OBJECTS AND CONCEPTS

Downloaded from https://academic.oup.com/aristoteliansupp/article-abstract/68/1/131/1774421 by guest on 25 October 2019

Dagfinn F0llesdal and David Bell

/—Dagfinn F0llesdal

INDIVIDUATION

I will begin this paper with a quote from Godel that seems
puzzling, but becomes more intelligible when looked upon from
a Husserlian perspective, a perspective which in a quite different
setting can also be found in Quine. Surprisingly, these three authors,
Godel, Quine and Husserl, who are often thought of as being poles
apart, will be seen to be kindred spirits.
After having developed this Husserlian perspective, I will dis-
cuss the various factors that come together in individuation and how
they are connected with the notion of laws of nature. At the end of
the paper I will relate these issues to the semantics of singular terms
and de re prepositional attitudes. Since I have written on these latter
topics before, that part of the paper will be just a brief summary.
I
Godel, Quine, Husserl. In the supplement that Godel added to
'What is Cantor's continuum problem?' when it was reprinted in
Benacerraf and Putnam, Philosophy of Mathematics: Selected
Readings in 1964, Godel states:
That something besides the sensations actually is immediately
given follows (independently of mathematics) from the fact that
even our ideas referring to physical objects contain constituents
qualitatively different from sensations or mere combinations of
sensations, e.g., the idea of object itself... Evidently, the 'given'
underlying mathematics is closely related to the abstract elements
contained in our empirical ideas. It by no means follows, however,
that the data of this second kind, because they cannot be associated
with actions of certain things upon our sense organs, are something
purely subjective, as Kant asserted. Rather they, too, may represent
an aspect of objective reality, but, as opposed to the sensations, their
132 I—DAGFINN F0LLESDAL

presence may be due to another kind of relationship between

Downloaded from https://academic.oup.com/aristoteliansupp/article-abstract/68/1/131/1774421 by guest on 25 October 2019

ourselves and reality.
Charles Parsons, whose work on the philosophy of mathematics
I find exceptionally illuminating, said at an earlier meeting of the
Aristotelian Society that he found this passage 'quite obscure'.1
However, I mink that there is a way of making sense of the passage,
and that it expresses in a condensed way a key to Godel's philo-
sophy of mathematics.
Godel's point in this passage is, I think, related to an insight that
Quine expresses in the following form, in two recent unpublished
manuscripts that he has permitted me to quote:
As Donald Campbell put it, reification of bodies is innate in man
and the other higher animals. I agree, subject to a qualifying
adjective: perceptual reification. I reserve ''full reification' and
'full reference' for the sophisticated stage where the identity of
a body from one time to another can be queried and affirmed
or conjectured or denied independently of exact resemblance.
Distinct bodies may look alike, and an identical object may
change its aspect. Such discriminations and identifications
depend on our elaborate theory of space and time and of
unobserved trajectories of bodies between observations.
(Quine, 'Reactions', manuscript p. 6, to be published in Paolo
Leonardi and Marco Santambrogio, eds., On Quine (Cambridge:
Cambridge University Press, 1994). In this and the next quotation
the italics are Quine's, the bold type is mine.)
I wonder whether a dog ever gets beyond this stage. He recognizes
and distinguishes recurrent people, but this is a qualitative matter
of scent. Our sophisticated concept of recurrent objects,
qualitatively indistinguishable but nevertheless distinct, in-
volves our elaborate schematism of intersecting trajectories in
three-dimensional space, out of sight, trajectories traversed
with the elapse of time. These concepts of space and time, or the
associated linguistic devices, are further requisites on the way to
substantial cognition. ('From stimulus to science', Lehigh Uni-
versity, Oct. 15, 1990, Franklin and Marshall College, April 17,
1992, manuscript p. 21.)

