptualize the trouble for Davidson is in the fact that he takes language to be both the way
that we measure the world and the manner by which we articulate our disagreements with each other. The
proposition thus ends up being both the unit of measurement and the means of disagreeing. Unless he wishes to
make disagreement nonexistent (which is not his approach) there has to be some relativity on his account on the
facts that could explain how differing languages measure the world differently. I too shall later appeal to the notion
of measurement as an apt analogy to elucidate semantics, but on my account, language is not the paradigmatic
manner by which we articulate disagreements. Rather, on my account, it is with texts that we articulate
disagreements, and these need not be linguistic in nature. Differing types of texts, that would permit differing types
of disagreements, would be analogous on my account to differing units of measuring the world, which can be clearly
understood as approaching differing constellations of meaning in semiotic phenomena, just as meters and feet
measure distance differently. Just as we can have a genuine disagreement with an objective resolution of how far it is
between two points, differing disagreements relative to differing types of texts will also be objectively resolvable.
However, it is probably best to understand the analogy between measures and texts of differing types as the analogy
between differing types of measure, such as the difference between the measure of distance and mass. There is thus

394
no real conflict between the differing types of texts, such as say the difference between the philosophical and
scientific text. The types of disagreements they articulate are distinct, and our various types of texts are all of them
complementary, just as measures of differing types of concerns, such as mass and distance, can be complementary
for certain projects.
11
Even if the meanings we associate with language are Platonic entities, we have no guarantee that all real languages
will express such meanings as a matter of linguistic fact and thus our ability to “translate” by matching words and
sentences based upon some putative account of cross-linguistic synonymy seems uncertain. The problem here is
that we do not translate third world entities: rather, we match expressions up in languages based upon their putative
ability to express meanings. But this ability is a natural fact about a language, and as languages are natural
phenomenon, our ability to rearticulate our ideas in any system of representation is contingent upon the linguistic
facts about a language and not guaranteed by a priori considerations, making translation mysterious and
indeterminate in many cases. If linguistic facts about a language were not constraints upon a translation, then we
could merely will translations into accuracy, without ever having to learn the language we are translating into.
Jerald Katz is one author who has proposed a return to a Platonic approach to semantics to overcome problems of
translation, but his solution fails because it cannot escape the brute facts of linguistic diversity. Jerald Katz argues
that Quine’s doctrine of the indeterminacy of translation rests upon the dubious presupposition that “radical
translation” models actual translation. Quine entertains the thought experiment of what is involved in translating a
language from scratch in Word and Object, and he calls this endeavour “radical translation.” Katz is correct for
pointing out that real translation does not resemble Quine’s thought experiment of radical translation (Katz 1993b,
177). Quine’s radical translation thought experiment is the dominant manner in which he argues for his
indeterminacy thesis. I shall show that real translation is not what is at all approximated by Quine’s thought
experiment in Chapters 2 and 3. Quine depicts translation as a matter of attempting to match up expressions across
languages according to their semantic properties. This is not primarily what translation is. Rather, the paradigm case
of translation is the process of reconstituting a text in a new semiotic medium that is semantically equivalent to its
original, but this equivalence is not a matter that reduces to the lexical equivalence of the constituents of the text sub
species aeternitatis. Katz is thus correct for thinking that Quine’s assumption that radical translation shows us
something about real translation is mistaken. However, Katz does not appear to reject the fundamental linguistic
assumption that translation is a matter of paring up words across languages with the same meaning. He writes:
In my story, the field linguist, faced with the choice of translating “gavagai” as “rabbit”, “rabbit stage”,
or “undetached rabbit part”, is not restricted to asking the informant about extensional properties and
relations. He or she can query the informant about intensional properties and relations, too. For
example, the linguist might ask whether “gavagai” bears the same relation to a native expression that
“finger” bears to “hand” or “handle” bears to “knife.” By asking informants about the ambiguity,
antonymy, synonymy, redundancy, superordination, and other sense properties and relations of relevant
examples, the linguist eventually obtains a body of objective data that, together with scientific
methodology, provide independent controls that, in principle, determine the English translation of
“gavagai.” (Katz 1993b, 176 )
“Gavagai” is Quine’s fictitious utterance that his natives speaking “Jungle” say when they see a rabbit, and for
Quine’s fictitious linguist it remains mysterious what this can mean, because the speech acts that make use of this
term are consistent with many possible interpretations. Katz believes allowing the linguist a slew of intensional
concepts allows them to narrow down the possible range of meanings of the word. Katz believes that his
intensionalism unseats naturalism and the linguistic turn and returns philosophy to the state it was in before the rise
of the linguistic turn (Katz 1990, ch. 8; 1993a, 134).
I have sympathy for Katz’s project, in so far as he understands himself as launching a critique of the linguistic turn.
However, returning philosophy to the state it was before the linguistic turn does not seem like progress to me.
Moreover I doubt that Katz has actually made such a reversion. On the point of translation, Katz shows himself to
be operating within the linguistic paradigm, even though he regards meaning as a metaphysical, Platonic affair, for
he regards the meaning as something that is determinately tracked by linguistic items such as words and sentences.
Linguists have always been able to quiz their informants, and still they cannot agree on translations of words, and

395
often it seems that words are incredibly and confusingly polysemous. This is a reality of attempting to translate
words that one realizes in real, cross-cultural scholarship, and this reality is left out of the picture in such
idealizations and thought experiments. At any rate, if translation is a mere matching up of words across languages,
then Katz’ solution is very ineffectual. Consider how impotent it would be to deal with the meaning of
philosophically interesting words like “real,” or “good,” or “knowledge,” (as though informants would all agree on
the semantics of these terms) or even theoretical items such as an electron, or perhaps so called natural kind terms
like “water,” which in many languages have multiple designations. Science as we know it would still be culturally
contingent and relative, in so far as we would be unable to find determinate correlates for theoretical terms in
Jungle, even though we fancy that their meaning is a type of Platonic form. Even with such a metaphysical
conception of meaning, we would not have any reason to think that our way of doing science was anything but an
expression of our culture.
12
Joseph LaPorte in his excellent Natural Kinds and Conceptual Change makes a similar observation. He writes: “even
if reference is determined causally rather than by descriptions, reference changes with theory change. Therefore the
causal theory of reference does not dispel the threat of incommensurability” (LaPorte 2004, 172).
The cultural relativity of the causal theory of reference is rather blatant in its classical articulation. The very thought
experiment that Putnam relies upon to motivate this thesis calls upon the case of aliens from another planet with a
language homophonic and interintelligible with ours whose reference is all the same distinct and fixed by the history
of its language. It so happens, on Putnam’s thought experiment that the aliens come from a planet that is chemically
distinct from ours, though it is alike in all superficial features. The aliens have a word “water” and so do we, and in
both cases we associate it with the same descriptive content (clear liquid, quenches thirst, etc.). Putnam uses this
famous “Twin Earth” thought experiment to motivate the notion that meaning is something fixed by reference to
the peculiarities of the history of a linguistic community. It may seem that what is doing the work here is the fact
that according to this famous thought experiment, the speakers of “Twinglish” come from a world that is
comprised of distinct chemical substances that have superficial similarities with Earth’s environment. However, the
argument turns just as much on the historical and cultural peculiarities of the speakers of English and Twinglish. It
is a contingent fact of our linguistic culture that we have a word for “water” whose ordinary extension is equivalent
with H2O. On the basis of this co-extensionality, and the co extensionality of “water” with chemical XYZ on Twin
Earth, Putnam concludes that “water” is “indexical,” which is to say that speakers of English refer to H2O when
they use “water” and speakers of Twinglish refer to XYZ when they use “water”(Putnam 1975, 152). They both
associate what Putnam calls the same stereotype with these words (which is the widely circulated short descriptor
for the referent of the term) but the respective experts in their respective communities will be able to discern that
the referents are indeed distinct. Kripke comes to the same conclusion, when he identifies “rigid designators” as
terms that identify the same item in every possible world and further claims that names are rigid designators.
According to his view, H2O and water are the same thing, and thus their names are rigid designators for the same
item in every possible world (Kripke 1980 [1972], 48, 128, 140). But cultural difference can make trouble for the
notion that “water” is determinately translatable, even though it may rigidly designate H2O.
13
The view that I shall defend will show us how any text of importance to us could be translated into any language
or symbolic system. We can, and do, add terms to languages when the need arises, or we redeploy older devices for
newer purposes. Thus, Sanskrit could come to have a term for the notion of “electrical charge,” though to my
knowledge, there is no established usage to this effect to date. The trouble with the traditional approach to meaning
and translation that I am bringing attention to is that it makes the contingencies of language itself the conditions of
meaning, truth and translation and thus real linguistic and cultural differences pose an incurable obstacle to
translation on such accounts. In reality, languages are not the fixed cites of meaning that philosophers dream of, but
are semantically changing and evolving social phenomena. Even a language that continues to be spoken today for
purely ritual and academic reasons, such as Sanskrit, evolves as its speakers needs and ideas evolve. For instance, see
the wonderful website, Spoken Sanskrit: <http://spokensanskrit.de/>.
14
The linguistic conceptualization of text-types is front in centre in Albrecht Neubert and Gregory M. Shreve’s
Translation as Text (1992). Something like a concern for text-types can be found in the work of the British Functional

396
Linguist, Michael Halliday in his notion of the thematic structure of texts. Halliday as well discusses the issue as
within the purview of linguistics (Halliday 1985).
15
It is noteworthy that the history of Western philosophy since the linguistic turn has been marked by a sharp lack
of curiosity for philosophical thought that is not Western. This contrasts sharply with the Ancient Greeks, who had
some interest in philosophy that was not theirs. For instance, Pyrrho of Elea was reputed to have made a trip to
India to study with the “gymnosophists” (naked wise people, who may have been the Jain ascetics who were naked,
though his philosophy seems more Buddhist than Jain) (cf. Flintoff 1980). There are also the records of
Megasthenes, envoy of the Macedonian army officer turned king, Seleucus Nicator to Indian kingdoms, including
the Indian emperor Chandragupta Maurya in the fourth century BCE. His Indica reflects the ancient willingness to
recognize philosophy as not a particularly Greek affair. It survives only in the fragments of other ancient authors:
Megasthenes makes a different division of the philosophers, saying that they are of two kinds--one of
which he calls the Brachmanes [Brahmins], and the other the Sarmanes [Sramanas]… The [Brahmin]
philosophers have their abode in a grove in front of the city within a moderate-sized enclosure. They live
in a simple style, and lie on beds of rushes or (deer) skins. They abstain from animal food and sexual
pleasures, and spend their time in listening to serious discourse, and in imparting their knowledge to
such as will listen to them. … Concerning generation, and the nature of the soul, and many other
subjects, they express views like those maintained by the Greeks. They wrap up their doctrines about
immortality and future judgment, and kindred topics, in allegories, after the manner of Plato. Such are
his statements regarding the Brachmanes… Of the Sarmanes he tells us that those who are held in most
honour are called the Hylobioi. They live in the woods, where they subsist on leaves of trees and wild
fruits, and wear garments made from the bark of trees. They abstain from sexual intercourse and from
wine. (fragments from Strabo, XV. 1. 58-60, collected in McCrindle 1877)
The early church father, Clement of Alexandria, in his Stromateis also shows a remarkable openness to the diffusion
of philosophy among all peoples,—on his account, Greek philosophy derived in great part from the philosophy of
the “barbarians,” which included the Egyptians, Druids, Jews, Persians and Indians (Clement 1991 I.xv). Plotinus
apparently was interested enough in Eastern philosophy that he joined an expedition of Gordian III against the
Persians in apparent hopes of making it to India, or at least Eastward to learn something of the philosophers there.
The similarities between Neo Platonic thought and certain strands of Indian philosophy have been long recognized
(Armstrong 1936; McEvilley 1980). Unfortunately, Gordian III was lynched by his own soldiers after some military
success, cutting short Plotinus’ sabbatical to the Orient (Oost 1958). In the modern period, several canonical
European philosophers had clear interest in responding to Eastern philosophy. Leibniz, for instance, was extremely
impressed by the Chinese script, and very interested in Chinese religion and philosophy (Leibniz 1994).
Schopenhauer, for what it is worth, was very influenced by Advaita VedBnta. Nietzsche held a fascination for Indian
thought, and particularly the Brahminical thought of Manu’s Dharma !stra, though it seems quite clear from recent
scholarly accounts that his actual grasp of things Indian was very weak (Brobjer 2004; Smith 2004). Yet he himself
seemed to have been impressed with what he thought was philosophically characteristic of India. Hegel dismisses
non-western philosophy as not philosophy at all, for it mixes religion and philosophy (Hegel 1990 [1825-1826]).
Hegel’s description of “Eastern” thought is ridiculous and constitutes a classic example of orientalist discourse that
depicts the East in a strange and exotified manner that does not correspond to the facts on the ground. (For
instance, one reads at length in Hegel about the “despot” that characterizes Oriental political rule.) But at least
Hegel addresses the possibility of philosophy that is not Western. Hegel needs to dismiss Indian philosophy, for his
Absolute Idealism is a rather blatant repackaging (with an interesting twist) of on an old theme one finds in the
Upanisads, namely an ultimate Being creating the universe in an evolutionary manner through self-contemplation—
if it is not dismissed, it would be rather apparent that Hegel’s view has its historical antecedents in the East, which
would not only ruin the Eurocentric nature of his narrative of thought unfolding in history, but it would also be
rather clear that his view can be understood as a synthesis of Upanisadic speculations and problems leftover by
Kant’s discussion of the antinomies. Not only would this undermine Hegel’s contribution to the history of
philosophy but it would also detract from the notion that there is a single grand narrative to be told about the
history of thought. Husserl too, is dismissive of Indian and Chinese philosophy as not real philosophy, for it lacks
“theory.” It is thus the white man’s burden to guide the whole world for only European thinkers have had a

397
synoptic theoretical perspective necessary for policy (Husserl 1965, 164-78). Bigoted as it is, at least Husserl felt the
need to recognize Eastern philosophy as a potential alternative source of philosophical knowledge. By Husserl’s
time, however, what little interest there is in philosophy from the East is for the most part an afterthought and not
part of the very project of philosophizing. Heidegger is reputed to have had some interested in Eastern philosophy
late in his career (cf. Zhang 2006). Karl Jaspers, something of an exception, has written on the Buddha and
Confucius (Jaspers 1962). There is the notable case of Derek Parfit and his followers who recognizes the Buddhist
view of the self as a precursor to his (Parfit 1984; Stone 1988). Yet, it is difficult to find analytic philosophers who
believe that there is something that can be learned by studying the history of Eastern philosophy. (Indeed, I know
of no other analytic philosopher aside for my self who has an active research project in Indian philosophy as such).
The notion that we might advance philosophical knowledge by studying it in a manner that is not directly tethered
to cultural context was taken seriously in the modern period for they rightly believed that philosophical knowledge
is not anthropological. The linguistic turn however makes philosophy barely distinguishable from a native
anthropology. It makes ethnocentrism into a research methodology. Cultural solipsism is a sad inevitability of this
move to language.
16
In their excellent edited volume, Naturalism in Question, Mario De Caro and David Macarthur note that
“naturalism”, though polysemous, appears to set out the following core doctrine:
“1. An Ontological Theme: a commitment to an exclusively scientific conception of nature;
2. A Methodological Theme: a reconception of the traditional relation between philosophy and science
according to which philosophical inquiry is conceived as continuous with science” (De Caro and
Macarthur 2004, 4)
I am sympathetic with 1, but I reject 2 if this implies that there is nothing like a distinctly philosophical realm of
inquiry or philosophical content.
17
I have Stuart Shanker to thank for pointing out the relevance of Ryle’s and Williams’ relationship to this issue.
18
For a discussion of the nefarious role of racial ideology in strong views about linguistic meaning, see Talbot
Taylor, “Normativity and Linguistic Form” (1997b, 157-158).
19
See his Language and Responsibility (Chomsky and Ronat 1979, 190). Chomsky’s views on language are in many
respects more sophisticated than the legions of philosophers who have looked upon Chomsky as vindicating
linguistics. Chomsky’s account of language operates at a level of abstraction that renders it both manageable to
systematize but quite irrelevant to the types of problems that we are engaging in. He writes “Certainly, it is true that
no individual speaks a well-defined language. The notion of language itself is on a very high level of abstraction. In
fact, each individual employs a number of linguistic systems in speaking. How can one describe such an amalgam?
(Chomsky 1975, 54). Indeed, we might wonder how anyone could determinately succeed in describing such an
amalgam with Chomsky and we should agree that it is quite impossible. Yet, the dominant mode of addressing
questions of semantics in the analytic tradition attempts just this, but for speech communities. There are other
reasons to note that Chomskyan insights about universal grammar and grammatical transformations will not help us
in our questions in translation. Chomsky’s transformations pertain to the relationships between the surface structure
of sentences in a language and the deep grammar of particular languages: the transformations do not pertain to
relationships across deep or surface grammars of distinct languages. I thus respectfully will leave aside Chomskyan
pursuits in universal grammar as something both interesting and of little consequence to our problem. The task of
overcoming cultural relativism and rendering translation determinate is the task of figuring out how we mediate real,
non-abstract languages in a linguistically and culturally neutral manner. In other words, we need to own up to the fact
that the challenge of linguistic diversity that philosophy must come to terms with in order to overcome problems for
objectivity are anthropological and empirical in nature, not abstract and rarefied. When we recognize that it is the real,
empirical language that we must address in translation, the notion that language is semantically determinate becomes
most implausible. But we will have other reasons to reject the linguistic turn that are more to the point. None of this
implies that discoveries in the realm of theoretical linguistics are necessarily irrelevant to translation. Our various
epistemic endeavours tend to complement each other and thus we should not be surprised if a certain clarity on some

398
theoretical issues should help sharpen certain choices in translation (cf. Malone 1988). But, there is no necessary
connection here. Given that translation is fundamentally a semiotic phenomenon, and not simply linguistic,
theoretical linguistics, and all linguistics, is of secondary importance to challenges of translation.
20
It is no surprise that “dharma” has come to be recast as a designate of “religion.” India’s colonial masters had an
interest in reconstruing indigenous moral thought in terms of religion so that it could be compartmentalized,
communalized and rendered irrelevant to questions of governance. Western ethical reasoning could thus fill the
seemingly patent indigenous void in moral reasoning and colonial intervention could be depicted as a moral saving
grace for a people who were construed as having no indigenous tradition of moral reflection. This is enough to lead
us to be suspicious about the notion of religion. But there are other reasons to be suspicious of the notion of
religion. The trouble with the notion of religion in studying non-Western intellectual traditions is not that religion is
a concept that derives from the West. The trouble is that it is fundamentally an incoherent notion that is used, from
what I can tell, to ghettoize philosophies that are non-Western. The fact that scholars of Indian and Chinese
philosophy must frequently seek out jobs in religious studies departments and that philosophy departments in the
west frequently do not offer advanced, graduate training in non-Western philosophy, is perhaps the best evidence
that the category is really good for very little. Certainly, we can classify phenomenon that is non-Western under the
heading of “religion,” but if there is nothing coherent or essential to this definition, it’s a matter of cherry picking
and mere fancy what ends up being called religious. Thus, Confucianism, Taoism and Buddhism (all atheistic) are
called religions, when it is very difficult to see what they have in common with Catholicism or Islam. Worse, from
what I can see, there is nothing semantically that corresponds to the notion of a “religious” text. Certainly, there are
texts that people who identify themselves as religious regard as holy or important, but from the question of
translation, there seems to be no common text-type. And, traditionally, votaries of specific faiths have resisted the
notion that their favourite text could be treated just like any other text. But this is what text-types demand: it is
because they are formal, do not specify authorship, but are concerned with some semantic features of a text to be
preserved in translation, that they can objectively preserve semantic content in translation. The notion of the
religious text, as it was developed in the West, through texts such as the Torah, the Bible and the Quran, is the very
antithesis to the notion of a text-type. Each text is thought by its adherents to be special and unique. But if such
specialness is a means of resisting the notion that they can be treated as merely one token of certain types of texts,
then the claim to specialness, if true, renders them semantically indeterminate (which is to say that one can get
anything one wants out of them, for there is no semantic fact of the matter as to what they mean, which could be
determinately set out in translation). The defenders of such texts however may be on to something: we could
translate such texts as histories, or alternatively as philosophy, or perhaps as some very early version of science, but
in all cases, the results would be quite impoverished. Indeed, they are quite short on argument and reasoning, and
the stress on the mysterious origin of doctrines contained therein suggests that they themselves resist being
understood as philosophy. This, unfortunately, undermines their claims to be foundations for values, for evaluative
and normative discourse, I shall argue, is simply an aspect of the philosophical text-type. We can contrast this with
the so called “religious” texts of the East: the Tao Te Ching, Confucius’s Analects, the Upanisads, the Bhagavad G7t!
(which only recently acquired the status of being a religious text when “Hindus”—a recently minted religious
group—wanted something identifiable and small to swear oaths upon), the PBli Canon, the Mah!y!na S8tra-s the Jain
S8tra-s—all of these texts are first and foremost texts of philosophy that contain arguments in favour of universal
and general theories on all matters of philosophical interest. It also happens that they are revered by large swaths of
people, but the importance that people attach to these texts has no obvious impact on how they can be translated.
If there was something like a text-type of religion, we could translate texts as both tokens of religion and tokens of
philosophy. However, I doubt very much that there is anything like a text-type of religion, given the ridiculous
heterogeny of texts that are identified as important for religious reasons: their importance is not semantic but
sociological. If there is no such thing as the religious text, then the distinction between religion and philosophy that
we see has been forwarded is dubious: it is like the distinction between water and an ocean. The distinction initially
served to insulate western religions from rational criticism. Later, it served to allow naturalistic philosophers a way
of not having to respond to philosophical arguments from those who reject materialist metaphysical platforms.
Now it serves to insulate philosophy as practiced in the west from philosophical ideas born elsewhere. But this is
simply part of the solipsism of philosophy as practiced in the linguistic paradigm. Ironically, it has allowed religions
to thrive and claim a moral high ground in moral epistemology—exactly what naturalistic philosophers often want

399
to refute. After providing a systematic account of the semantics of normative discourse, this high ground will appear
very poor. Religion might be reclaimed as a set of social practices bases upon the putative philosophical infallibility
of some text, person or tradition. But this sharpened notion of religion will serve to undercut religious claims to
infallibility, for religious texts will be treated consistently as types of philosophical texts, and all the institutional
considerations that go into philosophical criticism will be brought to bear upon religious texts. But this is as it must
be: claims of specialness are incompatible with the claims that such texts have important things to say on topics
such as ethics. This will be shown in the course of my argument.
21
I know of only one other active philosopher today who is also a translation theorist, aside from myself, and I met
him rather by accident (cf. Sarukkai 2002). The translation theory literature, before the new brand of “translation
studies” had taken over, was very sensitive to philosophical developments (cf. Jakobson 1992 [1959]; Nida 1964;
Steiner 1975). A rare exception in recent translation studies is the work of Anthony Pym, that is sensitive to
philosophical contributions to translation theory (Pym [2002]).
22
Several translation theorists make use of the notion of text-types (cf. Holmes 1988 [1972], 74-76; Laviosa-
Braithwaite 2001, 277-278). The recognition, for instance, that there are specialized areas of translation, such as
“scientific” or “technical” translation is an implicit recognition of the indispensability of text-types to translation. A
classic explication of the relevance of texts to the project of translation can be found in Albrecht Neubert and
Gregory M. Shreve’s Translation as Text (1992). Their discussion of the relevance of texts to translation is both
insightful and confused. On their account, texts are reducible to language and translation is always a linguistic affair.
If this true, then the addition of the notion of texts to translation theory is no advance on the prevailing linguistic
model of translation. For more on their confusions, see endnote 27 of this chapter.
23
This tripartite breakdown of different systems of translation is modelled after Roman Jakobson’s influential
scheme. However, Jakobson’s scheme is inaccurate if not unjust. According to Jakobson, only interlinguistic
translation counts as “translation proper” and moreover he relegates the translation of verbal languages into sign
languages to the category of intersemiotic translation, and does not realize that the category is better suited for
translation between different symbol system types. Sign language, whatever else it is, is linguistic in nature and thus
ought to be recognized as such (Jakobson 1992 [1959]).
24
Examples of obligatory features of grammars and their asymmetry include: one language demands that when
speaking about an item, some definite number be specified, while others lack such requirements; another language
demands that all nouns be gendered, while others do not.
25
The hidden premise in this argument, never acknowledged by its proponents, is that a translation must be
immediately intelligible to its target language readers. If this easy uptake requirement were abandoned, then literal
translation would be wholly feasible.
26
This argument of course presumes a restricted conception of both meaning and semantics. If the pragmatic
function of language was included in a theory of meaning, and if the equivalence of meaning were essential to
translation, then we would be led to translate the phrase “sweating like a pig” into a construction in the TL that
meant ‘perspiring copiously’. Critics of equivalence tend not to recognize this assumption.
27
It is difficult to blame any particular philosopher or translation theorist for the state of utter confusion and foggy
thinking on issues of meaning and translation over the last hundred years. The linguistic turn is so pervasive that its
absurdity is only paralleled by its diffusion. But the absurdity of the paradigm is most striking when it rears its head
in the comments of translation theorists. André Lefevere and Susan Bassnett assert (not argue) that to think about
translation in terms of equivalence is to adopt a model of translation after the fashion of St. Jerome. According to
Bassnett and Lefevere, “to elevate faithfulness to this central position, to the exclusion of many other factors, the
Jerome model had to reduce thinking about translation to the linguistic level only” (Bassnett and Lefevere 1998, 2) .
Apparently they too had never employed the notion of equivalence relative to a standard. This is truly odd, because the
normal way to employ the concept of equivalence is to assume a standard against which equivalence is measured and there
is no reason why this standard cannot be textual, and not linguistic. The same confusion between meaning, language, and

400
semantic equivalence is found in the work of Albrecht Neubert, and Gregory M. Shreve. Their account of the “text-
linguistic” model comes close to what I shall be arguing, however their account is incurably contradictory and absurd. They
write” “In the text-linguistic model meaning is not sentence-bound. The model locates and distributes meaning
equivalence throughout the text. Instead of being isolated in words and sentences, meaning is also carried globally in
the text” (Neubert and Shreve 1992, 23). Within a few sentences, they also make the following assertion: “In the
text-linguistic model, translation does not involve the transfer of meanings” (Neubert and Shreve 1992, 23). If
meaning equivalence is carried globally through the text, its translation will certainly involved a transfer of meaning.
The problem here is the same: meaning is conflated with language, and there is no way to think about equivalence in
meaning except in terms of language. This type of absurdity will rear its head again, but we shall see in Chapter 4
that philosophers too are guilty of this fallacy.
28
In other words, Nietzsche would thus prefer a translation of a text that maintained all of its aesthetic features at
the expense of its cognitive significance. Textual transformations that maintain cognitive significance at the expense
of the aesthetic features he calls “generalizations” and not translations.
29
Another reason for holding fast to the ideal of equivalence and fidelity in translation is to provide a normative
standard against which translations and adaptations can be distinguished. Adaptations have traditionally been
regarded as similar to translations, but not the same as translations because of the degree of deviance allowed an
adaptation. One might argue that, because of the recent critique of strict conceptions of equivalence in translations,
it might no longer be possible to draw a strong boundary between adaptations and translations (cf. Bastin 2001).
LET weighs into this debate by clarifying that the distinction between a translation and an adaptation is not to be
judged in terms of degree of divergence between the total form and content of a ST and TT, but with respect to
divergence from certain critical properties of the original. On LET, much that might have seemed an adaptation in,
say, the arts will end up being a translation, though some symbolic transformations will not. LET thus agrees with
those who believe that equivalence is the critical criterion that separates translations from non-translations (Neubert
1994b, 1994a; Pym 1992, 2004, 166).
30
While Nietzsche translates the fragment as “Whence things have their origin, there they must also pass away
according to necessity; for they must pay penalty and be judged for their injustice, according to the ordinance of
time.” and another philologically sensitive translator produces “But where things have their origin, there too their
passing away occurs according to necessity; for they pay recompense and penalty to one another for their
recklessness, according to firmly established time” (quoted in Heidegger 1984 [1946], 13) Heidegger concludes that
it should be rendered “…along the lines of usage; for they let order and thereby also reek belong to one another (in
the surmounting) of disorder” (Heidegger 1984 [1946], 57).
31
Derrida goes so far as to say that the problematic of translation is the problematic of philosophy: the attempt to
make writing determinate. He writes “It is a difficulty inherent in its very principle, situated less in the passage from
one language to another, from one philosophical language to another, than already, as we shall see, in the tradition
between Greek and Greek; a violent difficulty in the transference of a nonphilosopheme into a philosopheme. With
this problem of translation we will thus be dealing with nothing less than the problem of the very passage into
philosophy” (Derrida 1981 [1967], 71-2).
32
Derrida, in his characteristic ambivalence, does not seem to rule out the possibility that “pharmakon” can be
translated. He writes “All translations into languages that are the heirs and depositaries of Western metaphysics thus
produce on the pharmakon an effect of analysis that violently destroys it, reduces it to one of its simple elements by
interpreting it, paradoxically enough, in the light of the ulterior developments it itself has made possible. Such an
interpretative translation is thus as violent as it is impotent: it destroys the pharmakon but at the same time forbids
itself access to it, leaving it untouched in its reserve” (Derrida 1981 [1967], 99).
33
For an attempt to provide a positive account of translation from the perspective of Deconstruction, see Davis
(2001).

401
34
The view that translations can alter the manner in which we read originals is more consonant with Foucault’s idea
of the revisibility of the significance of the past, though neither Foucault nor Derrida seem to explicitly affirm this
as an option, as far as I can see. Gentzler however is convinced that the philosophies of Derrida and Foucault
secure such a possibility. See Gentzler on Deconstruction (2001, 145-186).
35
The American scholar of translation Lawrence Venuti is a prominent critic of this convention. One implicit
reason that Venuti is critical of this convention is that it contributes to the invisibility of the translator, which
contributes to the lack of accountability in translation. The main reason that Venuti supports granting the translator
the status of authorship is that authorship on his view is always a derivative affair and is thus no different from
translation (Venuti 1998, 43). While Venuti is correct that no author creates from scratch, there is certainly an
important difference between whatever creative license a translator deserves and the one afforded to the original
author. There is a greater role for creativity in translation in such text-types studied under the banner of “literature”
and here it may seem that Venuti’s proposed spotlight on the role of the translator is welcome. But an essay on the
Critique of Pure Reason (particularly one in relationship to a text such as the Critique of Judgment) makes little sense if it
is an essay on Norman Kemp Smith’s work, and not Kant’s work.
36
Those who understand translation as a central part of hermeneutics, or interpretation, may disagree. I have
already presented one reason in this chapter for us to distinguish between interpretations and translations, namely
the latter take on the textual identity of the ST while interpretations do not. I will address in greater detail
Gadamer’s argument for viewing translation as a special case of interpretation in the conclusion. I will argue that
Gadamer’s own diagnosis that translations are interpretations is fraught with tension, and does not match up to a
sound account of translation that all the same avoids the pitfalls of “Historicism” as he understands it. While
Gadamer’s views on this matter are influential, I shall argue that they are deeply muddled, and the lack of clarity for
his part is due to his commitment to the linguistic paradigm that a sound account of translation disavows.
37
In response to apparent problems with the system, Carnap himself moved to a type of semantic holism in “The
Methodological Character of Theoretical Concepts” (1956) according to which a concept (the semantic entity
behind a term) has its meaning relative to the theory and language it is a part of. Quine’s work was thus part of a
shift that already begun in Carnap’s thought.
38
The exegetical question of what Quine’s doctrine of indeterminacy of translation is has been a controversial issue,
which I shall not explore in great detail here (cf. Føllesdal 2001). I focus on a reading of the thesis in Kirk’s
Translation Determined (1986) that I find naturally comes off the pages of Quine’s Word and Object. Informed by this
reading, I take it that Quine has two explicit ways of making the argument for the indeterminacy of translation
thesis. One way is to argue that sentences in any language can be mapped on to themselves in an infinite number of
ways, but yet remain consistent with the totality of linguistic behaviour. Another way that he makes the argument is
by appeal to the field linguist’s task of radical translation, and the conclusion of the thought experiment: that field
linguists following the same empirically austere methodology can (and are likely) to arrive at incompatible manuals
of translation (Quine 1960, 27). Furthermore, the argument can be understood as making an epistemic claim, about
our inability to arrive at knowledge of what competing translations are correct in certain cases, and an ontological
claim about there being a lack of data to resolve such conflicts. Michael Friedman believes that Quine’s claim is
primarily ontological (i.e., that the epistemological problem is a result of the more basic ontological problem), and
moreover that Quine’s view is that translation is not only underdetermined by the relevant facts, but all possible
facts of nature conceived in terms of underdetermined theories of nature (Friedman 2001 [1975]). I am sympathetic
to this reading, and thus I believe that Chomsky’s criticism that Quine only shows that a theory of the meaning of a
language is underdetermined, but not indeterminate (Chomsky 1975, 182; 1980, 258 n.26), is incorrect as a
characterization of Quine’s thesis (though it’s not far off from a characterization of the force of his argument, I
think). However, we grant Quine too much if we allow him to set the terms of the debate in terms of the
ontological issue. My goal is to argue that the epistemological problem can be solved and that the ontological issue
is a distraction.

402
39
Quine wants to distinguish the ordinary underdetermination of empirical theories that we find in the sciences
from the problem that plagues linguistics. In the case of natural scientific theories, there is underdetermination,
Quine thinks, but we can be realists about what science has to teach us. However, if we take all that science could
ever have to teach us, underdetermined as it is, we still would not be able to determine translation (i.e., have
evidence that speaks in favour of only one translation manual to the exclusion of others). This is what the
indeterminacy of translation and meaning consists in, Quine thinks (Quine 1969b, 303). However, Quine’s
argument here is very weak, for he himself pins so much on the underdetermination of the linguist’s theory by
observable data—particularly stimuli—that the only way to make sense of Quine’s attempt to distinguish linguistics’
problems from those of physics is that in the case of physics, we are more or less institutionally unanimous on what
underdetermined theory we wish to be realists about whereas in linguistics we subscribe to several different
translation manuals. His certainty that linguistics is a different type of discipline from physics, which is incapable of
institutional unanimity, does not sit well with his self-professed radical empiricism—why is this not simply a
contingent truth? Where does the necessity of this difference come from? In his zeal to argue for “indeterminacy,”
Quine loses sight of his main premise however; that equivalence in translation seems nowhere to be found. If
translation could determine equivalence, the issue of indeterminacy would evaporate.
40
Quine expresses his scepticism about equivalence in translation while articulating his thesis of the Indeterminacy
of Translation. He writes: “The thesis is then this: manuals for translating one language into another can be set up
in divergent ways, all compatible with the totality of speech dispositions, yet incompatible with one another. In
countless places they will diverge in giving, as their respective translations of a sentence of the one language,
sentences of the other language which stand to each other in no plausible sort of equivalence however loose” (Quine
1960, 27, my italics). It is noteworthy that Quine’s criticism of equivalence is unqualified here. Yet, it is important to
note that there is a sense in which Quine is not criticizing equivalence in translation: empirical equivalence. Quine’s
point is to show that empirical equivalence does not guarantee the type of equivalence that we would normally
expect from two translations: semantic equivalence. His criticism of equivalence then seems best understood in
terms of intensional concerns. Within the context of translation, this seems to be the most natural way to
understand talk of equivalence however, which probably explains why Quine does not qualify his criticism of
equivalence in translation.
41
Quine is quite inconsistent in his treatment of the paradigm example of “Gavagai” and “(lo a) Rabbit.” In Word
and Object, he recognizes that these occasion sentences will likely not have the same stimulus meaning (Quine 1960,
37, 39). However, he later regards them as being intertranslatable observational sentences (Quine 1996 [1992], 452),
which are occasion sentences with stimulus meanings that are not affected by collateral information provided to the
native speaker (such as promptings, or conditioned stimuli that replace a natural stimulation that prompts assent or
dissent). Given his analogy of the problem of indeterminacy with regard to parts and wholes that I set out in this
same paragraph he ought not to regard any observational sentences as translatable. If some are, then translation as
such is not indeterminate.
42
Quine appears rather confused on the topic of the relationship between matters such as “ontological relativity”,
“the inscrutability of reference” and the indeterminacy of translation. Quine often treats these as distinct topics
(Quine 1996 [1992]), but yet they have a bearing on each other. For more on this issue, see Kirk (1986, 106-11.)
43
For ease of explication I have been glossing over a distinction that Davidson draws at one point between
interpretation and translation. According to Davidson, interpretation involves a relationship between two languages,
while translation involves a relationship between an object language (SL), a subject language (TL) and a third
metalanguage (Davidson 2001 [1973], 129). Davidson appears to be echoing Carnap’s view that translation is always
a translation from one language into the sub language of another language (Carnap 1937, 224). (The metalanguage
that contains the sublanguage provides the resources to talk about the translation qua translation.) It is of course
difficult to imagine that any translator thinks about translation this way. However, the distinction between the
various types of language is formal, and not meant to reflect natural language distinctions. Moreover, as Davidson
himself speaks about his employment of Tarski’s schema as enlightening translation and interpretation, the
distinction matters little to understand Davidson’s thought.

403
44
There is some textual grounds for thinking that Quine in Word and Object did not have such a restricted use for
charity. My explication here follows Davidson’s evaluation of Quine’s view, which Davidson formulates after
Quine’s 1960 Word and Object. See fn.54 on p.405.
45
Davidson’s revision of the principle of charity into the principle of rational accommodation, along with a holism
that he often espouses (Davidson 1996 [1974]) was foreshadowed by Richard Grandy’s “Reference, Meaning and
Belief” where he puts forward what he calls the principle of humanity. It is a “pragmatic constraint on translation, the
condition that the imputed pattern of relations among beliefs, desires and the world be as similar to our own as
possible” (Grandy 2001 [1973], 389 ). However, Grandy’s may be criticized for still giving the impression that the
aim in interpretation is to maximize agreement.
46
As noted, Quine later shies away from this strong position when he suggests that “Gavagai” and “(lo a) Rabbit”
are translatable as observation sentences (Quine 1996 [1992], 452). However, if he concedes such translations, then
translation is not indeterminate—as Kirk (1986) makes clear .
47
Fodor and Lepore make a similar charge in their “Is Radical Interpretation Possible?” (1994) against Davidson.
On their reading of Davidson, radical interpretation presupposes a state of epistemic poverty where interpreters do
not have access to institutional knowledge. Davidson in his response, “Radical Interpretation Interpreted” (1994),
protests that Fodor and Lepore have completely misunderstood his view. Davidson clarifies that for him, the
radical interpreter can have access to all sorts of knowledge, short of the intentions, beliefs and wants of an
interpreted subject (Davidson 1994, 125). I find it unclear however as to why Davidson should hold on to the
doctrine of the indeterminacy of translation if institutional knowledge is allowed. For, if nothing else, institutional
decisions, and knowledge of them, can by fiat resolve problems of indeterminacy by making decisions about
translation conventions. Moreover, Davidson seems to fail to address the fact that he wrote (as Fodor and Lepore
seize upon) that the only evidence that is respectable in the context of radical interpretation is that which we would
grant to the “virgin investigator” that he has “without his already being in possession of the theory it is supposed to
be evidence for” (Davidson 1996 [1974], 457). This type of information barred to him seems to be precisely the
institutional information that would make interpretation and translation determinate.
48
To a certain extent this goes against the grain of translation studies. University programs in translation studies
often treat specialized translation as one would treat specialization in medicine: something that one gains after
gaining competence in the field on the whole. The knowledge involved in being a generalist and a specialist, on the
common model, is essentially of the same sort (Kaiser-Cooke 1997, 288). This I believe is the wrong model for
translation, for there is little that will be common to all text-types. Fortunately, specialization in translation is
increasingly being given importance in translation studies programs (Maier and Massardier-Kenney 2001 [1993]). At
the moment, the idea of the expert in translation at-large continues to dominate the literature in translation studies.
This will continue to be the case until there is increased pressure for determinacy in translation. Part of the barrier
to such pressure is the widespread belief that no such thing is possible. For those academic areas concerned with
cultural representation and communication, the pressure is urgent, as it is in legal (Madsen 1997; Sarcevic 2000) and
scientific translation (Pinchuck 1977; Hann 1992; Wright and Leland D. 2001 [1993]). It is urgent in anthropology
as well (Wolf 1997) though anthropology to date has focused upon the problem of interpretation over translation.
49
The case of institutions, such as legal institutions or text-type institutions, contrasts sharply with the case of
linguistic practice. I address this contrast in greater detail, and in light of Wittgenstein’s argument for rule-following
paradoxes, in Chapter 7.
50
Kirk notes that Quine actually does not ever provide an example of the indeterminacy of translation (Kirk 1986,
xx). This is sufficient evidence to be suspicious of the thesis, I believe. The one example of “Gavagai” associated
with Quine’s thesis of the Indeterminacy of Translation was never meant to illustrate the thesis, according to Quine.
He writes: “…Hubert Dreyfus has California vanity plates on his Volkswagen Rabbit that spell ‘GAVAGAI’. The
word has become the logo of my thesis of indeterminacy of translation, and now it is making its way in a wider
world. Ironically, indeterminacy of translation in the strong sense was not what I coined the word to illustrate. It did
not illustrate that, for ‘Gavagai’ is an observation sentence, firmly translatable holophrastically as ‘(Lo, a) rabbit’. But

404
this translation is insufficient to fix the reference of ‘gavagai’ as a term; that was the point of the example” (Quine
1996 [1992], 452). It is the “indeterminacy” or “inscrutability” of reference that this example demonstrates.
51
Wittgenstein in the Preface to the Investigations writes: “The thoughts which I publish in what follows are the
precipitate of philosophical investigations which have occupied me for the last sixteen years. They concern many
subjects: the concepts of meaning, of understanding, of a proposition, of logic, the foundations of mathematics,
states of consciousness, and other things” (Wittgenstein 1958). MEANING and UNDERSTANDING indeed are front
and centre in this work.
52
Pym’s aim is to provide an account of the fault-lines between translation and the recent field of localization,
which concerns the adaptation of products and their textual elements to locals or varying markets. Pym admits that
the bulk of the argument is a reworking of his Translation and Text Transfer (1992). However, he also notes that
much has changed. In the 80s, the concern with translation was to initiate a conversation between cultural studies
and linguistics. In the 90s, the concern was to address issues of cultural sovereignty. Pym’s recent instalment is
aimed at addressing the pressures on translation and localization from globalization and internationalization.
53
“All the objective data he has to go on are the forces that he sees impinging on the native’s surfaces and the
observable behaviour, vocal and otherwise, of the native. Such data evince native “meanings” only of the most
objectively empirical or stimulus-linked variety” (Quine 1960, 28).
54
Some might argue that this tendency was already in Quine. For though most of his account of radical translation
did not defer to criterial considerations, in elaborating the principle of charity, he makes a case for its plausibility by
applying it to translation as such (Quine 1960, 59). This of course brings into question the extent to which Quine
only restricted the principle to the problem of determining logical connectives, as Davidson (Quine’s close student)
understands Quine after the release of Word and Object (cf. Davidson 2001 [1973]). While Quine may have at times
understood charity quite broadly, it is fair to note that he rarely makes use of it in his account of radical translation
(cf. Quine 1960, 26-79), and thus Davidson may rightly be regarded as further applying the principle of charity
where Quine neglected it.
55
Vermeer’s use of “skopos” is problematic. For the most part, the term appears to mean ‘purpose’ or ‘pragmatic
aim of translation’. However, he suggests that the term skopos has three different references: (a) “the translation
process, and hence the goal of this process; (b) the translation result, and hence the function of the translatum; (c)
the translation mode, and hence the intention of this mode” (Vermeer 1989, 171). It’s not clear what the third sense
of “skopos” has to do with the other closely related senses. However, his theory appears to be intelligible without
concern for the third sense.
56
An early application of Firth’s theory of language to translation is found in J.C. Catford’s A Linguistic Theory of
Translation. Catford shuns literal attempts at translation that preserve the “meaning” of the ST in the TT, calling this
project “trans-coding.” Rather, “SL and TL texts or items are translation equivalents when they are interchangeable in
a given situation” (Catford 1965, 42-49). Firth’s ideas have also made their way into subsequent scholarship through
the work of M.A.K. Halliday (cf. Halliday and Hasan 1989 [1985]).
57
The authors referred to here were part of a conference organized around the problem of translating Buddhist
texts—texts with clearly religious, philosophical and anthropological import. The dominant perspective of such
translators is that functionalist approaches are not appropriate to translate texts of their concern. Mullin is included
in these confirmatory references because, even though he wishes to playfully apply functionalist strategies to
translating Tibetan texts, he recognizes that in so doing his interest is not anthropology, religion or philosophy
(Mullin 1995). On a separate note, see Dorjee (1995) for a humorous example of what can go wrong if one opts for
functionalist approaches to translation in philosophy.
58
In chapter 17 of his Tractatus Theologico-Politicus Spinoza argues that in order to understand the difficult passages of
the Torah, the interpreter should keep in mind the historical particulars of the authors the text. I think this is
correct. When possible, we need to inform our translations in light of the relevant historical information, whether