Charles Parsons, 'Mathematical Intuition', Meeting of the Aristotelian Society held at

5/7 Tavistock Place on Monday, March 17,1980. Proceeding of the Aristotelian Society
N.S., 80 (1979-80), p. 146.
 OBJECTS AND CONCEPTS 133

I will now expand further on this idea of reification and its

Downloaded from https://academic.oup.com/aristoteliansupp/article-abstract/68/1/131/1774421 by guest on 25 October 2019

relevance to mathematics, starting out with a brief exposition of
Husserl's idea of Wesensschau. This idea was basic for Husserl's
philosophy of mathematics and will help us see more clearly some
of the issues connected with reification.
According to Husserl, as for Quine in the passages I just quoted,
we impose a structure on the world. The excitation of our senses is
not sufficient to uniquely determine what we perceive. This point
can be illustrated by the Jastrow/Wittgenstein duck/rabbit example.
The example is normally given in terms of a picture which can be
seen as a duck or a rabbit. Wittgenstein, unfortunately, transfers
this to perception and says that all seeing is seeing as. For Husserl,
perception does not involve something given, for example sense
data, that can be seen as something. To get this point right, let us
use a version of the duck/rabbit example where we are dealing with
a silhouette protruding over the horizon.
In this situation we can see a duck or a. rabbit. There is nothing
primitively given, like sense data, that can be taken in different
ways, seen as a duck or as a rabbit. Husserl argues against such
views and holds that there is nothing out there in the world that is
merely given and not structured by us.
Although there are no sense data our there, which are organized
into patterns through our structuring activity and which constrain
the ways in which we can see the world, we are not free to structure
the world in any way we might want. Perception is constrained, not
by sense data out there, but by certain experiences (hyle) which we
have when our senses are stimulated, to use a term that is as
appropriate for Husserl as it is for Quine. The stimulation, or
excitation, of the senses is not what we see or perceive. We perceive
physical objects directly, Husserl insists, without any intervention
of sense data. (In this sense, Husserl is a physicalist, with the
refinement, to which we shall return, that he maintains that not only
physical objects, but also persons, actions, etc., are experienced
directly. However, he is not a physicalist in the ontological sense;
there are many kinds of entities in the world, in addition to physical
objects, for example, abstract entities of various kinds.)
These experiences which we have when our senses are impinged
upon, constrain our structuring when we perceive, but they play no
such role when we think, fantasize, remember and so on.
134 I—DAGHNN F0LLESDAL

These constraints are, as we noted, far from sufficient to narrow

Downloaded from https://academic.oup.com/aristoteliansupp/article-abstract/68/1/131/1774421 by guest on 25 October 2019

the structuring possibilities down to one. Note that this under-
determination is there not only in the case of so-called 'ambiguous'
pictures or situations. All perception is like this, but we normally
do not notice it. That misperception is always possible, shows this.
When we discover that we have misperceived, we restructure what
we have in front of us in such a way that the new structure meets
the constraints of the present impingements upon our senses as well
as the past ones.
The structuring always takes place in such a way that the many
different features of the object are experienced as connected with
one another, as features of one and the same object. When, for
example, we see a rabbit, we do not merely see a collection of
coloured patches, various shades of brown spread out over our field
of vision (incidentally, even seeing coloured patches involves
intentionality, a patch is also a kind of object, but a different kind
of object than a rabbit). We see a rabbit, with a determinate shape
and a determinate colour, with the ability to eat, jump, etc. It has a
side that is turned toward us and one that is turned away from us.
We do not see the other side from where we are, but we see
something which has an other side.
It is this peculiarity of our consciousness that Husserl labels
intentionality, or directedness. That seeing is intentional, or object-
directed, means just this, that the near side of the object we have in
front of us is regarded as a side of a thing, and that the thing we see
has other sides and features that are co-intended, in the sense that
the thing is regarded as more than just this one side. The noema, to
use one of Husserl's technical terms, is the comprehensive system
of determinations that gives unity to this manifold of features and
makes them aspects of one and the same object.
It is important at this point to note that the various sides, ap-
pearances or perspectives of the object are constituted together with
the object. There are no sides and perspectives floating around
before we start perceiving, which are then synthesized into objects
when intentionality sets in. There are no objects of any kind,
whether they be physical objects, sides of objects, appearances of
objects or perspectives of objects, without intentionality. And
intentionality does not work in steps. We do not start by constituting
 OBJECTS AND CONCEPTS 135