405
this be archaeological, or textual. The more information that we have about the historical background of a text, the
better position we will be in to translate it. None of this implies that we do not require text-types. For instance, in
my account and translation of Indian philosophy, I bring to bear the text-type of philosophy in my account of
Indian ethics. However, my knowledge of the cultural background of the ancient Indians, including for instance
facts about their general cosmological outlook, and their way of life, as attested to by archaeological findings, is
important to explaining and rendering plausible how they could hold philosophical views that are very different
from ours on matters such as ethics. But this only takes on a significance once we have settled on questions of text-
type. (Spinoza 1998)
59
Nida provides references to many authors prior to him who voiced such functionalist leanings (cf. Nida 1964,
162-4).
60
Catford, however, avoids the following criticism by conceding to the critic that translation does not track
meaning (1965, 42). Catford’s functionalism thus does not call what is preserved in translation along functional lines
“meaning,” though it seems for all intents and purposes to track functional meaning.
61
The causal theory of meaning (cf. Monk 1990, 291) should be distinguished from an even more famous account
of meaning, restricted to proper names and nouns, proffered by Kripke and Putnam, and termed (perhaps
improperly) the “causal theory of names” or the “causal theory of reference” after Gareth Evan’s influential article
(Evans 1996 [1973]). Very closely related to the causal theory of meaning is what C.L. Stevenson advocated as the
dispositional theory of the meaning according to which the meaning of a sign is its disposition to cause certain
psychological reactions or conversely the dispositions of hearers to react in a certain way to a sign. According to
Stevenson, a dispositional account of the meaning locates more stability in meaning across a speaker than the
causal-psychological account for meaning does not vary accordion to varying psychological responses, but
according to tendencies of signs to evoke certain psychological responses (Stevenson 1944, 54-55).
62
David Kaplan distinguishes between “contexts” and “circumstances”, but the distinction seems to be
idiosyncratic, and not clearly respected by subsequent authors (Kaplan 1989, 494). For Kaplan, a “context” is the
situation of word usage, while a circumstance is a situation of word usage evaluation. This is usually distinguished in
the literature these days in terms of the “context of utterance” and the “context of evaluation.”
63
I’m not the first to notice a similarity between Quinian and Derridian arguments about meaning and translation.
Samuel C. Wheeler has written on this topic, but I find his treatment odd, forced, and selective. He too notes the
similarity, but he gets the point of contact wrong. He claims that Derrida rejects the polysemy of language (Wheeler
2000, 27). Derrida however recognizes the polysemy of concepts, particularly pharmakon (Derrida 1981 [1967], 71-2).
Derrida does seem to have an ontology of writing, however, that he says is not reducible to a polysemia (Derrida 1982
[1971]). Derrida is not consistent about his distinction between language and writing, and many times conflates the
two, and other times he differentiates the two, aligning the former with the traditional Contextualism that he
criticizes. In light of Derrida’s inconsistency, Wheeler’s gloss is hasty.
64
An exemplary case of determinate translation is to be found in the translation of Buddhist texts from Sanskrit
and Pali into Tibetan. This was accomplished by the Tibetan translator Lo-tsa-wa who made the difficult journey to
India at the beginning of the second millennium common era, where Indian gurus guided the Tibetan translation
project (Tsering 1995). The result is vast collections of the Bka’-’gyur and Bstan-’gyur. The translations are of such
quality that scholars continue to return to the Tibetan to shed light on the more obscure texts in Sanskrit. And, the
propriety of such translations continue to be the subject of critical inquiry today (cf. Tulku 1995).
65
My reading of Wittgenstein is influenced by the work of the linguist Talbot Taylor (Taylor 1997±, 74) and what
I’ve learned in a course on Wittgenstein with Prof. Stuart Shanker. On my view, Wittgenstein himself is rather
inconsistent about the role of determinate meaning in language. At times he wants to show that we do not require it,
at other times he brings in a rather strong sense of “grammar” that is supposed to solve all perplexity, suggesting
that indeed such grammar highlights a determinate meaning of language. This inconsistency arises because of

406
Wittgenstein’s commitment to the linguistic paradigm, which prevents him at times from understanding how his
deconstructive insights undermine such positive claims.
66
Lock in his Essay Concerning Human Understanding writes:
To make Words serviceable to the end of Communication, it is necessary that they excite, in the Hearer,
exactly the same Idea, they stand for in the mind of the Speaker. Without this, Men fill one another’s
Heads with noise and sounds; but convey not thereby their Thoughts, and lay not before one another
their Ideas, which is the end of Discourse and Language. (Locke 1979 [1690], III, ix, 6 )
The subject of telementation and its rather absurd presuppositions and implications are the subject of extended
discussion and criticism in Talbot Taylor’s work: (Taylor 1997±, 1992).
Taylor’s own positive view however is as extreme as the view he criticizes. Taylor correctly rejects the notion that
linguistic communication requires a shared system of meanings that are grasped before hand by interlocutors. He
however believes this implies that there are no facts about the meaning of expressions in a language. According to
Taylor, orthodox language theory regards reflexive comments as “‘folk’ hypotheses about the underlying facts of
language.” Orthodox linguistic theory thus treats reflexive statements as primitive versions of scientific propositions
(Taylor 1997a, 10). The result is that reflexive claims like “soporific means ‘tending to produce sleep’” are interpreted
as tracking something in reality—‘tending to produce sleep’—as the meaning or substance of what is claimed (Taylor
1997e). Orthodox linguistic theory, hence, regards reflexive discourse as always having descriptive content; just as
scientific claims about natural phenomenon are regarded as describing or qualifying something in reality. This is a
mistake according to Taylor: “such statements are not descriptions of states of affairs, but rather assertions, or
citations, of norms. In this case they are neither true nor false, in the sense of corresponding or not corresponding to some state
of affairs” (Taylor 1997±, 135).
According to Taylor, in fashioning metalinguistic comments in the model of proto-scientific propositions, orthodox
linguistic theory overlooks the true character of reflexive discourse: normativity. Taylor writes:
By means of “normative metadiscourse” (i.e., reflexive remarks used for normative purposes), we tell
others, or are told ourselves, how a given word “is to be” used, what a word “should” be taken to mean,
what rule we “ought” to be following, what the “correct” way to say X is, why one way of pronouncing a
word is “wrong,” what you “have to” say when someone compliments you, etc. Language users
ordinarily treat linguistic acts as something they value, as something whose characteristics matter to
them; and they convey-and enforce this attitude, in large part, by speaking of language and the
circumstances of its occurrence in normative and evaluative terms. (Taylor 1997a, 11)
One problem with this view is that it is not clear why a switch to a normative conception of language makes it any
less amenable to the type of descriptive linguistic enquiry for what the linguistic can track is not descriptive facts,
but rather normative facts, i.e., facts about what rules a community overwhelmingly believes language use should
conform to. The linguist’s job would thus be to describe these rules. Indeed, I think this is what linguists often do,
and their efforts thus can have some success, but to the extent that the phenomenon they are tracking is a shifting
target, their efforts will be undermined. More importantly, however, as linguistic meaning does not determine
translation, and as translation is the best account of the objectivity of meaning that we have, no account of linguistic
meaning can be complete, fully objective or accurate. It is not that there are not grammatical facts about a language
that a linguistic could uncover, but any such fact will be only a partial snapshot of the dialectic of meaning as it
unfolds in a community. Thus, the linguist’s account of a language will have a state of affairs to accord with, namely
the linguistic behaviour and metalinguistic speculation of speakers in a linguistic community, or in contrast
philological data, but this will always be a partial and of an indeterminate nature in matters of semantics. The
determinate factor is the text-type, but that it is outside the purview of the categories of linguistics. Texts after all
can be comprised of all manner of semiotic devices, not just linguistic devices and the text-types themselves are not
reducible to the categories of language, I argue (cf., III.1 Text-Types do not Reduce to Systematic or Literal
Features of a Language 148-152).

407
67
For an account of the Creationism vs. Darwinism debate that is still relevant today, see Michael Ruse’s edited
volume, But is it Science? (2008). It collects much historical material on this issue and transcripts from a court case
where Ruse was called on as an expert witness by the American Civil Liberties Association. The case revolved
around Arkansas State legislation that sought to include the teaching of Creationism in school science classes. Ruse
of course was on the side arguing that it should not be allowed in that forum. What is most interesting about this
debate is how Ruse had to oversimplify the case for rejecting Creationism in order to conform to the argument
being put forward by the ACLU lawyers. The ACLU lawyers were arguing that Creationism was not simply bad
science, but that it was not science at all. If it could only be established that Creationism was simply bad science,
then the US constitution would apparently allow it to be taught in high schools. If it was religion, however, it would
be barred. That it could be inspired by religion and be bad science was not a seeming option. Indeed, this seems to
be the right characterization of the thesis. In order to get the notion of “bad science” off the ground, it seems that
we need access to institutional considerations of what makes for good science. But good science in turn needs some
type of semantic grounding to be determinative. The linguistic paradigm has trouble with this I think. Kuhn’s view
is simply one example of the linguistic paradigm having difficulty accounting for the notion of institutional
standards of science that is distinct from a language. The earlier version of this difficulty can be found in the
thoughts of Karl Popper, who held that Darwinism is not a scientific theory, but a metaphysical research project
because the claims that Darwinism makes cannot be falsifiable because we cannot test the entire history of the
origins of species to see if things unfold as Darwin suggests (Popper 1988 [1977]). This muddle comes from
thinking that science is concerned with individual propositions. It is simply the atomistic correlate to Kuhn’s
holism and equally unable to explain why Darwin provides the better account. Ruse’s book has recently been
updated (Ruse and Pennock 2008) in light of a recent court case in Pennsylvania (Jones 2005).
68
TTS thus both solves and dissolves the traditional problem of the demarcation problem in the philosophy of science
(the problem of what separates science from non-science). In the linguistic paradigm the problem is unsolvable for
it aims to distinguish science from non-science in the abstract (not within texts of definite types) on the basis of the
content of scientific beliefs and theories. But as TTS and the problem of translation teaches us, words, and
sentences in the abstract have no determinate content. They are polysemous, and can have multiple significances.
Thus, for instance, scientific terminology and theories could just as easily comprise the substance of a play or a
poem as it can a scientific text. It is the text-type that we require to unlock the significance of such words and
sentences and the unlocked significance pertains to how it is to be preserved in translation. Here, the telos of the type
of text that one is translating is the only determinate factor. Thus, TTS dissolves the traditional demarcation
problem by pointing out that there is no determinate difference between “science” and “religion” if we attempt to
understand this apart from textual considerations. But it solves the demarcation problem in general by pointing out
that there are different types of texts, each with its own area of semantic concern. Science is one such text-type.
Religion, I believe, is not a text-type at all—this will be apparent by the end of this dissertation. For a review of the
problems with the traditional approach to the demarcation problem, see Laudan (Laudan 1988). For an updated and
slightly optimistic approach, see Resnik (Resnik 2000). Resnik opts for an institutional approach to demarcating
science, and I have some sympathy for this, but the textual issues appear to me to be primary, if we are going to be
able to show what science is cross-culturally and to solve the problems of translation.
69
Harris is unusual among linguists for recognizing an important asymmetry between speech and writing (Harris
2000). Yet, the distinction seems play no role in his analysis of important types of discourses like science. For
Harris, writing and speech are both linguistic and categories like “science” are linguistic categories (Harris 2005b). If
I am correct about the nature of translation, this is a mistake.
70
Davidson writes: “A metaphor implies a kind and degree of artistic success; there are no unsuccessful metaphors,
just as there are no unfunny jokes” (Davidson 1996 [1978], 416) Certainly in the case of metaphors, there are
institutional standards that are brought to bear that separate out great literature from bad literature such that one
can identify unsuccessful metaphors. If not, one couldn’t have university courses and programs devoted to training
creative writers and poets. Given the overwhelming success of some comedians relative to others, there does seem
to be degrees of success in jokes, such that some are not funny. Davidson’s linguistic orientation does not readily
dispose him to recognize such institutional standards as determinative.

408
71
Quine (1960, 77) cites a marvellous example of a case where an apparent natural kind term fails to meet the
standards of science. He cites Godfrey Lienhardt’s example of certain islanders who refer to their half-brothers as
pelicans (Lienhardt 1958). Scientists could use this term to stand for pelicans in the TL, but this would amount to a
type of regimentation of the term. Quine points out that the trouble in such cases is a trouble of the indeterminacy
of linguistic correlates. I agree. Does this reflect anything deep about difference in the thought of the islanders and
us? I think that the question is really nonsensical. The real issue is a matter of how we would translate texts with
their word for “pelican” and the issue would be settled by text-types. The notion that language maps the thought of
a people, or represents categories of thought is part of the problem of giving language too much importance, as
though it was the seat of semantic determinacy. It is the ideological move that, unfortunately, many of us are given
to.
72
For a very good review of philosophical commentaries on the topic of the relationship between philosophy and
literature, see the article “Recent Work: the Philosophy of Literature” (Read and Cook 2001). This article points to
several important contributions to this field (cf. Guetti 1993; Nussbaum 1995, 1990a, 1990b, 1986; Rajchman and
West 1985). Most interesting is their identification of two ostensibly literary works that explicitly aim at eliciting the
philosophical reflection of readers, one on the topic of animal rights (Coetzee and Gutmann 1999), and the other
that brings to light the stylistics of Wittgenstein’s Philosophical Investigations (cf. Silliman 1986, particularly his poem
“The Chinese Notebook”).
73
Gilbert Harman, has enthusiastically defended the indeterminacy of translation. See his “An Introduction to
‘Translation and Meaning,’ Chapter Two of Word and Object” (Harman 1969). Harman has also defended related
theses in the philosophy of language that cause trouble for translation (Harman 1993 [1993]). The underlying
commonality to problematic theses for translation is that meaning is simply a function of local contingencies, such
as language use or symbol usage in thought. A version of this thesis has been put forward recently under the
heading of “Conceptual Role Semantics” by Harman and Mark Greenberg: “Conceptual role semantics (CRS) is the
view that the meanings of expressions of a language (or other symbol system) or the contents of mental states are
determined or explained by the role of the expressions or mental states in thinking” (Greenberg and Harman 2006).
Greenberg and Harman argue in this article that Conceptual Role Semantics does not necessarily imply that the
meaning of an expression is its use, but only that its use explains or determines its meaning. All the same, Quinian
worries about translational determinacy arise on such an account, for there is no guarantee that expressions across
languages will have the same conceptual role. Indeed, we are most guaranteed that they do not, as the role that they
have is evaluated in relationship to their native environment.
74
Wong’s discussion of the topic of translation centres on identifying a principle to disambiguate speech acts. He
cites Quine as a leading translation theorist, but objects to Quine’s use of the principle of charity. Wong reads
Quine’s principle of charity as recommending that we choose interpretations of native speech acts that maximize
agreement between ourselves and native speakers (Wong 1984, 106). In its place, Wong recommends that we
instead adopt the principle of the best explanation: “a translation of a group's language should be such that the
content of imputed propositional attitudes, and the relations among them and the world, can be included within the
best explanation of the group's linguistic and nonlinguistic behaviour” (Wong 1984, 112). The advantage of this
approach, he argues is that “the principle permits us to use every bit of knowledge that could help in the effort to
produce the best explanation. This includes not only our psychological self-knowledge, but relevant parts of
neurophysiology, sociology, and anthropology” (Wong 1984, 113). He contrasts this approach to both Quinean and
Davidsonian approaches to translation, but Wong’s proposal does nothing to answer Quine’s basic complaint, that
if meaning of expressions are in part a response to culturally specific phenomenon, the notion that there is
something like objectivity in translation is a chimera. Differing to our best anthropology and sociology will not help
us out of this bind for they too will be saddled with the indeterminacy of translation, if Quine is right, rendering
very doubtful that they have anything objective to tell us about cultural difference. Wong’s own proposal does not
rule out alternative translations and he recognizes that there may be many “best explanations” of anthropological
behaviour (both linguistic and nonlinguistic). His contention is that any such best explanation must be consistent
with moral relativism (Wong 1984, 114). Without having addressed the deep worries about the objectivity of
meaning and translation, the consistency of moral relativism with our set of “best explanations” is no more

409
objective than apparent indeterminate translations. The problem in effect is not something that can be solved by
empirical research, if Quine is correct, for it has to do with a semantic asymmetry across languages that renders
translation indeterminate.
75
The type of relativism that TTS will show as mistaken is philosophical. A prominent version of such a relativism
is moral relativism, or the notion that the truth of moral claims or the substantive standards of morality are
relativized to locales. This variety of relativism is distinct the methodology of “Cultural Particularism” made famous
by Franz Boas. John Cook’s Morality and Cultural Difference (1999) presents a careful exploration of the relationship
between the methodology prescribed by famous anthropologists, such as Franz Boas, and the doctrine of moral
relativism. Cook’s careful investigation of Boaz’s thought and the possible formulations of moral relativism lead him
to the conclusion that the type of methodology adopted by anthropologists at Boaz’s behest does not support moral
relativism at all. However, he thinks that the careful investigation of cultural difference suggests that something like
a moral particularism is more fitting. Cook gives us more reason to conclude that the moral relativist has hastily
rushed to conclusions. Boaz, Cook notes, was much more interested in adopting a methodology that was
naturalistic than in taking a position in moral philosophy: his recommendation is that we are to study customs in
relation to a culture and not in relationship to our substantive convictions (Cook 1999, 72). My recommendation for
cross-cultural historians of philosophy is similar: we must find a way to study the evaluative thought of alien
cultures in a manner that clearly distinguishes our substantive views from those that we are studying, and so I
argued in Ethics and the History of Indian Philosophy. I also think, however, that if we have a proper count of
philosophical semantics that will allow us to objectively study the substantive views of thinkers from traditions that
are not ours (QI), we will also have the means by which we could decide moral controversies across cultures.

76
Moore accuses Bentham and Mill of this error (Moore 1903, 14).
77
Kripke writes, “Let's call something a rigid designator if in every possible world it designates the same object.”
(Kripke 1980 [1972], 48)
78
A similar project can be found in the work of David Brink who argues that we should favour Consequentialism
as a moral theory, for it can play a causal-explanatory role in the social sciences (Brink 1989, 132), though Brink
later criticizes Boyd’s argument for replacing deliberation with the causal history of term usage as determinative in
moral questions (Brink 2001). Peter Railton at many points appears to be on the same page with Boyd. However,
Railton seems to me to be far more careful about his non-analytic naturalism than Boyd. Railton holds that we
should adopt a naturalistic conception of ethics as a reformative definition and consequent of this reformation,
science can play a great role in helping us get at the truth in ethics (Railton 1986, 204; cf. Railton 1996). Boyd, in
contrast, seems to violate his non-analytic naturalism by first claiming that his homeostatic cluster property is what
governs our usage of moral terminology, though we may not know it (hence the role that science can play), but that
it all the same “defines” moral goodness (Boyd 1988, 203). There are other peculiarities about Boyd’s argument, in
my view. The strangest feature of the argument is that it assumes that mere causal governance of our moral
vocabulary absolves us from having to choose what conception of the good we should endorse. Likewise, his
argument does not seem to take seriously the possibility that we could reform moral language usage so that different
causes come to govern our usage in such a manner that over time we would come to refer to it with our moral
vocabulary and thus have new moral judgments that are true, in his sense. And, moreover, it does not seem that his
causal account has the resources to explain why we should not sufficiently reform our practices in favour of
alternative conceptions of the good. In light of recent findings in genetics, moreover, Boyd’s notion that there is
something like a common human good, naturalistically defined, may not be true, or far too meagre to ensure the
type of welfare he is interested in. See Freeman (2006) and Shianna and Willard (2006). I don’t see that any of these
concerns are addressed in Boyd’s recent work on ethics (cf. Boyd 2003a, 2003b). Brink proposes an alteration. On
his account we should understand the causal constraint so that it is mediated by a deliberative constraint: “…a
natural property N causally regulates a speaker’s use of moral term ‘‘M’’ just in case his use of ‘‘M’’ would be
dependent on his belief that something is N, were his beliefs in dialectical equilibrium” (Brink 2001, 169). This may
avoid some of these problems that I identify if it is successful. But, the problem with this account is that one’s

410
beliefs may be in dialectical equilibrium, and moreover one may have the belief that something is N, but N may not
play any actual role in causing one’s beliefs. This is most blatant in the case of false beliefs about N. I and my tribe
may believe that the volcano dubbed “Ughu,” visible from our village, is an angry God who demands human
sacrifices once a year. Moreover, our beliefs are in dialectical equilibrium, given our cultural mores and what
traditional, scientific knowledge we have. Yet it would be difficult to establish on naturalistic grounds that the
natural property N, namely the Volcano protruding from the landscape, causally regulates our use of “Ughu” when
me and my tribe say that Ughu demands a human sacrifice. A similar criticism of Brinks’ fix can be found in Heimir
Geirsson’s (2005, 367-368), but with a different example. Naturalist moral semantics has a fair share of problems,
but the most sure reason to reject it is for semantic reasons, I believe.
79
Lumping McDowell in with Naturalists of Boyd’s variety may seem odd because McDowell wishes to stress the
great variability and flexibility in our response to natural phenomenon. He is famous for arguing that moral
properties, though supervening upon natural properties, do not reduce to patterns at the primary quality level
(McDowell 1981, 144-145). I shall argue in the next section that their positions are not that different, and moreover
the linguistic assumptions of both Boyd’s and McDowell’s account land them into trouble, regardless of their
metaphysical assumptions about moral semantics.
80
He writes: “The right response to the claim that all our assessments of truth are made from the standpoint of a
"conceptual system" that is inescapably our own is not to despair of our grip on reality but to say, with Hilary
Putnam, "Well? We should use someone else's conceptual system? It is pointless to chafe at the fact that what we
believe is what we believe” (McDowell 1998a, 128). Elsewhere, he affirms the same relativism with less gusto but
with just as much trouble: “But, without abandoning a fundamentally Aristotelian outlook, we can let the question
arise whether the space of reasons really is laid out as it seems to be from the viewpoint of a particular shaping of
practical logos. What we must insist is that there is no addressing the question in a way that holds that apparent
layout in suspense, and aims to reconstruct its correctness from a vantage-point outside the ways of thinking one
acquired in ethical upbringing. This allows for radical ethical reflection, as Aristotle himself seems not to. But, like
any reflection about the credentials of a seeming aspect of logos, this reflection must be Neurathian; we cannot
escape the burden of reflective thought-the obligation to weigh, by the best lights we have, the credentials of
considerations purporting to appeal to reason-by a fantasy of having some suitable first-natural facts force
themselves on us in a way that would bypass the need for thought” (McDowell 1998b, 189). The trouble for
McDowell, and all proponents of linguistic or cultural accounts of ethical knowledge, is explaining how persons like
the Earthlings and Twin Earthlings could be thought to be in the same boat that one alters in Neurathian fashion.
Textual institutions provide us this (and they also allow us not to worry so much about altering our boat in
Neurathian fashion, if the boat is the textual institution), but it provides us this common ground by deprivileging
the role of upbringing or the concepts furnished by one’s language (whatever they are).
81
Indeed, Stevenson’s requirement for disagreement in attitude is stronger than the way I described it, and makes it
impossible for people who have never met: “Two men will be said to disagree in attitude when they have opposed
attitudes to the same object, one approving of it, for instance, and the other disapproving of it, and when at least
one of them has a motive for altering or calling into question the attitude of the other” (Stevenson 1944, 3). Here,
the problem is not that the two parties must coordinate their attitude relative to one item (which may not happen in
cross-contextual circumstances---it certainly wont happen on different planets) but that one party must have a
motive for altering or calling into question the other party’s attitude, which seems only possible if the one party
knows of the other party.
82
The classical articulation of this view is in Rawls’ “The Independence of Moral Theory” (Rawls 1975) Moral
theory, for Rawls, is distinct from moral philosophy, and it concerns peculiarly normative matters that metaphysics
and semantics have little to add. Rawls is both correct and incorrect. He is correct to the extent that the semantics
of moral concepts is not directly reducible to metaphysical or semantic concepts—this QI supports. However, he is
clearly incorrect for thinking that moral theory does not itself concern metaphysical and semantic issues, for the
bulk of what is in a moral theory against which the world is to be judged just are matters of metaphysics and
semantics. Rawls of course was quite blind to this, and it allowed him to sneak into his picture of ethics an explicit

411
anthropocentrism that clearly takes a stand on the metaphysics of moral agents. Dogs and hibiscus plants have no
place in the original position, for Rawls. If it is possible, behind the veil of ignorance, that some of us might end up
being pigs, I’m sure vegetarianism would have been a major argumentative end of a Theory of Justice. I for one would
have applauded this line of argument. Rawls recognizes that his account of justice as fairness leaves out questions of
our duties to animals, but this admission does nothing to attenuate the metaphysically skewed manner in which he
approaches issues of justice (Rawls 1971, 15). More recently, Putnam seems to come close to arguing that
metaphysics is irrelevant to ethics in his Ethics without Ontology (2004).
83
The view I am criticizing here is different from the one I am arguing for. I am arguing that there is such a thing
as institutional expertise that can lead us to moral knowledge. However, on my account, such institutional expertise
need not be something that people at large defer to in their moral deliberations any more than people at large in the
US defer to scientists in their deliberations on the origins of the universe. On Boyd’s account, however, people at
large in a society must defer to such experts to mediate their moral beliefs in order for there to be an objective
referent that causes their moral beliefs. Objectivity on my account is truly institutional, whereas on Boyd’s account,
it is a matter of sociolinguistics.
84
I have Professor Henry Jackman to thank for the example of “happiness” and “eudaimonia” as highlighting the
contentions of the ethical naturalist.
85
Of course, there is a deeper story about genetic diversity in the human population that is now being uncovered,
but there is no indication in these new findings about genetic diversity that such diversity correlates easily with
cultural groupings. See Freeman (2006) and Shianna and Willard (2006). No doubt, such differences at the
individual level will contribute to some differences in what convictions people are likely to hold, but genetics does
not obviously determine every fait that awaits a person. Some persons who are genetically predisposed to heart
disease can alter their life style and avoid heart disease, for instance. People’s ideas and beliefs has some role to play
in our explanation of their practices.
86
The issue has come under discussion by several authors (DePaul 2000; Kornblith 1998; Goldman and Pust 1998).
See also work in “Experimental Philosophy” (Weinberg, Nichols, and Stich 2001).
87
If such a matter as European heritage should seem to be irrelevant to the philosophical views we hold, we can
turn to a wonderful study in experimental philosophy conducted by Weinberg, Nichols and Stich. They designed a
survey of undergraduate students at Rutgers University. The students were presented the famous counter examples
divined by Ed Gettier to the now standard account of knowledge as justified true belief, where it turns out that a
believer has a true belief by pure luck, with no causal connection with what makes the belief true (Gettier 1963).
Weinberg, Nichols and Stich found that students of Asian descent tended not to have the intuitions that Gettier was
seeking, while the students of European descent tended to have them. And, moreover, the students in question
were all native English speakers. Equally interesting was the finding that people of South Asian descent differed
even more greatly than East Asians to people of European descent (Weinberg, Nichols, and Stich 2001). Brilliant! It
should thus come as little surprise that I (as someone of South Asian descent though a native English speaker born
in Canada) don’t have Gettier intuitions. Here’s one possible explanation. In the Abrahamic traditions, what
legitimizes a message from, or an incarnation of, God is a causal connection to God. Thus, regardless how bizarre
or ridiculous the claim (as my friend, the Swedish Sanskritist, Martin Gansten puts it, there is only one completely
blameless person in all of history—the son of God, who is God, but different, it’s a mystery—and that person had
to be tortured slowly to death for everyone’s benefit), we are to take the message as veridical because it comes from
God in high Heaven. In the Asian tradition, there is no obvious or exceptional causal-connection requirement for a
message to be divine, or for someone to be an incarnation of the Buddha or Vishnu. All that matters is that the
message in question, or the putative incarnation in question, have all the necessary and sufficient conditions for the
message or person to be divine This is why there is no sensible or easy line to be drawn between religion and
philosophy in Asia (both in China and in India), though it has proved useful in the West. In the Asian context,
divinity itself becomes a matter for philosophical scrutiny. However, the Western tradition, for accidents of history,
has gone down another path. Perhaps this has been fortuitous, as it may in part be responsible for the rise of
modern science as we know it in the West (for the causal requirement turns our attention to causes, and a desire to

412
weed out lucky guesses), and, ironically, made the Abrahamic religions intellectually without a home in science or
philosophy. But, to make a long story not so long, I don’t share something that is supposed to be foundational with
most of my colleagues, namely the intuition that the causal origins of convictions always matter in the assessment of
knowledge. This is not the only “intuition” on which I tend to defer from many of my colleagues. However, the
Moral Twin Earth argument, and much work in moral semantics, is driven by appeals to intuitions. But really, I
don’t think we should or need to be appealing to such intuitions in philosophy—unless we are hell-bent on
maintaining a certain cultural exclusivity to the exercise. To date, it has seemed that we have little else to go on for
failing an appeal to intuitions, there is no foundational input for philosophical thinking. I believe that my text-type-
theoretic account provides us an alternative way to do philosophy, which I outline in the final chapter. Contrary to
the prevailing doom and gloom in the field, the approach I advocate can show us a way out of relativism.
88
An excellent review essay of the accomplishments and developments in formal semantics is by Barbara Partee
(1997). What is apparent from her essay is that the developments in this field are purely technical. Perhaps most
amusing is the state of recent formalist attempts to address the question of meaning, distinct from syntax.
Apparently, formal semanticists think that headway can be made by specifying the properties of items referred to by
words. I wish them luck: they will have a rude awakening when they attempt to address the semantics of moral
vocabulary and stumble unwittingly into a quagmire of controversy that has been ranging for over a hundred years
in the wider philosophical tradition. Reading Partee’s article, I get the sense that formal semanticists have a great
sense of accomplishment, for little reward, and usually what is accomplished is a very arduous re-invention of the
wheel. Montague’s grammar, which was supposed to demonstrate that English is a formal language, only
accomplishes this task by holding that there is no strict equivalence between English language sentences and the
formal language (Montague 2002 [1973]). If it were not for the great difficulty in accomplishing such formal systems
to do so little, my sense is that formal semanticists would have no sense of accomplishment at all.
89
The bigotry of the major figures of the linguistic turn is not often enough discussed. For information on Frege’s
radical right-wing politics, and his anti-Semitism, see Michael Dummett’s admirable Preface to his Frege’s Philosophy of
Language (Dummett 1973, xii). Heidegger’s involvement in the Nazi party is well known (cf. Ott 1993). What is less
known is Wittgenstein’s self-loathing as a Jew. Wittgenstein presents us with the classic case of a victim of racism
who has internalized racism. This is most evident in his Culture and Value (1980). German speaking philosophers do
not have the market cornered on racism. Indeed, we can find this type of thing in Hume as well (cf. Ten 2002). But
it is these three Germanic philosophers who have had a profound effect on the very notion of conceptual analysis
in the twentieth century on both sides of the Atlantic. It is thus odd that the connection between their intolerance
to diversity and their optimism about our ability to reflect upon our own thoughts and our culture’s language use as
a means of yielding conceptual truths has not been noticed. I take my pluralistic institutional approach to semantics
to be a corrective and check against such pernicious forces. Conceptual analysis is not out of the picture, but one
that defers to culture and language is.
90
Allan Gibbard’s ambivalence is a recent change of heart. In his earlier Wise Choices, Apt Feelings, he rejected the
notion that moral sentences could be true or false (Gibbard 1990, 8-10, 92).
91
Gibbard underwrites this view by distinguishing between concepts, some of which maybe non-natural, and
properties, all of which are natural on his view. Some natural properties can be understood as being a property
signified by a non-natural concept, such as moral concepts, without the property itself being non-natural. As
Gibbard now understands moral concepts as essentially plan laden, calling a natural property GOOD is in effect to
recognize it as the thing to be desired (Gibbard 2003b, 6, 22, 32-6, 43). A moral property, on this account, is not the
name of a type of property, but rather the identification of a property as important relative to our interests.
92
See Mark Schroeder (forthcoming).
93
Frank Jackson and Philip Petit in an influential paper have argued that Expressivism fails to deliver on its anti-
descriptivism. I think that their characterization of Expressivism is a bit unfair. They fail to recognize the quasi-
realist nuance of genuinely Expressivist positions in the recent literature and rather reduce it to the crude
Emotivism of Ayer’s Language Truth and Logic that outright rejects the cognitive aspects of atomic normative

413
sentences (Jackson and Pettit 1998, 239-240). There is an important connection between non-cognitivist versions of
Expressivism and Emotivism (see footnote 95) but there are also important differences. For a response to Jackson
and Pettit on behalf of the Expressivist, see Schroeder (forthcoming).
94
While I take the terminology of “mentalism” from Schroeder (forthcoming), he argues that a feature of the
Expressivist doctrine of mentalism is that it also interprets the content of descriptive propositions corresponding to
beliefs in the same way. The argument that I shall develop in the latter part of the paper will not depend upon
indicting this feature of Expressivism.
95
A notable feature of the earlier “non-cognitive” versions of Expressivism is its kinship with Emotivism and the
notion that affect can semantically anchor normative concepts. Early on, in 1980, Blackburn tells us that he takes
“an emotivist starting point: we see the meaning of moral utterances as essentially exhausted by their role in
expressing the speaker’s attitude. I have argued elsewhere that a surprising degree of quasi-realism is consistent with
that view” (Blackburn 1993 [1980], 14). Nearly twenty years later, Blackburn’s view is not so different, though he is
careful to emphasize that ethics includes “the full dynamic range of our practical natures” (Blackburn 2000 [1998],
18). While Gibbard appears to emphasize the prescriptivist aspect of the recent non-cognitivist tradition in his view
that ethical concepts are essentially plan laden, and concerned with the “thing to do,” he follows Ewing in analyzing
“good” in terms of desire or what is to be sought (Gibbard 2003b, 22, 142). One possible differentia of the newer
“cognitivist” variety of Expressivism may be its move to shun such affective analyses of ethical concepts.
96
This is, in a nutshell, Gibbard’s view of translating normative discourse (cf. Gibbard 2003b, 166), though
Gibbard’s view contains the additional complexity of a specialized planning language that would putatively mediate
the translation of natural languages (Gibbard 2003b, 141). Earlier, R.M. Hare articulated such a picture without
appeal to a specialized language (Hare 1952, 146-9). Gibbard’s view also contains a nuanced account of thick
concepts that blend evaluation and description in such a way that it is possible for their expressed content to be so
out of sink with our own planning theories that we can neither agree nor disagree with the judgements expressed
with such concepts, but merely demure from making them ourselves. Gibbard seems to suggest that such
judgements are untranslatable (Gibbard 2003a, 2003b, 170). The argument that I will make against Expressivism’s
ability to translate normative discourse is broader, and centres around what I call Expressivism’s “linguistic”
orientation. As I shall argue, Expressivism cannot account at all for the translation of normative discourse.
97
In the translation theory literature it is now customary to defer to the tripartite account of translation presented
by Roman Jakobson. According to Jakobson, translation might proceed intralingually (via a process of rewording,
which involves conversions of signs into other signs within the same language), interlingually (across languages), and
intersemiotically (across symbolic systems, such as in the case of translation from a full language into a narrow system
of icons). Jakobson thought that only interlingual translation was “translation proper.” (cf. Jakobson 1992 [1959])
My sense is that an adequate theory of translation will be able to account for all three varieties of translation, and I
believe that the one I shall present does. However, for the purposes of this paper, I shall focus predominantly on
the linguistic versions of translation and assume that we are primarily concerned with interlingual translation.
98
A similar point about the loss of reflexivity in translation was later made by W.D. Hart in his observation that
one cannot preserve self-reference, reference and truth-value simultaneously in a translation. Hence, if the sentence
“The first word of this very sentence has three letters” were translated metaphrastically into French, it would be
false, for the first word of the sentence would have two letters (Hart 1970).
99
Quine claims that he never intended his thesis of the inscrutability or indeterminacy of reference of words,
exemplified by his famous example of “Gavagai” to highlight the problem of the indeterminacy of translation.
Robert Kirk shows that these theses are more closely related than Quine thinks (cf. Kirk 1986, 106-132).
100
Quine’s indeterminacy of translation thesis is supposed to apply just as much to domestic, intralingual translations as
interlingual translations. If resources of a language are defined relative to their role in a native language, then any
two resources in that language will have a unique relationship to the rest of the language, making intralingual
equivalence questionable as well.

414
101
See for instance Pym (2004, 58) and Madsen (1997, 287) for more on the sour reception the notion of
equivalence receives in translation studies these days. The trouble is not that the notion of semantic equivalence is
alien to the project of translation, but most theorists are quite confused about how to achieve this. And thus, even
those who rail against equivalence in translation end up affirming some sophisticated account of what equivalence
in translation amounts to (cf. Holmes 1988, 37, 43 fn.10). Cooler heads have noticed that there are many different
criteria of equivalence (Koller 1989 [1979], 100-1). Some effort has been made to understand equivalence in terms
of text-types (Neubert and Shreve 1992). But unfortunately the semantic importance of this move has not generally
been grasped, even though it is granted in the recent literature that different projects of translation require different
translation methodologies. It is as though translation theory continues to operate under what I identify as the
linguistic approach to meaning.
102
Translation theorists who endorse functional translation tend to be a bit indiscriminate in their enthusiasm. The
original impetus for this approach derives from the British Functionalist school of linguistics, which takes meaning
on the whole to be best understood by reference to its affective and effective role in the social and conversational
context (Firth 1964 [1930], 110; cf. Halliday and Hasan 1989 [1985]). The position will be familiar to philosophers
of language who will recognize it as the position advanced by Bertrand Russell in his Analysis of Mind (1921), C. K.
Ogden and I.A. Richard in their The Meaning of Meaning (1923) and much later by C.L. Stevenson (1944, 54-55).
Wittgenstein himself entertained this approach in the Blue and Brown Books, but later came to appreciate the
problems it poses. In his unpublished notes, he writes “If I wanted to eat an apple, and someone punched me in the
stomach, taking away my appetite, then it was this punch that I originally wanted” (quoted in Monk 1990, 291).
Function cannot be the same as meaning, all the time. However, good cases can and have been made for regarding
the aesthetic and affective function of language as essential to literature and it is in this limited role that it seems
quite unassailable. For over enthusiastic endorsements of functionalist translation methods that make a good case
for its applicability in the literary context only, see Vermeer (1989) and Bassnett (Bassnett 2002, 26-7). Eugene Nida
has been famous for advocating a functional approach to translating the Bible (cf. Nida 1964, 1997), but his
justifications for applying this methodology in the case of religious texts can only be understood as underscoring a
desire to make them more palatable to target audiences, which may not be the most honest approach or the most
appropriate method to think about meaning in the context of religious texts.
103
I take the idea of text-types from the translation studies literature (cf. Holmes 1988 [1972], 74-76; Laviosa-
Braithwaite 2001, 277-278; Neubert and Shreve 1992). It is now a permanent part of the translation studies
literature and is especially exemplified in the space the discussion has made for “specialized” or “technical”
translation (cf. Hann 1992; Wright and Leland D. 2001 [1993]; Vermeer 1989; Pym 2004; Pinchuck 1977; Maier and
Massardier-Kenney 2001 [1993]; Madsen 1997; Sarcevic 2000).
104
The first translation theorist to draw attention to the fact that translation is best thought of not as an effort to
match up lists of words and sentences with one language against another was likely Eugene Nida who
recommended that we think of translation instead as a process of decomposition and reconstitution (Nida 1964,
68). Nida’s work preceded the important theoretical innovation of the text-type and thus his discussion of this
process is not made with any explicit reference to the plurality of semantic modalities that translation must distil
from STs, relative to translational projects. Translation theory would have to wait for some time before that type of
subtlety was articulated (cf. Koller 1989 [1979]). Nonetheless, our earlier important works of translation theory are
articulated around the fact, apparent to any translator, that there are several different translational objectives one can
work around, each aiming at a distinct type of equivalence (cf. Dryden 1992 [1680]). In each case, the process of
translation is a type of filter of semantic modalities relative to a translational objective.
105
I have heard it said that Kant is easier to understand in English than in German, but this is undocumented
hearsay. There are however well known cases where translation of texts has improved intelligibility. One case is that
of the translation of early Buddhist philosophical texts into Tibetan. Scholars continue to consult the authoritative
Tibetan translations, produced jointly by medieval Indian and Tibetan scholars, as they shed light on otherwise
difficult to understand Sanskrit texts (cf. Tsering 1995).

415
106
We could think of the various translational anchors of term usage as elucidated by Prototype Theory, explicated
by George Lakoff in Women, Fire, and Dangerous Things (1987), especially in its account of “radial categories.” Each
radial category has a relationship with the central model that is not equivalent with the entire category, but provides
a type of conversational and conceptual anchor to various radial extensions. Mother would be an example of a
central model, defined by certain properties, and various radial categories, such as FOSTER MOTHER, and WORKING
MOTHER can be understood in relationship to the central model. From the perspective of translation, the
relationship a radial category has with the central model is not necessarily important. The significance of that
relationship is constrained by the text-type, and quite often what is important is not its relationship to the central
model at all but rather what the radial category does in the text relative to the text-type. Each such radial model takes
on an independent importance of its own from the stand point of the translator, endowed with semantic properties
that can only be understood in terms of its function within the text being translated, or being created in translation.
We should be surprised if translation did not sever language in this manner, for if it had no way to abstract some
features as important and relegate others as inessential, translation would be impossible, given the uniqueness of
each language.
107
Thomas Kuhn’s view, in the postscript of the second edition of the Structure of Scientific Revolutions (1970) is that
members of differing scientific paradigms should be understood as members of differing linguistic communities and
thus the problem of communication between them should be understood as a problem of translation (Kuhn 1970,
175, 205). Kuhn makes the common category mistake prevalent in writings on translation, which consists of
thinking of translation as a matter of pure linguistic conversion. This is actually not translation at all, but
simultaneous interpretation (Gile 2001). Scientists of differing paradigms do not necessarily speak a different
language. They often speak the same language, and they can author texts in the same language, but call upon
different technical terminology to articulate their differing scientific theories. The problem of cross paradigmatic
communication is not a problem for translation at all, but one of attempting to understand a theory that one does
not agree with. This is of course no big problem: philosophers do it all the time without having to become
translators. Kuhn finds himself in the position of explaining cross paradigmatic communication by his reliance upon
Quine who he cites as the expert on translation (Kuhn 1970, 202 fn.17).
108
TTS can take on board a fair bit of traditional views on the semantics of words, including the notion that the
semantics of natural kind terminology is informed by the knowledge and practices of experts (cf. Putnam 1975).
109
In their polemic with Contextualist semanticists, Herman Cappelen and Ernest Lepore write: “To use the
dubious metaphor of language as a tool for a moment: if words are tools, then they had better be pretty easy to use
because they don't come to us with instruction manuals and even if they did, there would be no time for us to
consult these instruction manuals when we're steeped in the middle of a fast and furious conversation” (Cappelen
and Lepore 2005, 112). They believe that their theory explains how it is that we can engage in such fast and furious
conversations. On their account speakers of English understand the minimal content of a proposition “semantically
expressed” by all utterances of the sentence, and this is set out in a T sentence: e.g., “Rudolph has a red nose” is
true if and only if Rudolph has a red nose. The latter is the “minimal content” of the sentence (Cappelen and Lepore
2005, 152). However, Cappelen and Lepore have no story to tell about how speakers go from this minimal content
to understanding conversational implicature of such utterances in contexts. On their view, utterances of sentences
express many propositions and they call this their theory of Speech Act Pluralism (see their chapter 13). TTS can
explain how we manage to draw the right types of inferences about what people say in certain specialized contexts,
and this is because we are able to treat utterances as virtual texts. TTS is thus an alternative to Contextualist theories
of meaning that explain conversational implicature by contextual factors. TTS cannot be a Contextualist theory of
meaning because it takes texts to be the primary bearers of meaning. The context that a text appears in does not
alter its meaning, for this is a matter that is fixed by text-type institutions that are not bound by any particular
cultural setting.
110
These are merely our contemporary manifestations of text-type institutions. In time past and in other cultures,
text-type institutions would have had different presentations, and would likely have been attached to institutions
such as state bureaucracies and monasteries.

416
111
It is not an implication of TTS that we always manage to produce translations that are in accordance with
institutional standards for text-types. Many times we fall short of the mark, and this is particularly the case when
translation is relegated to translators who themselves are not conversant with the institutional standards of a text-
type.
112
I’ve chosen to focus on Plato in my examples because of his place in the tradition and the lack of controversy
surrounding the non-naturalism that he articulates through Socrates in the Republic. What may be controversial is
that Plato himself held the view he puts in the mouth of Socrates. What is not controversial is that the Republic
argues for non-naturalism, through Socrates, and through various devices such as the Divided Line. Any acceptable
account of moral semantics should be able to explain how we can translate Plato’s text and the position we thus
conventionally attribute to him. If it fails at this task, there is something very wrong with the position. But other
philosophical positions and contrasts will do. We could contrast a conservative Christian view that the human
foetus has an inherent property that demands our respect, versus a deflationary view that argues that we should
simply not abort human foetuses. I suspect we could also draw upon Kant, and contrast the traditional Kant, heavy
in metaphysics, with a deflationary position that regards the metaphysics as mere noise. The translational contrasts
between such pairs will be difficult for Expressivism to handle.
113
Blackburn explains “good” in terms of “admiration”(Blackburn 2000 [1998], 69-70). Gibbard similarly explains
it as “desirable” (Gibbard 2003b, 16-17).
114
Gibbard writes, “Tweedledum rationally forms all the right convictions of what he ought to do, and he reliably
acts on them. His brother Tweedledee, though, is a convinced egoistic hedonist: he thinks that one ought always to
do what most furthers one's own prospects for pleasure and lack of displeasure. This is not the rational view to
have, let's imagine.… [Some] take Tweedledee's disagreements with his brother at face value: the two, they say,
disagree sharply and systematically on what a person ought to do. I myself think there's not much difference
between them: they agree remarkably in thinking their way to decisions, and disagree only on what to say about it
and on the words with which to think about it. Their disagreement is verbal; they disagree on what words to mouth.
They have no serious difference between them on what to do and why.” (Gibbard 2003b, 12-13, my emphasis)
115
Blackburn writes, “Minimalism seems to let us end up saying, for instance, that 'kindness is good' represents the
facts. For 'represents the facts' means no more than: 'is true'. It might seem, then, that our investigations have ended
up with a position remarkably like that of Moore. The ethical proposition is what it is and not another thing; its
truth means that it represents the ethical facts or the ethical properties of things. We can throw in mention of
reality: ethical propositions are really true … Or, really true, or really factually true, or really in accord with the
eternal harmonies and verities that govern the universe, if we like that kind of talk. We can add flowers without end: 'it is
good to be kind to children' conforms to the eternal normative structure of the world. For this means no more than
that it is good to be kind to children. And rather than saying that we hold that it is good to be kind to children, we
can if we like say that as we hold that it is good to be kind to children our minds are in harmony with the eternal
normative order of things. For this just means the same.” (Blackburn 2000 [1998], 79, my emphasis)
116
Prior to Kaplan, the most well known account of indexicals was likely Hans Reichenbach’s “token reflexivity”
account (Reichenbach 1947). Like Kaplan, Reichenbach regarded indexicals as governed by semantic rules that
specify their referents relative to contexts, but his account dealt with tokens of indexicals—specific indexical
words—while Kaplan’s account deals with types. For a useful introduction to this difference, and a defence of
Reichenbach’s approach, see Garcia-Carpintero (1998) who defends Reichenbach’s account against the charge that
token reflexivity cannot account for synonymy between indexicals. For our concern, this debate matters little, as the
resources that Kaplan has to offer those interested in value semantics are limited to aspects that a token-indexical
account will find unproblematic.
117
In everyday conversation—i.e., the realm of indeterminate meaning—we translate out of statements the
occurrence of a term governed by the character of “I”. This is what we do, for instance, when we attempt to derive
a sentence that more literally captures the so called “proposition” of a sentence employing “I”. For instance, “I
am hungry” said by me can be translated as “Shyam Ranganathan is hungry” or even “hungry is the author of this

417
text” or perhaps even “the doctoral candidate is hungry”. Philosophers typically argue these days that the
controversy over which is the proper translation can be resolved by deferring to the context of utterance—this, I
take it, is what is at stake in attempting to determine the propositional content of an utterance. However, contexts
cannot be translated. Only texts about contexts can be translated. The failure to appreciate the textual dimension of
translation thus renders rules about the translation of words in isolation—i.e., their definitions, or rules for linguistic
use—indeterminate in force. Once we appreciate the textual dimension in translation (and the truism that texts
must be translated according to type), appeals to context as the determiners of translational sense can be seen for
what they are: ineffectual.
118
If autobiography constitutes a text-type, a case might be made that “I” is the inscriptional aspect of the text-type
feature for at least one text-type. I however have my doubts that autobiography constitutes such a text-type, for it
seems to be a special case of the biography. Pronouns would be important from a translational perspective in such
texts, but their importance would shift depending upon the character of the biography.
119
Dreier might argue that this is not quite true, for a spring board for articulating his theory is a dissatisfaction
with Gibbard’s view that propositions are sets of possible worlds (Gibbard 1990, 96). Dreier was correct to note
that this is an unsatisfactory account as propositions are far more fine grained than sets of possible worlds (cf.
Soames 1988 [1987], 199). Dreier’s solution is to understand moral judgments not as functions of norms to possible
worlds, as Gibbard seems to have, but as functions from norms to ordinary propositions (Dreier 1999, 562). The
character of value terms is this function on Dreier’s account. While propositions, and their nature, may have
something to do with semantics, it is clear that considerations that motivate Dreier’s transformation of Gibbard’s
Expressivism have nothing to do with general questions of meaning, but merely overcoming a problem from
Gibbard.
120
David Lewis (1979) may have been the first in recent times to clearly offer such an indexical account of
knowledge of epistemic and moral concepts.
121
Benjamin Lee Whorf’s classic paper on the thesis of linguistic relativity is instructive here. Whorf argues that
languages, which he conceptualizes in a Kantian manner as “systems”, furnish speakers with a framework that
exerts a powerful influence on their ability to conceptualize alternative metaphysics. His focus is on the Hopi
concept of TIME, in contrast to the Western or European. He argues that there are very many differences between
these two linguistic frameworks that make the European objectification of time into a quasi-spatial dimension
incomprehensible to the Hopi (Whorf 2001 [1939]). Yet, Whorf is able to translate the Hopi terminology for time
into English to his satisfaction, all the while violating the linguistic expectations of his native language and the
language of his scholarship. If such linguistic expectations really set limits upon intelligibility, this should not be
possible. Similarly, if ordinary language really did provide us with determinate criteria of semantics, we would not be
able to understand alien views expressed through its means. However, we can, as when we are presented with the
claim that Buddhists conceive of reality as comprised of ethicals. The problem that we encounter is not one of
understanding the claim, but of a kind of cognitive dissonance. Davidson has provided a classic refutation of such
incommensurability theses, in his argument that the very fact of comprehending a putative disagreement between
one’s own way of seeing things and the alien perspective places the alien perspective within the wider context of
one’s own view on things (Davidson 2001 [1974]). Interestingly, his advice on how to make sense of disagreement
involves an implicit reduction of cognitive dissonance, in so far as it involves interpretively enlarging the shared
ground against which the disagreement is compared. Translation of philosophy cannot literally lay bare such an
expanded agreement, though the instruction and teaching of novel philosophical views will no doubt rely upon this
strategy.
122
The question of how to distinguish moral claims from other types of evaluative claims was the subject of much
discussion at an earlier time, under the heading of “metamorality.” In the topic of axiological or philosophical
differentiation, the question of the differentia of ethical and moral theories is the one issue that has received a fair
bit of commentary in the literature, at least at one point. While this question is no longer hotly contested, the earlier
debate has left us with many theoretical options. Such options include the view that ethics pertains to important or
overriding norms (Cooper 1970; Hare 1971, 169; McDowell 1995 [1978], 26)—views that could be called the

418
Importance Thesis and Overridingness Thesis. Another related pair of theses regarding the nature of ethics and the
meaning of “morality”—the Universalizability Thesis and Categoricality Thesis—holds that the moral pertains to
norms that are universalizable (Hare 1955), or those that are categorical (Kant 1956 [1785], 82). According to what
might be called the Conduct Thesis, ethics is concerned solely with conduct, in so far as it can be right or wrong
(Dewey and Tufts 1929, 1; Seth 1928, 1; McKenzie 1929, 1; Dasgupta 1961, 4). A similar view is that morality or
ethics pertains only to (pro) social matters, which is the Social Content Thesis—a view championed by William
Frankena (1973, 6-9). There is also the Blame Inclination Thesis that holds that morality or ethics has to do with
matters that we are inclined to blame people for (Skorupski 1993, 126); the Conformity Thesis that holds that
morality is something that there is or that we desire conformity on (Sprigg 1970 [1964], 137; Hart 1961); and the
Punishment Thesis that holds that morality concerns matters we wish punishment to enforce (Mill 1965, 246 vol.x;
Sprigg 1970 [1964], 137; Hart 1961). Simon Blackburn appears to favour the Punishment thesis (with an emphasis
on negative emotions) in his account of the distinguishing character of ethical claims (cf. Blackburn 2000 [1998], 13-
17). Rhetorically, the winner in this debate has been the social content view, championed by authors such as
William Frankena. However it is a pyrrhic victory. Indeed, the social concern about ethics and moral philosophy has
dominated the academic study of ethics in the West for centuries and is well instanced in the trend of the recent
literature in ethics. This position has the significant disadvantage of not accounting for individualized ethical
perspectives, according to which ethics concerns the examined life. W.D. Falk argued that the social conception of
ethics and the individual, introspective varieties of ethics are artefacts of the two originally distinct traditions that
blended in the West: the Judeo-Christian tradition and the Greek tradition (Falk 1963). Some authors have seen the
clash between these two traditions as so pronounced as to suggest that what we have here is a distinction between
morality (the social, other regarding) and the ethical (introspective, rational life perspective) (cf. Williams 1985, 6).
The notion that ethics/morality pertains to normative theories of social significance, though very broad, is the
common ground between these competing trends in our text-type institution of philosophy. It can naturally
embrace both introspective and extraverted ethics, both virtue theory and deontic morality, through an essential
concern for the social, even if this registers solely in terms of a concern that one be viewed by one’s peers in a
certain light. It helps show that Socrates, Aristotle and Kant are, in their various ways, moral philosophers. On the
full view that I am advocating, mere social significance is not sufficient for a theory to be moral in nature. It must
also be normative in nature.
123
Under discussion here is not the difficult or controversial case, but rather the uncontroversial case that is
underrepresented in our theorizing about axiological differentias. The problem of difficult or controversial cases, I
believe, is resolved differently, in a conservative manner. Failing such conservatism, there is no sense to the idea
that institutional norms set standards for translation. If we are in doubt as to how to subsume a certain
philosophical theory or tradition within the wider institutional practice, we look for normative direction from the
uncontroversial cases to set the conditions under which the novel or difficult cases are accepted. Thus, for instance,
in my research into the place of moral philosophy in the Indian tradition (a controversial topic) I looked to the
Western moral philosophical tradition in search for a common, axiological differentia that underwrote moral
discourse in the Western tradition (a relatively uncontroversial topic in the sense that no one doubts that moral
philosophy is well represented in the West’s tradition). I then extended this definition of moral semantics to the case
of the Indian tradition and was able to argue that “dharma” has the same meaning as “ethics” or “moral”
(Ranganathan 2007a). If we are open and honest in this endeavour, and do not attempt to rig our account of the
axiological differentia of a field of philosophy so that it favours our substantive goals in normative ethics or
philosophy, radical outcomes are possible that are surprising given the conservative methodology that gives rise to
them. The reason that radical and inclusive outcomes are possible given this conservative methodology is that there
is often far more theoretical diversity in any given philosophical tradition than is often appreciated.
124
The account presented here does leave open the possibility that we could understand the truth of normative
theories in terms of their ability to measure up to the totality of reality in its full metaphysical and normative glory
(inclusive of empirical contingencies), which we may also call “the world.” This manner of grounding the truth of
value theories and claims would be mediated by what I call the “objectivity clause.”

419
125
As in the case of moral theories, there is a sense in which metaphysical theories are judged by their ability to
capture the totality of reality: this is mediated through what I identify as the “objectivity clause.”
126
The model of translation that I argue for in chapters 2 and 3 has no problem with preserving token-reflexivity in
translation. This is because meaning, on its account, is primarily a textual phenomenon. Thus technical terms may
be important into a translation to preserve token reflexivity or target language resources can be conscripted into a
token reflexive role through the device I labelled “text moulding” in chapter 2. Failing this textual approach to
meaning and translation, translation remains indeterminate, I argued.
127
Judith Jarvis Thomson provides considerations that support the observation that GOOD is axiologically
underdetermined, but she draws the conclusion from this observation (that the topic of judgments of goodness
requires elaboration) that there is no property of goodness (Thomson 1997). This is hasty. That one can still have a
theory of the Good though it is axiologically underdetermined is comprehensible on the model of QI. It may be
difficult for us to come to grips with what this ultimate theory of goodness is, but it does not follow that such a
theory is impossible or that it cannot specify some property that all good things must share. In the conclusion, I
provide considerations for arriving at objective knowledge in all areas of philosophy. An ultimate theory of
goodness may be possible. But it is more likely that we will arrive at true propositions in various areas of axiological
concern that would provide the raw data for an ultimate theory of goodness. Whether there is any theory that could
systematize such objectively true philosophical propositions is not clear.
128
A related issue to the question of the translation of value discourse is how one captures the semantics of deontic
logic that recognizes the obligatory, the forbidden and the permissible. The obligatory and the forbidden can be
explained on the model that I’ve provided: the obligatory would be specified by any deontic theory that is a theory of
action, which is substituted in for the theory variable in a normative concept, while the forbidden would be the
complement of the theory. PERMISSIBILITY however cannot be straightforwardly explained in terms of the model of
normativity that I have been deferring to for it does not have a direction of fit. This is only to recognize, however,
that PERMISSIBILITY is not really a value concept, though it is relevant to value discourse. Rather, “permissible” is a
thin term that refers to an item identified by a theory whose axiological differentia consists of identifying items that
are not accounted for either by normative theories or their complements. A theory of the permissible would have an
unusual direction of fit: it would have a theory of permissibility to value theory direction of fit. A theory of the
permissible would ultimately have to be grounded in a normative theory as a special set of items that it lists as good
not to criticize, but not as items that are to be approved of outright. The permissible would be recognized as
beyond reproach, but also not explicitly as good.
129
Jonathan Dancy, in an informative article, attempts to put forward a seemingly incomprehensible account of
thick concepts that at once seems to rely upon the traditional description of thick concepts as part evaluation and
part description but rejects it (Dancy 1995). Dancy’s objection appears to be that there is no determinate evaluative
or descriptive content in thick terms, but something corresponding to both the evaluative and descriptive are to be
found in thick terms, to the extent that “evaluation” and “description” are coherent notions (which he rejects). The
article is very informative for presenting a survey of the main views on the topic. However, I find his positive view
characteristically obscure. At any rate, the translational approach to the topic that I am encouraging is at odds with
Dancy’s seemingly linguistic approach.
130
I shall say more about this later in this chapter. See VII.2.1 TTS and Insensitive Semantics, pp. 319-324, and in
particular pp. 322-324.
131
Hare provides an account of translating “good” that is much like the one that Dreier wants to proffer, quoted
here on p. 258. This is what Hare has to say on the topic: “Let us suppose that a missionary, armed with a grammar
book, lands on a cannibal island. The vocabulary of his grammar book gives him the equivalent, in the cannibals’
language, of the English word ‘good’. Let us suppose that, by a strange coincidence, the word is ‘good’. And let us
suppose, also, that it really is the equivalent? that it is, as the Oxford English Dictionary puts it, ‘the most general
adjective of commendation’ . . . If the missionary has mastered his vocabulary, he can, so long as he uses the word
evaluatively and not descriptively, communicate with them about morals quite happily. They know that when he

420
uses the word he is commending the person or object that he applies it to. The only thing they find odd is that he
applies it to such unexpected people, people who are meek and gentle and do not collect large quantities of scalps;
whereas they themselves are accustomed to commend people who are bold and burly and collect more scalps than
the average. We thus have a situation which would appear paradoxical to someone who thought that ‘good’ . . . was
a quality-word like ‘red’. Even if the qualities in people which the missionary commended had nothing in common
with the qualities which the cannibals commended, yet they would both know what the word ‘good’ meant. If
‘good’ were like ‘red’, this would be impossible; for then the cannibals’ word and the English word would not be
synonymous . . . It is because in its primary evaluative meaning ‘good’ means neither of these things, but is in both
languages the most general adjective of commendation, that the missionary can use it to teach the cannibals
Christian morals” (Hare 1952, 146-9).
132
David Kaplan distinguishes between “contexts” and “circumstances”, but the distinction seems to be
idiosyncratic, and not clearly respected by subsequent authors (Kaplan 1989, 494). For Kaplan, a “context” is the
situation of word usage, while a circumstance is a situation of word usage evaluation. This is usually distinguished in
the literature these days in terms of the “context of utterance” and the “context of evaluation.”
133
An SL is the source language, or the language that a source text (ST) is written in. The TL is the target language,
or the language that the target text (TT) is written in.
134
The Routledge Encyclopedia of Translation Studies contains a not so careful article on the topic, that while drawing the
distinction between interpreting and translation, noting that interpreters have been at the forefront of drawing the
distinction, calls interpreting a kind of translation. See “Conference and Simultaneous Interpreting” (Gile 2001).
135
“Grammar does not tell us how language must be constructed in order to fulfil its purpose, in order to have
such-and-such an effect on human beings. It only describes and in no way explains the use of signs. The rules of
grammar may be called ‘arbitrary’, if that is to mean that the aim of the grammar is nothing but that of the language.
If someone says ‘If our language had not this grammar, it could not express these facts’—it should be asked what
‘could’ means here.” (Philosophical Investigations I §§496-97)
136
Wittgenstein still resorted to the notion of “grammar” in On Certainty, though much less and to much less
positive effect. In general, he seemed to appreciate the indeterminacy of context, and the purely rhetorical function
of normative concepts such as “know” and “certainty” in ordinary discourse. He writes, “What is ‘learning a
rule’?—This. What is ‘making a mistake in applying it’?—This. And what is pointed to here is something
indeterminate” (On Certainty 28).
137
I owe this historical insight into the development of Wittgenstein’s thought to Stuart Shanker.
138
In Davidsonian language, we could say that what linguistic context presents to us is a problem of interpretation
and interaction, not translation, which relies upon determinacy in meaning. We avoid the thorny issue of
determinacy of meaning within the context by stipulating significant agreement as the base from which interlocutors
attempt to linguistically interact. I find this to be an improvement on the very bare recommendations of
Wittgenstein that we simply follow rules in practice. However, I take this Davidsonian lesson on radical
interpretation to be restricted in significance to the problem of simultaneous interpretation, and not translation,
which I’ve argued is a textual process. Something like the principle of charity is operative in translation as well, but
it is subservient to the text-type of translation, that constrains the type of stipulated agreement that a translator must
share with an author of the ST, so that it is agreement in a particular, textual fashion, and not widespread or
extensive agreement. On my reading, Davidson is a Contextualist. However, he is the only Contextualist to attempt
to explain how objectivity and cross-linguistic understanding is possible on Contextualist grounds—a project that
Quine gives up on.
139
I have been focusing on the Wittgensteinian insights into meaning and context for they are far more helpful
than the ones we find in Austin. Austin is disappointingly unrigorous on the topic of translation. He confronts the
issue of translation in his ““[\]^_ and `ab[cd^_c[ in the Ethics of Aristotle” (1979b [1939]). But he does so

421
not as a translation theorist but someone debating the merits of various terminological translations, as though the
very issue of translation is not problematic, particularly in light of his Contextualism. In his “Are There A Priori
Concepts”, he comments in passing that translation will vary according to context, but he doesn’t tell us how or
why (Austin 1979a [1939], 39-40). According to Austin, the proper answer the question of the meaning of a word or
sentence is “explaining its syntactics and demonstrating its semantics” (Austin 1979 [1940], 60). But given this
account, it is absolutely unclear how translation is supposed to determinately proceed, for the answer to the
semantic question, on his view, is inextricably context-bound.
140
It would be incorrect to conclude that texts have no pragmatic function. A text of a marriage vow clearly does.
However, according to the view I am advocating, this is not essential to its meaning. If it were, the mere presence of
the text in any context would marry someone. But clearly it does not. Yet, in these contexts where the text of the
marriage vow does nothing, it still has the determinate meaning it does, by virtue of its translatability according to
TTS and its equivalence with other instances of the same text in other contexts.
141
I owe this insight into the distinction between mere social practices and institutions, and the distinction between
practical and expert authorities to Leslie Green, who pressed the distinction in a conversation I had with him.
However, Prof. Green was quite convinced that there are no text-type institutions. On his reckoning, if there were
an institution of poetry there would be practical authorities with the power to get us to read some poems and not
others. This gets the type of power that text-type experts have wrong. Text-type institutions serve the purpose of
defining text-types, and teaching those who are not in the institution how to read texts of a certain sort. They thus set
the conditions for what will be counted as an instance of the text-type they preside over. The practical authority of
text-type experts can be felt in numerous ways, from their influence on what texts are counted as instances of a text
of their type, what texts are published under their rubric, to penalties and rewards they can bestow on students. Like
all institutional power, these penalties and rewards have effects far beyond the narrow field of the institution.
142
Wittgenstein’s description of examples of language games might make it seem as though games have an
institutional structure, but this impression disappears when we realize the games are context-bound slices of social
interactions. As “games”, they have no higher purpose than the edification, or felicity, of the participants. The
problem, of course, is finding examples of genuine language games. Any social interaction can be conceived of as a
language game, but this would be to ignore the wider institutional incursions into the “game.” Wittgenstein gives
the following examples of language games: Giving orders, and obeying them; Describing the appearance of an
object, or giving its measurements; Constructing an object from a description (a drawing); Reporting an event;
Speculating about an event; Forming and testing a hypothesis; Presenting the results of an experiment in tables and
diagrams; Making up a story; and reading it; Play-acting; Singing catches; Guessing riddles; Making a joke; telling it;
Solving a problem in practical arithmetic; Translating from one language into another; Asking, thanking, cursing,
greeting, praying (Philosophical Investigation § 23).
143
The argument I am presenting here might seem to run afoul of what Anthropologists do, for it seems that they
are charged with the task of determining questions like what Canadian etiquette, objectively, is. Careful
Anthropologists of recent time, it seems to me, are deeply aware of the problem of interpreting social practices and
thus often avoid making claims about the objectivity of their interpretations.
144
Kaplan writes: “The group of words for which I propose a semantical theory includes the pronouns ‘I’, ‘my’,
‘you’, ‘he’, ‘his’, ‘she’, ‘it’, the demonstrative pronouns ‘that’, ‘this’, the adverbs ‘here’, ‘now’, ‘tomorrow’, ‘yesterday’,
the adjectives ‘actual’, ‘present’, and others.” (Kaplan 1989, 489). He continues: “What is common to the words or
usages in which I am interested is that the referent is dependent on the context of use and that the meaning of the
word provides a rule which determines the referent in terms of certain aspects of the context. The term I now
favour for these words is ‘indexical’. Other authors have used other terms; Russell used ‘egocentric particular’ and
Reichenbach used ‘token reflexive’. I prefer ‘indexical’ (which, I believe, is due to Pierce) because it seems less
theory laden than the others, and because I regard Russell’s and Reichenbach’s theories as defective” (Kaplan 1989,
490).

422
145
The lesson of the answering machine “I’m not here now”, as well as counterfactual, embedded uses of
indexicals (“If I were you, I would…”) is likely that indexicality is far more complicated a phenomenon than Kaplan
had initially explained it, and that contextual factors such as the context of utterance and the context of evaluation
must be taken into account (cf. Predelli 1998). Particularly, it suggests that there is nothing determinate about the
relationship between the character of a term and its meaning. This should not surprise us for the meaning of
language on the whole is indeterminate, unless it is cotextualized.
146
Cappelen and Lepore note that one author has explicitly taken exception with the view that English contains no
monsters (Schlenker 2003) but they claim that the types of putative monsters that English may contain are of little
help to the Contextualist (Cappelen and Lepore 2005, 120 fn. 24). That the issue is controversial ought to give
Cappelen and Lepore reason for pause.
147
By this I do not mean to imply that animals do not understand human language. They clearly do. I discovered
this vividly when travelling through Italy: none of the dogs responded to my English. However, the more general
point can be made thus: once pragmatic felicity is understood as the foundation of linguistic competence, there is
little reason to assume that animals could not understand language.
148
One might attempt to understand the work of Frege (1988 [1918-19]), and some of the early positivists as an
attempt to deal with context insensitivity in meaning, but the history of philosophy generally judges such attempts
as failures. The trouble is that the early Analysts had not understood the difficulties that pragmatics makes for a
theory of communication. Austin and Wittgenstein attempt to embrace these problems, but in so doing give up on
the possibility of context insensitivity in semantics. At the time it may have seemed like a good idea, but it is the
type of idea that only seems good if one ignores issues such as globalization and immigration. Derrida, for his part
does not give up on the idea of context insensitivity in semantics, but his execution of a solution to the problem
fails in part because he does not distinguish between pragmatics and semantics.
149
“…what I have been talking about is the possibility of death. It is not a thesis on our mortality. I am not saying ‘We
are mortal’ or ‘Death is inevitable’. I would be inclined to think so! But that is not the point. The point is that for a
sentence such as ‘I am dead’ to be a sentence— an intelligible, meaningful sentence—it has to be implied that I may be
absent and that it can continue to function. The functioning of the sentence doesn’t require my being present to it.
On the contrary, the functioning of the sentence implies the possibility of my being radically ‘on leave’ so to
speak, radically absent. So, when I speak of death in this case, it is just a figure to refer to this absence, to refer to
the structural conditions of possibility for the sentence to be performed, understood and repeated. Thus, it is not
a thesis about death. It is simply that for the sentence, spoken or written… to be understood it must be possible
for its link with the origin to be interrupted” (Derrida 2000, 400-401).
150
Hobbes writes: “But whatsoever is the object of any man’s appetite or desire, that is it which he for his part
calleth ‘good’; and the object of his hate and aversion, ‘evil’; and of his contempt, vile and inconsiderable. For these
words of ‘good’, ‘evil’, and ‘contemptible’ are ever used with relation to the person that useth them: there being
nothing simply and absolutely so; nor any common rule of good and evil to be taken from the nature of the objects
themselves; but from the person of the man, where there is no Commonwealth; or, in a Commonwealth, from the
person that representeth it; or from an arbitrator or judge, whom men disagreeing shall by consent set up and make
his sentence the rule thereof.” (Leviathan 6.7)(Hobbes 1994 [1651])

151
The most famous author to make a case that philosophy is best treated as a peculiarly European affair is Richard
Rorty. For more on this, see Anindita Niyogi Balslev’s interesting published correspondence with Rorty on
comparative philosophy (Balslev and Rorty 1991) and Wei Zhang’s analysis of this correspondence (Zhang 2006).
Zhang writes, “…, locating himself in a cross-cultural context, Rorty’s position of anti-essentialism turned
essentialist.” (Zhang 2006, 2). In correspondence with Balslev, Rorty denies that philosophy is a particular type of
genre. He writes:
My own view is that philosophy is not a genre of discourse, but simply a genealogical linkage connecting
certain past figures with certain present figures - not a thread running through the rope, in Wittgenstein’s

423
 figure, but just a way of noting that there is an ancestral relation of overlapping fibers. That was my
point when I said that philosophy is delimited not by form or matter but by tradition. However, in the
passage from Consequences of Pragmatism to which you refer, I did refer to philosophy as a ‘literary genre.’
That was a mistake. For the word genre suggests format, and I did not mean to do that.
So my answer to your question ‘Will you grant that the very concept of philosophy is a generic concept?’
must, I think be ‘no.’ In the sense in which I think you intend the term ‘generic concept’, I take it,
whether something is instance of that concept can be established without reference to historical or
cultural context. This may be true of pictorial art or of music, in the sense that these are distinct from
written words in obvious, inter-cultural and transhistorical ways. But when it comes to distinguishing
among written words, I do not think that we have a way of dividing up texts which meets your
requirements.
To get an interesting classification of written texts, one needs to answer the question: what other texts
are relevant to this one? Answering this question often does help one block out the written word into
areas. Chemical treatises cluster together, for example, as do love stories. But the most interesting texts,
usually, are the ones which Geertz describes as blurring genre-divisions. Most of the truly original and
history-changing texts are of this sort - they are texts which were, on their first appearance, rather unlike
anything that had previously been seen. (Think of Plato’s Dialogues, for example, or of Machiavelli’s The
Prince, or Hegel’s Phenomenology of Spirit.) The really important texts are the ones that render our old
classifications unsatisfactory and force us to think up new ones. (Balslev and Rorty 1991)
It is odd that Rorty should think that there is some type of distinction to be drawn between music and pictures, on
the one hand, and texts on the other, for they are all semiotic phenomena. Rorty’s clear mistake, which is not
peculiar to him and very common, is to confuse texts with types. Texts may themselves transverse more than one
type, but this is not inconsistent with them falling under a type. Indeed, such texts fall under not one but two or
more types. Texts never really blur the distinction between text-types. It is we who blur the distinction when we do
not distinguish between tokens of texts and their types. A text in the Indian tradition being a token of philosophy is
simply a matter of whether we could translate it as such, and such a feat would purposively not aim to capture all
the features of the text we are translating: only the philosophical features.
152
John Rawls is famous for bringing to light the idea of reflective equilibrium, though he finds it in the work of
Nelson Goodman (Goodman 1954, 63-68). Rawls describes it as a process of dialectically mediating convictions in
light of evidence that may come to unseat initial convictions (Rawls 1971, 42-3). “Wide” reflective equilibrium, as
Hales describes it, concerns the mediation of convictions according to all relevant evidence, and not simply a
narrow concern (Hales 2006, 42).

424
References

Aldrich, Virgil A. “Illocutionary Space”. Philosophy and Phenomenological Research (1971): 22 15-28.
Alm, David. “Moral Conditionals, Non-Cognitivism, and Meaning”. The Southern Journal of Philosophy 38, no. 3 (2000): 355-377.
Anscombe, G. E. M. “On Brute Facts”. Analysis 18 (1958): 69-72.
———. Intention. 2nd ed. New York: Cornell University Press, 1963. 1957.
Appiah, Kwame Anthony. “Thick Translation.” In The Translation Studies Reader, ed. L. Venuti, 389-401. New York: Routledge,
2004 Callaloo. 16, (1993): 808-19.
Armstrong, A. H. “Plotinus and India”. The Classical Quarterly 30, no. 1 (1936): 22-28.
Ashton, E. B. “Translating Philosophy”. Delos 6 (1971): 16-29.
Austin, J. L. Are There A Priori Concepts? 3d ed., Philosophical Papers, ed. J. O. Urmson and G. J. Warnock. Oxford: Oxford
University Press, 1979a. (1939).
———. “ΕΓΑΘΟΝ and ΕΥ∆ΑΙΜΟΝΙΑ in the Ethics of Aristotle.” In Philosophical Papers, eds. J. O. Urmson and G. J. Warnock,
1-31. Oxford: Oxford University Press, 1979b, (1939).
———. “The Meaning of a Word.” In Philosophical Papers, eds. J. O. Urmson and G. J. Warnock, 55-75. Oxford: Oxford
University Press, 1979, (1940).
Austin, John Langshaw. How to Do Things with Words. 2nd ed., William James lectures. 1955. Cambridge: Harvard University Press,
2000. 1962.
Ayer, A.J. Language Truth and Logic. New York: Dover Publications, 1946. 1936.
Azzouni, Jody. “Thick Epistemic Access: Distinguishing the Mathematical from the Empirical”. Journal of Philosophy (1997): 94(9)
472-484.
Balslev, Anindita Niyogi, and Richard Rorty. Cultural Otherness: Correspondence with Richard Rorty. Shimla: Indian institute of
Advanced Study, 1991.
Bassnett, Susan. Translation Studies. 3rd ed. London; New York: Routledge, 2002.
Bassnett, Susan, and André Lefevere, eds. Constructing Cultures: Essays on Literary Translation. Topics in translation ; 11. Clevedon;
Philadelphia: Multilingual Matters, 1998.
Bastin, Georges S. “Adaptation.” In Routledge Encyclopedia of Translation Studies, trans. M. Gregson, eds. M. Baker and K.
Malmkjæar, 5-8. London ; New York: Routledge, 2001.
Blackburn, Simon. “Morality and Thick Concepts (II—Through Thick and Thin)”. Aristotelian Society Supp 66 (1992): 285-99.
———. “Truth, Realism and the Regulation of Theory.” In Essays in Quasi-Realism, 15-34. New York: Oxford University Press,
1993 Midwest Studies in Philosophy. v, (1980).
———. Ruling Passions: a Theory of Practical Reasoning. Oxford; New York: Oxford University Press, 2000. Oxford University
Press, 1998.
Boghossian, Paul A. “Analyticity Revisited”. Nous 30 (1996): 360-91.
———. “Analyticity.” In A Companion to the Philosophy of Language, eds. B. Hale and C. Wright, 331-68. Oxford: Blackwell, 1999.
Boyd, Richard N. “How to be a Moral Realist.” In Essays on Moral Realism, ed. G. Sayre-McCord, 181-228. Ithaca, New York:
Cornell University Press, 1988.
———. “Finite Beings, Finite Goods: The Semantics, Metaphysics and Ethics of Naturalist Consequentialism, Part I”. Philosophy
and Phenomenological Research 66, no. 3 (2003a): 505-553.
———. “Finite Beings, Finite Goods: The Semantics, Metaphysics and Ethics of Naturalist Consequentialism, Part II”.
Philosophy and Phenomenological Research 67, no. 1 (2003b): 24-47.

425
Brink, David Owen. Moral Realism and the Foundations of Ethics. Cambridge Studies in Philosophy. Cambridge; New York: Cambridge
University Press, 1989.
———. “Realism, Naturalism, and Moral Semantics”. Social Philosophy and Policy 18, no. 2 (2001): 154-176.
Brobjer, Thomas H. “Nietzsche’s Reading About Eastern Philosophy”. Journal of Nietzsche Studies, no. 28 (2004): 3-27.
Burton, Stephan L. “Thick' Concepts Revised”. Analysis (1992): 52(1) 28-32.
Cappelen, Herman, and Ernest Lepore. Insensitive Semantics: a Defense of Semantic Minimalism and Speech Act Pluralism. Malden, MA:
Blackwell, 2005.
Carnap, Rudolf. The Logical Syntax of Language. London: K. Paul, Trench, 1937.
———. Testability & Meaning. New Haven, Connecticut: Whitlock's, 1954. Philosophy of Science. 3-4, 1936.
———. “The Methodological Character of Theoretical Concepts.” In Minnesota Studies in the Philosophy of Science, eds. H. Feigl and
M. Scriven, vol. 1. Minneapolis: University of Minnesota Press, 1956.
———. “Carnap's Intellectual Biography.” In The Philosophy of Rudolf Carnap, ed. P. A. Schilpp, Library of living philosophers. 11, 3-
86. La Salle, Ill.: Open Court, 1963.
———. The Logical Structure of the World [and] Pseudoproblems in Philosophy. Translated by R. A. George. Berkeley: University of
California Press, 1967. Der logische Aufbau der Welt. Berlin: Weltkreis-Verlag, 1928.
Catford, J. C. A Linguistic Theory of Translation: an Essay in Applied Linguistics. London: Oxford University Press, 1965.
Chesterman, Andrew. “Ethics of Translation.” Translation as Intercultural Communication:, vol. 20, 147-160. Amsterdam ;
Philadelphia:: J. Benjamins, 1997.Originally presented at Selected Papers from the EST Congress, Prague 1995.
Chomsky, Noam. “Quine‘s Empirical Assumptions.” In Words and Objections: Essays on the work of W.V. Quine, eds. D. Davidson,
W. V. O. Quine and K. K. J. Hintikka, 53-68. Dordrecht: D. Reidel, 1969.
———. Reflections on Language. 1st ed. New York,: Pantheon Books, 1975.
———. Rules and Representations. Oxford: B. Blackwell, 1980.
Chomsky, Noam, and Mitsou Ronat. Language and Responsibility: Based on Conversations with Mitsou Ronat. Translated by J. Viertel. 1st
American ed. New York: Pantheon Books, 1979.
Clement. Stromateis. Translated by J. Ferguson. Washington, D.C.: Catholic University of America Press, 1991.
Coetzee, J. M., and Amy Gutmann. The Lives of Animals. University Center for Human Values series. Princeton, NJ: Princeton
University Press, 1999.
Cohen, Stewart. “Knowledge and Context”. Journal of Philosophy 83 (1986): 574-583.
———. “Knowledge, Context, and Social Standards”. Synthese 73 (1987): 3-26.
———. “How to Be a Fallibilist”. Philosophical Perspectives 2 (1988): 581-605.
———. “Contextualist Solutions to Epistemological Problems: Scepticism, Gettier, and the Lottery”. Australasian Journal of
Philosophy 76, no. 2 (1998): 289-306.
———. “Knowledge, Assertion, and Practical Reasoning”. Philosophical Issues 14, no. 1 (2004): 482-491.
———. “Knowledge, Speaker and Subject”. The Philosophical Quarterly 55, no. 219 (2005): 199-212.
Confucius. “Selection from the Analects.” In A Source Book in Chinese Philosophy, trans. W.-T. Chan, ed. W.-T. Chan, 18-48.
Princeton, N.J.,: Princeton University Press, 1963.
Cook, John W. Morality and Cultural Difference. Oxford: Oxford University, 1999.
Cooper, David Edward. Meaning. Montreal: McGill-Queen's University Press, 2003.
Cooper, Neil. “Morality and Importance.” In The Definition of Morality, eds. G. Wallace and A. D. M. Walker, 91-97. London:
Methuen and Co., 1970.
Copp, David. “Realist-Expressivism: A Neglected Option for Moral Realism”. Social Philosophy and Policy 18, no. 2 (2001): 1-43.
Dancy, Jonathan. “In Defense of Thick Concepts”. Midwest Studies in Philosophy XX (1995): 261-273.
Daniels, Peter T. “The Study of Writing Systems.” In The World's Writing Systems, eds. W. Bright and P. T. Daniels, 3-18. New
York: Oxford University Press, 1996.

426
Dasgupta, Surama. Development of Moral Philosophy in India. New York: Frederick Ungar Publishing Co., 1961.
Davidson, Donald. “A Coherence Theory of Truth and Knowledge.” In Truth and Interpretation: Perspectives on the Philosophy of
Donald Davidson, ed. E. Le Pore, 307-319. Cambridge: Blackwell, 1986, Read at a colloquium organized by Richard
Rorty as part of the 1981 Stuttgart Hegel Congress.
———. “Radical Interpretation Interpreted.” In Logic and Language, ed. J. E. Tomberlin, Philosophical Perspectives 8, 121-128.
Atascadero, Calif.: Ridgeview Pub. Co., 1994.
———. “Belief and the Basis of Meaning.” In The Philosophy of Language, ed. A. Martinich, 456-464. New York: Oxford
University Press, 1996 Synthese. 27, (1974): 309-323.
———. “The Objectivity of Values.” In El Trabajo Filosofico de Hoy en el Continente, ed. C. Gutieirez. Bogata: Editorial ABC,
1996.Originally presented at 13th Inter-American Congress of Philosophy, at Universidad de los Andes (1994).
———. “What Metaphors Mean.” In The Philosophy of Language, ed. A. Martinich, 415-426. New York: Oxford University Press,
1996 On Metaphor. Edited by Sheldon Sacks, 29-46. Chicago: University of Chicago Press, 1978.
———. “Objectivity and Practical Reason.” In Reasoning Practically, ed. E. Ullmann-Margalit, 17-26. New York: Oxford
University Press, 2000a.
———. “Truth Rehabilitated.” In Rorty and His Critics, ed. R. Brandom, Philosophers and Their Critics; 9, 65-73. Malden, Mass.:
Blackwell Publishers, 2000b.
———. “Inscrutability of Reference.” In Inquiries into Truth and Interpretation, 227-242. Oxford: Claredon Press (Oxford
University Press), 2001 The Southwestern Journal of Philosophy. 10, (1979): 7-19.
———. “On Saying That.” In Inquiries into Truth and Interpretation, 93-108. Oxford: Claredon Press, 2001 Synthese. 19, (1968-9):
130-46.
———. “On the Very Idea of a Conceptual Scheme.” In Inquiries into Truth and Interpretation, 183-198. New York: Oxford
University Press, 2001 Proceedings and Addresses of the American Philosophical Association. 47, (1974): 5-20.
———. “Radical Interpretation.” In Inquiries into Truth and Interpretation, 125-140. Oxford: Claredon Press (Oxford University
Press), 2001 Dialectica. 27, (1973): 313-28.
———. “Reality Without Reference.” In Inquiries into Truth and Interpretation. Oxford: Clarendon Press (Oxford University Press),
2001 Dialectica. 31, (1977): 247-53. First given in different form at a Semantics and Linguistics Workshop at the
University of Western Ontario, London, Ontario in April 1972.
———. “Truth and Meaning.” In Inquiries into Truth and Interpretation, 17-36. Oxford: Claredon Press (Oxford University Press),
2001 Synthese. 17, (1967): 304-23.
Davis, Kathleen. Deconstruction and Translation. Translation Theories Explained, 8. Manchester, UK; Northampton, MA: St. Jerome
Publishers, 2001.
Davis, Wayne A. “Are Knowledge Claims Indexical?” Erkenntnis: An International Journal of Analytic Philosophy 61, no. 2-3 (2004):
257-281.
De Caro, Mario, and David Macarthur. Naturalism in Question. Cambridge, Mass.: Harvard University Press, 2004.
Deleuze, Gilles, and Félix Guattari. What is Philosophy? Translated by H. Tomlinson and G. Burchell. European Perspectives. New
York: Columbia University Press, 1994. Qu'est-ce que la philosophie? : Les Editions de Minuit, 1991.
DePaul, Michael R. “Linguistics Is Not a Good Model for Philosophy (Commentary on Nenand Miscevic).” In Southern Journal of
Philosophy: The Role of the A Priori (and of the A Posteriori) in Epistemology, Spindel Conference Supplement 38, eds. D.
Henderson and T. Horgan, 113-20, 2000.
DeRose, Keith. “Solving the Skeptical Problem”. The Philosophical Review 104, no. 1 (1995): 1-52.
———. “The Problem with Subject-Sensitive Invariantism”. Philosophy and Phenomenological Research 68, no. 2 (2004): 346 -350.
———. “Bamboozled by Our Own Words': Semantic Blindness and Some Objections to Contextualism”. Philosophy and
Phenomenological Research (2005a).
———. “The Ordinary Language Basis for Contextualism, and the New Invariantism”. The Philosophical Quarterly 55, no. 219
(2005b): 172-198.
Derrida, Jacques. Of Grammatology. Translated by G. C. Spivak. 1st American ed. Baltimore: Johns Hopkins University Press, 1974.
De la Grammatologie.

427
———. “Plato's Pharmacy”. In Dissemination, trans. B. Johnson. Chicago: University of Chicago Press, 1981.“La pharmacie de
Platon,” Tel Quel. 32 and 33, (1967): 3-48 and 18-59.
———. “Signature, Event, Context.” In Margins of Philosophy, trans. A. Bass, 307-330. Chicago: University of Chicago Press,
1982: First presented at the Sociétés de philosophie de langue francaise, Montreal, August, 1971.
———. “Response to Baldwin”. Ratio 13, no. 4 (2000): 400-407.
———. “What is a "Relevant" Translation.” In The Translation Studies Reader, trans. L. Venuti, ed. L. Venuti, 423-446. New York:
Routledge, 2004 Critical Inquiry. 16, (2001): 174-200.
Derrida, Jacques, and Christie V. McDonald. The Ear of the Other: Otobiography, Transference, Translation: Texts and Discussions with
Jacques Derrida. New York: Schocken Books, 1985.
Descartes, René. “Letter to Mersenne of 20 November 1629.” In Descartes: Philosophical Letters, ed. A. J. P. Kenny. Oxford:
Clarendon Press, 1970, 1629.
Dewey, John, and James H. Tufts. Ethics. New York: Henry Holt and Co, 1929.
Donnellan, Keith S. “Reference and Definite Descriptions”. Philosophical Review 75 (1966): 281-304.
Dorjee, Tenzin. “Styles and Problem Areas in Translating from Tibetan into English.” In Buddhist Translations: Problems and
Perspectives, ed. D. Tulku, 222-228. New Delhi: Manohar, 1995.Originally presented at International Seminar on
"Buddhist Translations: Problems and Perspectives" (1990 : Delhi, India).
Dreier, James. “Internalism and Speaker Relativism”. Ethics 101 (1990): 6-26.
———. “Transforming Expressivism”. Nous 33, no. 4 (1999): 558-572.
———. “Metaethics and the Problem of Creeping Minimalism”. Philosophical Perspectives, 18, Ethics 18 (2004): 23-41.
Dryden, John. “On Translation [From the Preface to his translation of Ovid's Epistles].” In Theories of Translation: an Anthology of
Essays from Dryden to Derrida, eds. R. Schulte and J. Biguenet, 17-31. Chicago: University of Chicago Press, 1992 Ovid's
Epistles. 1680.
Dummett, Michael A. E. Frege: Philosophy of Language. 1st U.S. -- ed. New York: Harper & Row, 1973.
———. Origins of Analytical Philosophy. Cambridge, Mass.: Harvard University Press, 1994.
———. “Preface.” In (Edmund Husserl's) Logical Investigations, trans. J. N. Findlay, vol. 1, xvii-xix. London; New York: Routledge,
2001.
Evans, Gareth. “Causal Theory of Names.” In The Philosophy of Language, ed. A. Martinich, 271-283. New York: Oxford
University Press, 1996 Aristotelian Society: Supplementary Volume 47. (1973): 187-208.
Everett, Daniel L. “Cultural Constraints on Grammar and Cognition in Pirahã: Another Look at the Design Features of Human
Language”. 46, no. 4 (2005): 621-636.
Falk, W. D. “Morality, Self and Others.” In Morality and the Language of Conduct, eds. H. N. Castaneda and G. Nakhnikian, 25-66.
Detroit: Wayne State University, 1963.
Finlay, Stephen. “Value and Implicature”. Philosophers' Imprint 5, no. 4 (2005): 1-20.
Firth, J. R. The Tongues of Men, and Speech. London: Oxford University Press, 1964. 1930.
Flintoff, Everard. “Pyrrho and India”. Phronesis 25 (1980): 88-106.
Fodor, Jerry, and Ernest Lepore. Holism: A Shopper's Guide. Cambridge: Blackwell, 1992.
———. “Is Radical Interpretation Possible?” In Logic and Language, ed. J. E. Tomberlin, Philosophical Perspectives 8, 101-119.
Atascadero, Calif.: Ridgeview Pub. Co., 1994.
Føllesdal, Dagfinn, ed. Indeterminacy of Translation. Philosophy of Quine. New York: Garland Pub., 2001.
Foot, Philippa. “Moral Arguments.” In Virtues and Vices, and Other Essays in Moral Philosophy, Values and Philosophical Inquiry, 96-
109. Oxford: B. Blackwell, 1978 Mind. 67, (1958).
Foucault, Michel. “What is an Author?” In The Foucault Reader, eds. M. Foucault and P. Rabinow, 101-120. New York: Pantheon
Books, 1984.
Foucault, Michel, and Donald F. Bouchard, eds. Language, Counter-Memory, Practice: Selected Essays and Interviews. Ithaca, N.Y.:
Cornell University Press, 1977.

428
Frankena, William K. “Recent Conceptions of Morality.” In Morality and the Language of Conduct, eds. H. N. Castaneda and G.
Nakhnikian, 1-24. Detroit: Wayne State University, 1963.
———. Ethics. 2nd ed. Englewood Cliffs NJ: Prentice Hall., 1973.
———. “Has Morality an Independent Bottom; Lecture III: Three Questions About Morality”. Monist 63, no. 1 (1980a).
———. “Is Morality a System of Ordinary Oughts; Lecture II: Three Questions About Morality”. Monist 63, no. 1 (1980b): 27-
47.
———. “Must Morality Have an Object; Lecture I: Three Questions About Morality”. Monist 63, no. 1 (1980c): 3-26.
Freeman, J. L. “Copy number variation: new insights in genome diversity”. 16 (2006): 949-961.
Frege, Gottlob. “Begriffsschrift, a Formula Language, Modeled upon that of Arithmetic, for Pure Thought.” In Frege to Gödel, ed.
J. v. Heijenoort. Cambridge, MA: Harvard University Press, 1967 Begriffsschrift, eine der arithmetischen nachgebildete
Formelsprache des reinen Denkens. Halle: L. Nebert, 1879.
———. The Foundations of Arithmetic: a Logico-Mathematical Enquiry into the Concept of Number. Translated by J. L. Austin. 2nd rev. ed.
Oxford: Basil Blackwell, 1980. Die Grundlagen der Arithmetik, eine Logisch Mathematische Untersuchung über den Begriff der
Zahl. Breslau: W. Koebner, 1884.
———. “Thoughts.” In Propositions and Attitudes, trans. P. Geach and R. H. Stoothoff, eds. N. U. Salmon and S. Soames, 33-55.
Oxford; New York: Oxford University Press, 1988“Der Gedanke,” Beträge zur Philosophie des deutschen Idealismus. 1,
(1918-19): 58-77.
Friedman, Michael. “Physicalism and the Indeterminacy of Translation.” In Indeterminacy of Translation, ed. D. Føllesdal, Philosophy
of Quine, vol. 3, 165-186. New York: Garland Pub., 2001 Nous. 9, (1975): 353-374.
Gadamer, Hans-Georg. “Culture and the Word.” In Hermeneutics and the Poetic Motion Translation Perspectives V, trans. D. J. Schmidt,
ed. D. J. Schmidt, 11-24, 1990.
———. Truth and Method. Translated by J. Weinsheimer and D. G. Marshall. 2nd Revised English Language ed. New York:
Continuum, 1996. Wahrheit und Methode. Tubingen: J.C.B. Mohr (Paul Seibeck), 1960.
Garcia-Carpintero, Manuel. “Indexicals as Token-Reflexives”. Mind: A Quarterly Review of Philosophy 107, no. 427 (1998): 529-563.
Gaudet, Eve. Quine on Meaning. Continuum Studies in American Philosophy. London ; New York: Continuum, 2006.
Geertz, Clifford. “Thick Descriptions: Towards an Interpretive Theory of Culture”. In The Interpretation of Cultures, 3-30. New
York: Harper and Row, 1973.
Geirsson, Heimir. “Moral Twin-Earth and Semantic Moral Realism”. Erkenntnis: An International Journal of Analytic Philosophy 62,
no. 3 (2005): 353-378.
Gentzler, Edwin. Contemporary Translation Theories. 2nd ed. Toronto: Multilingual Matters, 2001.
Gettier, Edmund. “Is Justified True Belief Knowledge?” Analysis 23 (1963): 121-123.
Gibbard, Allan. Wise Choices, Apt Feelings: a Theory of Normative Judgment. Cambridge, Mass.: Harvard University Press, 1990.
———. “Reasons Thin and Thick”. Journal of Philosophy 100, no. 6 (2003a): 288-304.
———. Thinking How to Live. Cambridge, Mass.: Harvard University Press, 2003b.
Gibson, Roger. “Radical Translation and Radical Interpretation.” In Routledge Encyclopedia of Philosophy, ed. E. Craig, vol. 8, 25-35.
London: Routledge, 1998.
Gile, Daniel. “Conference and Simultaneous Interpeting.” In Routledge Encyclopedia of Translation Studies, eds. M. Baker and K.
Malmkjæar, 40-45. London; New York: Routledge, 2001.
Goldman, Alvin , and Joel Pust. “Philosophical Theory and Intuitional Evidence.” In Rethinking intuition: the Psychology of Intuition
and Its Role in Philosophical Inquiry, eds. M. R. DePaul and W. Ramsey, 179-97. Lanham, Md.: Rowman & Littlefield,
1998.
Goodman, Nelson. Fact, Fiction & Forecast. London: University of London Athlone Press, 1954.
Grandy, Richard. “Reference, Meaning and Belief.” In Indeterminacy of Translation, ed. D. Føllesdal, Philosophy of Quine, vol. 3, 385-
452. New York: Garland Pub., 2001 Journal of Philosophy. 70, (1973): 439-52.
Greenberg, Mark, and Gilbert Harman. “Conceptual Role Semantics.” In Oxford Handbook of the Philosophy of Language, eds. E.
Lepore and B. Smith, 295-322. Oxford: Oxford University Press, 2006.

429
Groth, Miles. Translating Heidegger. Contemporary Studies in Philosophy and the Human Sciences. Amherst, N.Y.: Humanity Books, 2004.
Guetti, James L. Wittgenstein and the Grammar of Literary Experience. Athens: University of Georgia Press, 1993.
Halbfass, Wilhelm. Tradition and Reflection: Explorations in Indian Thought. Albany, N.Y.: State University of New York Press, 1991.
Hales, Steven D. Relativism and the Foundations of Philosophy. Cambridge, Mass.: MIT Press, 2006.
Halliday, M. A. K. An Introduction to Functional Grammar. London: Edward Arnold, 1985.
Halliday, M. A. K., and Ruqaiya Hasan. Language, Context, and Text: Aspects of Language in a Social-Semiotic Perspective. 2nd ed.,
Language education. Oxford: Oxford University Press, 1989. Deakin University Press, 1985.
Hann, Michael. The Key to Technical Translation. Amsterdam/Philadelphia: J. Benjamins, 1992.
Hare, R. M. “Universal Prescriptivism.” In A Companion to Ethics, ed. P. Singer, Blackwell companions to philosophy, 451-463. Oxford,
UK ; Cambridge, Mass.: Blackwell Reference, 1993.
Hare, Richard M. The Language of Morals. Oxford: Clarendon Press, 1952.
———. “Universalizability”. Proceedings of the Aristotelian Society (1955): 295-312.
———. Freedom and Reason. London: Oxford University Press, 1971.
Harman, Gilbert. “An Introduction to ‘Translation and Meaning,’ Chapter Two of Word and Object.” In Words and Objections:
Essays on the Work of W.V. Quine, eds. D. Davidson and K. K. J. Hintikka, 14-26. Dordrecht: D. Reidel, 1969.
———. “Meaning Holism Defended.” In Holism: A Consumer Update, ed. J. A. Fodor. Amsterdam: Rodopi, 1993 Grazer
Philosophische Studien. 46, (1993): 163-171.
Harris, Roy. The Language Myth. London: Duckworth, 1981.
———. Rethinking Writing. Bloomington: Indiana University Press, 2000.
———. “Integrationism.” Available from http://www.royharrisonline.com/integrationism.html (2005a).
———. The Semantics of Science. London ; New York: Continuum, 2005b.
Hart, H.L.A. “Justice And Morality”. In The Concept Of Law, 151-179. Oxford: Clarendon Press, 1961.
Hart, W. D. “On Self-Reference”. Philosophical Review 79 (1970): 523-528.
Hawthorne, John. Knowledge and Lotteries. Oxford; New York: Clarendon; Oxford University Press, 2004.
Hegel, Georg Wilhelm Friedrich. Lectures on the History of Philosophy: the Lectures of 1825-1826. Translated by R. F. Brown, J. M.
Stewart and H. S. Harris. 3 vols., ed. R. F. Brown. Berkeley: University of California Press, 1990. 1825-1826.
Heidegger, Martin. Being and Time. Translated by J. Macquarrie and E. Robinson. New York: Harper, 1962.
———. “The Anaximander Fragment”. In Early Greek Thinking, trans. D. F. Krell and F. A. Capuzzi., 13-58. San Francisco:
Harper & Row, 1984.“Der Spruch der Anaximander,” (1946).
Heisenberg, Werner. Physics & philosophy : the revolution in modern science. Harper torchbooks. Science library. New York: Harper & Row,
1962.
Hobbes, Thomas. Leviathan: with Selected Variants from the Latin edition of 1668. ed. E. M. Curley. Indianapolis: Hackett Pub. Co.,
1994. 1651.
Hodes, Harold T. “Ontological Commitment: Thick and Thin.” In Meaning and Method: Essays in Honor of Hilary Putnam. New
York: Cambridge University Press, 1990.
Holland, Sean. “Dispositional Theories of Value Meet Moral Twin Earth”. American Philosophical Quarterly 38, no. 2 (2001): 177-
195.
Holmes, James S. “The Cross-Temporal Factor in Verse Translation”. In Translated!: Papers on Literary Translation and Translation
Studies, 35-44. Amsterdam: Rodopi, 1988.
———. “The Future of Translation Theory: A Handful of Theses”. In Translated!: Papers on Literary Translation and Translation
Studies, 99-102. Amsterdam: Rodopi, 1988. Originally presented at International Symposium on Achievements in the Theory of
Translation held in Moscow and Yerevan, 23-30 October, 1978.
———. “The Name and Nature of Translation Studies.” In Translated!: Papers on Literary Translation and Translation Studies,
Approaches to translation studies ; v. 7, 67-80. Amsterdam: Rodopi, 1988.Originally presented at the Translation Section of

430
 the Third International Congress of Applied Linguistics (in a contracted form), held in Copenhagen, 21-26 August,
1972.
Horgan, Terence, and Mark Timmons. “New Wave Moral Realism Meets Moral Twin Earth”. Journal of Philosophical Research 16
(1991): 447-465.
———. “Troubles for New Wave Moral Semantics: The Open Question Argument' Revived”. Philosophical Papers 21, no. 3
(1992a): 153-175.
———. “Troubles on Moral Twin Earth: Moral Queerness Revived”. Synthese 92, no. 2 (1992b): 221-260.
———. “From Moral Realism to Moral Relativism in One Easy Step”. Critica 28, no. 83 (1996): 3-39.
———. “Copping Out on Moral Twin Earth”. Synthese 124, no. 1 (2000a): 139-152.
———. “Nondescriptivist Cognitivism: Framework for a New Metaethic”. Philosophical Papers 29, no. 2 (2000b): 121-153.
———. “Cognitivist Expressivism.” In Metaethics after Moore, eds. T. Horgan and M. Timmons. Oxford; New York: Clarendon
Press; Oxford University Press, 2006.
Horwich, Paul. Meaning. Oxford; New York: Clarendon Press; Oxford University Press, 1998.
Humberstone, I. L. “Direction of Fit”. Mind 101 (1992): 59-84.
Husserl, Edmund. “Philosophy and the Crisis of European Man.” In Phenomenology and the Crisis of Philosophy, trans. Q. Lauer. New
York: Harper & Row, 1965.
Jackman, Henry. “Intuitions and Semantic Theory”. Metaphilosophy 36, no. 3 (2005): 363-380.
Jackson, Frank, and Philip Pettit. “A Problem for Expressivism”. Analysis 58, no. 4 (1998): 239-251.
Jakobson, Roman. “On Linguistic Aspects of Translation.” In Theories of Translation: an Anthology of Essays from Dryden to Derrida,
eds. R. Schulte and J. Biguenet, 144-151. Chicago: University of Chicago Press, 1992 On Translation. Edited by Reuben
Brower, Cambridge Mass.: Harvard University Press, 1959.
Jaspers, Karl. The Great Philosophers. Translated by R. Manheim. ed. H. Arendt. New York: Harcourt Brace & World, 1962.
Johnston, Mark. “Dispositional Theories of Value”. Proceedings of the Aristotelian Society 63, no. supp (1989): 139-174.
Jones, Judge. “TAMMY KITZMILLER, et al. vs. DOVER AREA SCHOOL DISTRICT, et al.,”. IN THE UNITED STATES
DISTRICT COURT FOR THE MIDDLE DISTRICT OF PENNSYLVANIA, no. Case No. 04cv2688 (2005).
Kaiser-Cooke, Michéle. “Murder in the Laboratory—Termhood and the Culture Gap.” In Translation as Intercultural
Communication: Selected Papers from the EST Congress, Prague 1995, eds. M. Snell-Hornby, Z. Jettmarova and K. Kaindl,
Benjamins translation library, v. 20, 285-290. Amsterdam ; Philadelphia: J. Benjamins, 1997.
Kant, Immanuel. Groundwork of the Metaphysics of Morals. Translated by H. J. Paton. New York: Harper Torchbooks, 1956. 1785.
———. Critique of Judgment. Translated by J. H. Bernard. Hafner Library of Classics. New York; London: Hafner Press, Collier
Macmillan, 1974. 1790.
———. Critique of Practical Reason. Translated by M. Gregor. Cambridge Texts in the History Of Philosophy, ed. K. Ameriks and D. M.
Clarke. New York: Cambridge, 1997. 1788.
Kaplan, David. “On Demonstratives.” In Themes from Kaplan, eds. J. Almog, J. Perry, H. K. Wettstein and D. Kaplan, 481-563.
New York: Oxford University Press, 1989.
Katz, Jerrold J. The Metaphysics of Meaning. Representation and Mind. Cambridge, Mass.: MIT Press, 1990.
———. “Precis of Metaphysics of Meaning.” In Naturalism and Normativity, ed. E. Villanueva, 128-34. Atascadero, CA: Ridgeview
Pub. Co., 1993a.
———. “Reply to Gibson.” In Naturalism and Normativity, ed. E. Villanueva, 174-179. Atascadero, CA: Ridgeview Pub. Co.,
1993b.
———. Sense, Reference, and Philosophy. Oxford ; New York: Oxford University Press, 2004.
Kirk, Robert. Translation Determined. Oxford Oxfordshire; New York: Clarendon Press; Oxford University Press, 1986.
Koller, Werner. “Equivalence in Translation Theory.” In Readings in Translation Theory, trans. A. Chesterman, ed. A. Chesterman,
200. Helsinki: Oy Finn Lectura, 1989 Einfiihrung in die Ubersetzungswissenschaft. 186-191. Heidelberg: Quelle & Meyer
Verlag, 1979.

431
Kornblith, H. “The Role of Intuition in Philosophical Inquiry: An Account with No Unnatural Ingredients.” In Rethinking
intuition: the Psychology of Intuition and Its Role in Philosophical Inquiry, eds. M. R. DePaul and W. Ramsey, 129-41. Lanham,
Md.: Rowman & Littlefield, 1998.
Kraut, Richard. “Two Conceptions of Happiness”. The Philosophical Review 88, no. 2 (1979): 167-197.
Kripke, Saul A. Naming and Necessity. Cambridge, Mass.: Harvard University Press, 1980.“near verbatim transcription of three
lectures that Kripke delivered in Princeton 1970,” 1972.
———. Wittgenstein on Rules and Private Language; An Elementary Exposition. Cambridge, Massachusetts: Harvard University Press,
1982.
Kuhn, Thomas S. The Structure of Scientific Revolutions. 2nd ed. Chicago: University of Chicago Press, 1970.
———. “Commensurability, Comparability, Communicability.” In Philosophy of Science Association, 1982, eds. P. D. Asquith and T.
Nickles, vol. 2, 669-88. East Lansing, Mich: Philosophy of Science Association, 1983.
Lafont, Cristina. The Linguistic Turn in Hermeneutic Philosophy. Studies in contemporary German Social Thought. Cambridge, Mass.: MIT
Press, 1999.
Lakoff, George. Women, Fire, and Dangerous Things: What Categories Reveal About the Mind. Chicago: University of Chicago Press,
1987.
Lance, Mark Norris, and John O'Leary-Hawthorne. The Grammar of Meaning: Normativity and Semantic Discourse. Cambridge studies in
philosophy. Cambridge, [England] ; New York, NY: Cambridge University Press, 1997.
LaPorte, Joseph. Natural Kinds and Conceptual Change. Cambridge studies in philosophy and biology. Cambridge; New York: Cambridge
University Press, 2004.
Laudan, Larry. “The Demise of the Demarcation Problem.” In But is it Science?: The Philosophical Question in the Creation/Evolution
Controversy, ed. M. Ruse, 337-350. Buffalo, N.Y.: Prometheus Books, 1988.
Laviosa-Braithwaite, Sara. “Universals of Translation.” In Routledge Encyclopedia of Translation Studies, eds. M. Baker and K.
Malmkjær, 288-291. London ; New York: Routledge, 2001.
Leibniz, Gottfried Wilhelm. “From Of the Art of Combination”. In Logical Papers: a Selection Translated and Edited with an Introduction by
G.H.R. Parkinson, ed. G. H. R. Parkinson, 1-11. Oxford: Clarendon, 1966. De Arte Combinatoria. 1666.
———. Writings on China. Translated by D. J. Cook and H. Rosemont. Chicago, Ill.: Open Court, 1994.
Lewis, David. “Scorekeeping in a Language Game”. Journal of Philosophical Logic 8 (1979): 339--359.
———. “Elusive Knowledge”. Australasian journal of philosophy 74, no. 4 (1996): 549-567.
Lienhardt, Godfrey. “Modes of Thought.” ed. E. E. Evans-Pritchard, 95-107. Oxford: Basil Blackwell, 1958.
Lilly, Reginald. “Translator’s Introduction.” In (Heidegger's) The Principle of Reason, Studies in Continental Thought, vii-xix.
Bloomington: Indiana University Press, 1991: Originally delivered as a lecture to the University of Freiburg as Der Satz
vom Grund, 1956.
Locke, John. An Essay Concerning Human Understanding. ed. P. H. Nidditch. Oxford; New York: Clarendon Press; Oxford
University Press, 1979. 1690.
Lycan, William G. Modality and meaning. Studies in linguistics and philosophy ; v. 53. Dordrecht ; Boston: Kluwer Academic, 1994.
Mackie, J. L. Ethics: Inventing Right and Wrong. Harmondsworth ; New York: Penguin, 1977.
Madsen, Dorte. “A Model for Translation of Legal Texts.” In Translation as Intercultural Communication: Selected Papers from the EST
Congress, Prague 1995, eds. M. Snell-Hornby, Z. Jettmarova and K. Kaindl, Benjamins translation library, v. 20, 291-299.
Amsterdam ; Philadelphia: J. Benjamins, 1997.
Maier, Carol, and Francoise Massardier-Kenney. “Toward an Expanded Pedagogy of Specialized Translation.” In Scientific and
Technical Translation, eds. S. E. Wright and J. W. Leland D., American Translators Association Scholarly Monographs, Vol.6,
151-161. Amerstam/Philidelphia: John Benjamins, 2001, 1993.
Malone, Joseph L. The Science of Linguistics in the Art of Translation: some Tools from Linguistics for the Analysis and Practice of Translation.
SUNY Series in Linguistics. Albany, N.Y.: State University of New York Press, 1988.
Manu. The Laws of Manu (Manavadharmashastra). Translated by G.Buhler. Vol xxv. Sacred Books of the East, ed. M. Müller. Oxford:
Oxford University Press, 1886.
McCrindle, J. W., ed. Ancient India as Described by Megasthenes and Arrian. Calcutta and Bombay: Thacker, Spink, 1877.

432
McDowell, John. “Non-Cognitivism and Rule-Following.” In Wittgenstein: To Follow a Rule, eds. S. H. Holtzman and L. M.
Christopher, 141-162. London: Routledge and K Paul, 1981.
———. “Are Moral Requirements Hypothetical Imperatives.” In Meta-Ethics, ed. M. Smith, 103-120. Aldershot; Brookfield,
USA: Dartmouth, 1995 Proceedings of the Aristotelian Society. 52, 13–29. 1978.
———. “Aesthetic Value, Objectivity, and the Fabric of the World.” In Mind, Value, and Reality, 112-130. Cambridge: Harvard
University Press, 1998a.
———. “Two Sorts of Naturalism.” In Mind, Value, and Reality, 167-197. Cambridge: Harvard University Press, 1998b.
———. “Values and Secondary Qualities”. In Mind, Value, and Reality. Cambridge: Harvard University Press, 1998c.
McEvilley, Thomas. “Plotinus and Vijnanavada Buddhism”. Philosophy East and West 30, no. 2 (1980): 181-193.
McKenzie, John S. A Manual of Ethics. 6th ed. London: University Tutorial Press, 1929.
Megrab, R.A. “Standards of Textuality and the Translation of Hadith.” In Translating Sensitive Texts: Linguistic Aspects, ed. K.
Simms, Approaches to Translation Studies; 14, 231-238. Amsterdam; Atlanta: Rodopi, 1997.
Messick, Brinkley Morris. The Calligraphic State: Textual Domination and History in a Muslim Society. Comparative studies on Muslim
societies ; 16. Berkeley: University of California Press, 1993.
Mill, John Stuart. Collected Works of John Stuart Mill. 33 vols., ed. J. M. Robson, F. E. Mineka, M. Filipiuk and J. O‘Grady. Toronto:
University of Toronto Press, 1965.
Miller, Alexander. Philosophy of Language. ed. . London: UCL Press, 1998.
Miller, Seumas. “Social Institutions.” ed. E. N. Zalta,The Stanford Encyclopedia of Philosophy. May 28 2007. Available from
http://plato.stanford.edu/archives/spr2007/entries/social-institutions/ (2007).
Monk, Ray. Ludwig Wittgenstein: the Duty of Genius. London: Jonathan Cape, 1990.
Montague, Richard. “The Proper Treatment of Quantification in Ordinary English.” In Formal Semantics: the Essential Readings,
eds. P. Portner and B. H. Partee, 17-34. Oxford: Blackwell, 2002 Approaches to Natural Language. 221-242. Dordrecht:
Reidel, 1973.
Moore, George Edward. Principia Ethica. Cambridge: University Press, 1903.
Mullin, Glenn H. “The Author Versus the Subject as Focus in Tibetan Translation.” In Buddhist Translation: Problems and
Perspectives, ed. D. Tulku, 215-221. New Delhi: Manohar, 1995.Originally presented at International Seminar on
"Buddhist Translations: Problems and Perspectives" (1990 : Delhi, India).
Murdoch, Iris. The Sovereignty of Good. London: Routledge & K. Paul, 1970.
Napper, Elizabeth. “Styles and Perspectives in Translation.” In Buddhist Translation: Problems and Perspectives, ed. D. Tulku, 35-42.
New Delhi: Manohar, 1995.Originally presented at International Seminar on "Buddhist Translations: Problems and
Perspectives" (1990 : Delhi, India).
Neubert, Albrecht. “Competence in Translation: a Complex Skill, How to Study it and Teach it.” In Translation Studies: an
Interdiscipline, eds. M. Snell-Hornby, F. Pèochhacker and K. Kaindl, Benjamins translation library v. 2, viii, 438 p.
Amsterdam ; Philadelphia: J. Benjamins, 1994a.
———. “‘A House of Many Rooms’: The Range of Translation Studies.” In Translating Slavery: Gender and Race in French Women's
Writing, 1783-1823, eds. D. Y. Kadish and F. Massardier-Kenney, Transation Studies 2. Kent, Ohio: Kent State
University Press, 1994b.
Neubert, Albrecht, and Gregory M. Shreve. Translation as Text. Translation studies ; 1. Kent, Ohio: Kent State University Press,
1992.
Newmark, Peter. “Communicative and Semantic Translation.” In Readings in Translation Theory, trans. A. Chesterman, ed. A.
Chesterman, 173-87. Helsinki: Oy Finn Lectura, 1989.
Nida, Eugene Albert. Toward a Science of Translating, with Special Reference to Principles and Procedures Involved in Bible Translating. Leiden:
E.J. Brill, 1964.
———. “Translating a Text with a Long and Sensitive Tradition.” In Translating Sensitive Texts: Linguistic Aspects, ed. K. Simms,
Approaches to Translation Studies; 14, 189-196. Amsterdam; Atlanta: Rodopi, 1997.
Nietzsche, Friedrich. “On the Problem of Translation.” In Theories of Translation: an Anthology of Essays from Dryden to Derrida, trans.
P. Mollenhauer, eds. R. Schulte and J. Biguenet, 68-70. Chicago: University of Chicago Press, 1992 Die Fröhliche
Wisssenchaf. 1882.

433
Nussbaum, Martha Craven. The Fragility of Goodness: Luck and Ethics in Greek Tragedy and Philosophy. Cambridge; New York:
Cambridge University Press, 1986.
———. Love's Knowledge: Essays on Philosophy and Literature. New York: Oxford University Press, 1990a.
———. The Poetics of therapy : Hellenistic Ethics in its Rhetorical and Literary Context. Edmonton, Alta.: Academic Printing & Pub.,
1990b.
———. Poetic Justice: the Literary Imagination and Public Life. Boston, Mass.: Beacon Press, 1995.
Odenstedt, Anders. “Gadamer on Context-Dependence”. The Review of Metaphysics 57 (2003).
Ogden, C. K., and I. A. Richards. The Meaning of Meaning: a Study of the Influence of Language upon Thought and of the Science of Symbolism.
San Diego: Harcourt Brace Jovanovich, 1923.
Oost, Stewart Irvin. “The Death of the Emperor Gordian III”. Classical Philology 53, no. 2 (1958): 106-107.
Ott, Hugo. Martin Heidegger: a Political Life. Hammersmith, London; New York, N.Y.: HarperCollinsPublishers; BasicBooks, 1993.
Oxford University. “context, n.” In Oxford English Dictionary. http://www.oed.com/: Oxford University Press, 2006.
Pappas, Nickolas. Routledge Philosophy Guidebook to Plato and the Republic. 2nd ed., Routledge philosophy guidebooks. London ; New York:
Routledge, 2003.
Parfit, Derek. Reasons and Persons. Oxford Oxfordshire: Clarendon Press, 1984.
Partee, Barabara H. “The Development of Formal Semantics.” In The Handbook of Contemporary Semantic Theory, ed. S. Lappin, 11-
38. Oxford, OX, UK ; Cambridge, Mass., USA: Blackwell Reference, 1997.
Phillips, David. “The Middle Ground in Moral Semantics”. American Philosophical Quarterly 35, no. 2 (1998): 141-155.
Pinchuck, Isadore. Scientific and Technical Translation. London: A. Deutsch, 1977.
Popper, Karl. “Darwinism as a Metaphysical Research Program.” In But is it Science?: The Philosophical Question in the
Creation/Evolution Controversy, ed. M. Ruse, 144-155. Buffalo, N.Y.: Prometheus Books, 1988 Philosophy of Science. 44,
638-61. no. 4 (1977).
Predelli, Stefano. “'I Am Not Here Now'”. Analysis 58, no. 2 (1998): 107-115.
Putnam, Hilary. “The Meaning of Meaning”. Minnesota Studies in the Philosophy of Science VII (1975): 131-193.
———. “Why Reason Can't Be Naturalized”. Synthese 52, no. 1 (1982): 3-23.
———. The Many Faces of Realism. La Salle, Ill.: Open Court, 1987.
———. Representation and Reality. Cambridge, Mass.: MIT Press, 1988.
———. “Reply to Gerald Massey.” In Indeterminacy of Translation, ed. D. Føllesdal, Philosophy of Quine, vol. 3, 357-366. New York:
Garland Pub., 2001 Philosophical Topics. 20, (1992): 399-402.
———. The Collapse of the Fact/Value Dichotomy and Other Essays. Cambridge, MA: Harvard University Press, 2002.
———. Ethics Without Ontology. Cambridge, Mass.: Harvard University Press, 2004.
Pym, Anthony. “Translation Studies and Western Philosophy (Forthcoming: Pre-print draft 3.0).” In The Companion to Translation
Studies (Multilingual Matters), eds. P. Kuhiwczak and K. Littau., 2002.<http://www.fut.es/~apym/on-
line/philosophy.pdf>
———. Translation and Text Transfer: an Essay on the Principles of Intercultural Communication. Frankfurt am Main: Peter Lang, 1992.
———. The Moving Text: Localization, Translation, and Distribution. Vol 49. Benjamins Translation Library,. Amsterdam ; Philadelphia:
John Benjamins Pub. Co., 2004.
Quine, Willard Van Orman. Word and Object. Cambridge: MIT Press, 1960.
———. “Ontological Relativity, the Dewey Lectures 1968”. Journal of Philosophy 65 (1968): 185-212.
———. “Natural Kinds.” In Ontological Relativity, and Other Essays, 114-138. New York,: Columbia University Press, 1969a.
———. “Reply to Chomsky.” In Words and Objections: Essays on the work of W.V. Quine, eds. D. Davidson, W. V. O. Quine and K.
K. J. Hintikka, 302-11. Dordrecht: D. Reidel, 1969b.
———. “Two Dogmas of Empiricism.” In Classics of Analytic Philosophy, ed. R. R. Ammerman, 196-213. Indianapolis: Hacket,
1990 Philosophical Review. 60, no. 1 (1951): 20-43.

434
———. “Meaning.” In The Philosophy of Language, ed. A. Martinich, 446-455. New York: Oxford University Press, 1996 Pursuit of
Truth. 37-59. Cambridge, MA: Harvard University Press, 1992.
Quinn, Philip L. “Moral Obligation, Religious Demand, and Practical Conflict.” In Rationality, Religious Belief, and Moral
Commitment; New Essays in the Philosophy of Religions, eds. R. Audi and W. J. Wainwright, 195-212. London: Cornell
University Press, 1986.
Railton, Peter Albert. “Moral Realism”. Philosophical Review 95 (1986): 163-207.
———. “Made in the Shade: Moral Compatibilism and the Aims of Moral Theory.” In On the Relevance of Metaethics: New Essays on
Metaethics, eds. J. Couture and K. Nielsen, Canadian Journal of Philosophy. Supplementary volume ; 21, 79-106. Calgary:
University of Calgary Press, 1996.
Rajchman, John, and Cornel West. Post-Analytic Philosophy. New York: Columbia University Press, 1985.
Ranganathan, Shyam. Ethics and the History of Indian Philosophy. Delhi: Motlial Banarsidass, 2007a.
———. Patanjali: Yoga Sutra: with a Translation, Commentary and Introduction. Delhi: Penguin, 2007b.
Rawls, John. A Theory of Justice. Cambridge, Mass.: Harvard University Press, 1971.
———. “The Independence of Moral Theory”. Proceedings and Addresses of the American Philosophical Association 48 (1975): 5-22.
———. Political Liberalism. Paperback ed., John Dewey Essays in Philosophy; no. 4. New York: Columbia University Press, 1996.
Read, Ruper, and Jon Cook. “Recent Work: the Philosophy of Literature”. Philosophical Books 42, no. 2 (2001): 118-131.
Recanati, François. “Literal/nonliteral”. Midwest Studies in Philosophy 25 (2002): 264-274.
———. Literal Meaning. Cambridge ; New York: Cambridge University Press, 2004.
Reichenbach, Hans. Elements of Symbolic Logic. New York: Free Press, 1947.
Resnik, David B. “A Pragmatic Approach to the Demarcation Problem”. Studies in History and Philosophy of Science Part A 31, no. 2
(2000): 249-267.
Ridge, Michael. “Ecumenical Expressivism: Finessing Frege”. Ethics (2006): 302-336.
Ross, W. D. “The Basis of Objective Judgments in Ethics”. International Journal of Ethics 37, no. 2 (1927): 113-127.
———. The Right and the Good. Oxford: The Clarendon press, 1963. 1930.
Ruse, Michael. “Biological Species: Natural Kinds, Individuals, or What?” British Journal of the Philosophy of Science 38 (1987): 225-
242.
Ruse, Michael, and Robert Pennock, eds. But is it Science?: The Philosophical Question in the Creation/Evolution Controversy. Buffalo,
N.Y.: Prometheus Books, 2008.
Russell, Bertrand. The Analysis of Mind. London: Allen and Unwin, 1921.
Ryle, Gilbert. “Thinking and Reflecting”. In Collected Papers, 2, 465-479. London: Hutchinson, 1971a. University Lectures. no. 18
Saskatoon: University of Saskatchewan, 1968.
———. “The Thinking of Thoughts: What Is 'Le Penseur' Doing?” In Collected Papers, 2, 480-496. London: Hutchinson, 1971b.
University Lectures. no. 18 Saskatoon: University of Saskatchewan, 1968.
———. Collected Papers. 2 vols. London: Hutchinson, 1971†.
Sacks, Harvey, and Gail Jefferson. Lectures on Conversation. Oxford, UK ; Cambridge, Mass.: Blackwell, 1995.
Sager, Juan C. “Terminology—applications.” In Routledge Encyclopedia of Translation Studies, eds. M. Baker and K. Malmkjæar, 251-
255. London ; New York: Routledge, 2001.
Sallis, John. On Translation. Studies in Continental Thought. Bloomington: Indiana University Press, 2002.
Salmon, Nathan U. “Two Conceptions of Semantics.” In Semantics vs. Pragmatics, ed. Z. G. Szabó, 317-328. Oxford; New York:
Clarendon Press; Oxford University Press, 2005.
Sarcevic, Susan. “Creativity in Legal Translation: How Much is Too Much?” In Translation in Context: elected contributions from the
EST Congress, Granada, 1998, eds. A. Chesterman, N. Gallardo San Salvador and Y. Gambier, Benjamins translation
library, v. 39, 281-292. Amsterdam ; Philadelphia: John Benjamins, 2000.
Sartre, Jean Paul. “Existentialism is a Humanism.” In Existentialism from Dostoevsky to Sartre, trans. P. Mairet, ed. W. A. Kaufmann,
345-369. Toronto: Meridian, 1975 L’existentialisme est un humanisme. London: Methuen and Co., 1964.

435
Sarukkai, Sundar. Translating the World: Science and Language. Lanham: University Press Of America, 2002.
Savory, Theodore Horace. The Art of Translation. 1st American ed. Boston: The Writer, 1968.
Scheffler, Samuel. “Morality Through Thick and Thin”. Philosophical Review 96, no. 3 (1987): 411-434.
Schleiermacher, Friedrich Daniel Ernst. “From On the Different Methods of Translating.” In Theories of Translation: an Anthology of
Essays from Dryden to Derrida, trans. W. Bartscht, eds. R. Schulte and J. Biguenet. Chicago: University of Chicago Press,
1992 Friedrich Schleiermacher sämmtliche Werke, Dritte Abtheilung: Zur Philosophy. 2, 207-45. Berlin: Reimer, 1938.
Schlenker, Philippe. “A Plea for Monsters”. Linguistics and Philosophy 26 (2003): 29–120.
Schmidt, Dennis. “Some Reflections on Translating Philosophy.” In Translation Perspectives, vol. 1, 28-34. Binghamton: SUNY
Press, 1984.
Schopenhauer, Arthur. “On Language and Words.” In Theories of Translation: an Anthology of Essays from Dryden to Derrida, trans. P.
Mollenhauer, eds. R. Schulte and J. Biguenet. Chicago: University of Chicago Press, 1992“Über Sparch unt Worte,”
Parerga und Paralipomena. 1800.
Schroeder, Mark. “Expression for Expressivists.” Philosophy and Phenomenological Research. Available from http://www-
rcf.usc.edu/~maschroe/research/ (forthcoming).
Scott, W. Richard. Institutions and Organizations. 2nd ed., Foundations for organizational science. Thousand Oaks, Calif.: Sage
Publications, 2001.
Searle, John. “Literal Meaning”. Erkenntnis 13 (1978): 207-24.
———. “The Background of Meaning.” In Speech Act Theory and Pragmatics, eds. F. Kiefer and M. Bierwisch, Synthese Language
Library, vol. 10, xii, 317. Dordrecht, Holland; Boston, Hingham, MA: D. Reidel; Kluwer Boston, 1980.
Seth, James. A Study of Ethical Principles. 8th ed. London: W. Blackwood and Sons, 1928.
Shianna, Kevin V., and Huntington F. Willard. “Human Genomics: In search of Normality”. 444, no. 7118 (2006): 428-429.
Silliman, Ronald. The Age of Huts. New York, N.Y.: Roof, 1986.
Silverstein, Michael. “Language Structure and Linguistic Ideology.” In The Elements: A Parasession on Linguistic Units and Levels, eds.
P. R. Clyne, W. F. Hanks and C. L. Hofbauer, 193-247. Chicago: Chicago Linguistic Society, 1979.
———. “Metapragmatic Discourse and Metapragmatic Function.” In Reflexive Language: Reported Speech and Metapragmatics, ed. J.
A. Lucy, 33-58. Cambridge, England: Cambridge U Press, 1993.
———. “The Limits of Awareness.” In Linguistic Anthropology : a Reader, ed. A. Duranti, 382-401. Malden, Mass.: Blackwell
Publishers, 2001“The Limits of Awareness,” Sociolinguistic Working Paper No.84. Austin: Southwest Educational
Development Laboratory, 1981.
Skorupski, John. “The Definition of Morality”. Philosophy 35(Supp) (1993): 121-144.
Smart, John Jamieson Carswell. “Quine's Philosophy of Science.” In Words and Objections: Essays on the Work of W.V. Quine, eds.
D. Davidson and K. K. J. Hintikka, 3-13. Dordrecht: D. Reidel, 1969.
Smith, David. “Nietzsche’s Hinduism, Nietzsche’s India: Another Look”. Journal of Nietzsche Studies, no. 28 (2004): 37-56.
Smith, Michael. The Moral Problem. Cambridge: Blackwell, 1994.
Snell-Hornby, Mary, ed. Translation Studies: an Integrated Approach. Amsterdam; Philadelphia: J. Benjamins, 1988.
Soames, Scott. “Direct Reference, Propositional Attitudes, and Semantic Content.” In Propositions and Attitudes, eds. N. U. Salmon
and S. Soames, 197-239. Oxford; New York: Oxford University Press, 1988 Philosophical Topics. 15, (1987): 47-87.
———. Philosophical analysis in the twentieth century. 2 vols. Princeton, N.J.: Princeton University Press, 2003.
Sperber, Dan, and Deirdre Wilson. Relevance: Communication and Cognition. Oxford: Blackwell, 1986.
Spinoza, Benedictus de. Tractatus Theologico-Politicus. Translated by S. Shirley. Indianapolis, Ind.: Hackett Pub., 1998.
Sprigg, T. L. S. “Definition of a Moral Judgment.” In The Definition of Morality, eds. G. Wallace and A. D. M. Walker, 119-145.
London: Methuen and Co., 1970 Philosophy. 301–22. (1964).
Sreenivasan, Gopal. “Understanding Alien Morals”. Philosophy and Phenomenological Research 62, no. 1 (2001): 1-32.
Stanley, Jason. “Context and Logical Form”. Linguistics and Philosophy 23 (2000): 391-434.
———. “On the Linguistic Basis for Contextualism”. Philosophical Studies 119 (2004): 119-146.

436
———. Knowledge and Practical Interests. Oxford: Oxford University Press, 2005.
Stanley, Jason, and Timothy Williamson. “Quantifiers and Context-Dependence”. Analysis 55.4 (1995): 291-295.
Steiner, George. After Babel: Aspects of Language and Translation. London ; New York: Oxford University Press, 1975.
Stevenson, Charles L. Ethics and Language. New Haven: Yale University Press, 1944.
Stone, Jim. “Parfit and the Buddha: Why There Are No People”. Philosophy and Phenomenological Research 48, no. 3 (1988): 519-532.
Sweetser, Eve. From Etymology to Pragmatics: Metaphorical and Cultural Aspects of Semantic Structure. Cambridge studies in linguistics ; 54.
Cambridge ; New York: Cambridge University Press, 1990.
Tappolet, Christine. “Through Thick and Thin: Good and its Determinates”. Dialectica 58, no. 2 (2004): 207-221.
Tarski, Alfred. “The Semantic Conception of Truth: and the Foundations of Semantics”. Philosophy and Phenomenological Research 4,
no. 3 (1944): 341-376.
Taylor, Talbot J. Mutual Misunderstanding: Scepticism and the Theorizing of Language and Interpretation. Post-contemporary interventions.
Durham: Duke University Press, 1992.
———. “Enculturating Language.” In Theorizing language: Analysis, Normativity, Rhetoric, History, 1-26. Amsterdam ; New York:
Pergamon, 1997a.
———. “Normativity and Linguistic Form.” In Theorizing language: Analysis, Normativity, Rhetoric, History, 139-168. Amsterdam ;
New York: Pergamon, 1997b.
———. “Why We Need a Theory Of Language.” In Theorizing language: Analysis, Normativity, Rhetoric, History, 169-182.
Amsterdam ; New York: Pergamon, 1997c.
———. Theorizing language: Analysis, Normativity, Rhetoric, History. 1st ed. Amsterdam ; New York: Pergamon, 1997±.
Ten, C.L. “Hume’s Racism and Miracles”. Journal of Value Inquiry 36 (2002): 101–107.
Thomson, Judith Jarvis. “The Right and the Good”. The Journal of Philosophy 94, no. 6 (1997): 273-298.
Timmons, Mark. Morality Without Foundations : a Defense of Ethical Contextualism. New York: Oxford University Press, 1998.
Traugott, Elizabeth Closs, and Richard B. Dasher. Regularity in Semantic Change. Cambridge studies in linguistics; 96. Cambridge, U.K;
New York: Cambridge University Press, 2002.
Travis, Charles. “On What is Strictly Speaking True”. Canadian Journal of Philosophy 15, no. 2 (1985): 187-229.
———. “Meaning's Role in Truth”. Mind 100, no. 451-66 (1996).
Tsering, Migmar. “Co-Translations: Its Value and Problems.” In Buddhist Translation: Problems and Perspectives, ed. D. Tulku, 205-
214. New Delhi: Manohar, 1995.Originally presented at International Seminar on "Buddhist Translations: Problems
and Perspectives" (1990 : Delhi, India).
Tulku, Doboom, ed. Buddhist Translation: Problems and Perspectives. International Seminar on "Buddhist Translations: Problems and
Perspectives" (1990 : Delhi, India). New Delhi: Manohar, 1995.
van Roojen, Mark. “Morality without Foundations: A Defense of Ethical Contextualism (Review)”. The Philosophical Review 110,
no. 2 (2001): 283-286.
van Roojen, Mark. “Knowing Enough to Disagree: A New Response to the Moral Twin Earth Argument.” In Oxford Studies in
Metaethics, ed. R. Shafer-Landau, vol. I. Oxford: Oxford University Press, 2006.
Velleman, J. David. “The Possibility of Practical Reason”. Ethics 106, no. 4 (1996): 694-726.
Venuti, Lawrence. The Scandals of Translation: Towards an Ethics of Difference. London ; New York: Routledge, 1998.
Vermeer, Hans J. “Skopos and Commission in Translational Action.” In Readings in Translation Theory, trans. A. Chesterman, ed. A.
Chesterman, 173-87. Helsinki: Oy Finn Lectura, 1989.
Wedgwood, Ralph. “Conceptual Role Semantics for Moral Terms”. Philosophical Review 110, no. 1 (2001): 1-30.
———. “Meaning of 'Ought'.” In Oxford Studies in Metaethics, ed. R. Shafer-Landau, vol. 1, 127-160. Oxford: Clarendon Press,
2006.
Weinberg, Jonathan, Shaun Nichols, and Stephen Stich. “Normativity and Epistemic Intuitions”. Philosophical Topics 29 (2001):
429-60.

437
Wheeler, Samuel C. Deconstruction as Analytic Philosophy. Cultural memory in the present. Stanford, Calif.: Stanford University Press,
2000.
Whorf, Benjamin Lee. “The Relation of Habitual Thought and Behavior to Language.” In Linguistic Anthropology : a Reader, ed. A.
Duranti, Blackwell Anthologies in Social and Cultural Anthropology, 363-381. Malden, Mass.: Blackwell Publishers, 2001
Edited by J. B. Carroll, (1939).
Wiggins, David. “A Sensible Subjectivism”. In Needs. Values. Truth. Oxford: Basil Blackwell, 1991.
Wilder, Hugh T. “Quine on Natural Kinds”. Australasian Journal of Philosophy 3 (1972): 263 - 270.
Williams, Bernard. Ethics and the Limits of Philosophy. Cambridge: Harvard University Press, 1985.
———. “Truth in Ethics.” In Truth in Ethics, ed. B. Hooker, 19-34. Cambridge: Blackwell, 1996 Ratio. 8, no. 3 (1995): 227-242.
Wilson, Deirdre, and Dan Sperber. “Linguistic Form and Relevance”. Lingua 90 (1993): 1-25.
Wilson, N.L. “Substance Without Substrata”. Review of Metaphysics XII, no. 4 (1959): 521-539.
Wittgenstein, Ludwig. Philosophical Investigations. Translated by G. E. M. Anscombe. 2nd ed. New York: Macmillan, 1958.
———. Preliminary Studies for the “Philosophical investigations”: Generally Known as the Blue and Brown Books. 2nd ed. New York: Harper
& Row, 1960. 1958.
———. Culture and Value. ed. G. H. v. Wright and H. Nyman. Chicago: University of Chicago Press, 1980.
Wolf, Michaela. “Translation as a Process of Power: Aspects of Cultural Anthropology in Translation.” In Translation as
Intercultural Communication: Selected Papers from the EST Congress, Prague 1995, eds. M. Snell-Hornby, Z. Jettmarova and K.
Kaindl, Benjamins Translation Library, v. 20, 123-133. Amsterdam ; Philadelphia: J. Benjamins, 1997.
Wolff, Robert P. “Kant's Debt to Hume Via Beattie”. Journal of the History of Ideas 21, no. 1 (1960): 117-123.
Wong, David B. Moral Relativity. Berkeley: University of California Press, 1984.
Wright, Crispin. Truth and Objectivity. Cambridge, Mass. ; London: Harvard University Press, 1992.
Wright, Sue Ellen, and Jr. Wright Leland D. Scientific and Technical Translation. 2001. American Translators Association Scholarly
Monographs, Vol.6, 1993.
Zhang, Wei. Heidegger, Rorty, and the Eastern thinkers: a Hermeneutics of Cross-Cultural Understanding. SUNY series in Chinese philosophy
and culture. Albany: State University of New York Press, 2006.

438