six sides and then synthesize these into a die; we constitute the die

Downloaded from https://academic.oup.com/aristoteliansupp/article-abstract/68/1/131/1774421 by guest on 25 October 2019

and the six sides of it in one step.
The word 'object' must, as we have noted, be taken in a very
broad sense. It denotes not only physical things, but also, as we have
seen, animals, and likewise persons, events, actions, processes and
changes, and sides, aspects and appearances of such entities. There
are also abstract objects; we shall come to these later.
We should also note that when we experience a person, we do
not experience a physical object, a body, and then infer that a person
is there. We experience a full-fledged person, we are encountering
somebody who structures the world, experiences it from his or her
own perspective. Our noema is a noema of a person. Seeing persons
is no more mysterious than seeing physical objects, no inference is
involved in either case. When we see a physical object we do not
see sense data or the like and then infer that there is a physical object
there, our noema is the noema of a physical object. Similarly, when
we see an action, what we see is a full-fledged action, not a bodily
movement from which we infer that there is an action.

II

The world and the past. We constitute not only the different
properties of things, but also the relation of the thing to other
objects. If, for example, I see a tree, the tree is conceived of as
something which is in front of me, situated among other trees, seen
by other people than myself, etc. It is also conceived of as some-
thing which has a history: it was there before I saw it, it will remain
after I have left, perhaps it will eventually be cut and transported to
some other place. However, like all material things, it does not
simply disappear from the world.
My consciousness of the tree is in this way also a consciousness
of the world in space and time in which the tree is located. My
consciousness constitutes the tree, but at the same time it constitutes
the world in which the tree and I are living. If my further experience
makes me give up the belief that I have a tree in front of me because,
for example, I do not find a tree-like far side or because some of my
other expectations prove false, this affects not only my conception
of what there is, but also, as we noted above, my conception of what
has been and what will be. Thus in this case, not just the present, but
136 I—DAGFINN F0LLESDAL

also the past and the future are reconstituted by me. To illustrate how

Downloaded from https://academic.oup.com/aristoteliansupp/article-abstract/68/1/131/1774421 by guest on 25 October 2019

changes in my present perception lead me to reconstitute not just the
present, but also the past, Husserl uses an example of a ball which I
initially take to be red all over and spherical. As it turns, I discover
that it is green on the other side and has a dent:
the sense of the perception is not only changed in the momentary
new stretch of perception; the noematic modification streams back
in the form of a retroactive cancellation in the retentional sphere
and modifies the production of sense stemming from earlier phases
of the perception. The earlier apperception, which was attuned to
the harmonious development of the 'red and uniformly round', is
implicitly 'reinterpreted' to 'green on one side and dented'.2

in
Values, practicalfunction. So far, I have mentioned only the factual
properties of things. However, things also have value properties,
and these properties are constituted in a corresponding manner,
Husserl says. The world within which we live is experienced as a
world in which certain things and actions have a positive value,
others a negative. Our norms and values, too, are subject to change.
Changes in our views on matters of fact are often accompanied by
changes in our evaluations.
Husserl emphasizes that our perspectives and anticipations are
not predominantly factual. We are not living a purely theoretical
life. According to Husserl, we encounter the world around us
primarily 'in the attitude of the natural pursuit of life', as 'living
functioning subjects involved in the circle of other functioning
subjects' . 3 Husserl says this in a manuscript from 1917, but he has
similar ideas about the practical both earlier and later. Thus in the
Ideas (1913) he says:
this world is mere for me not only as a world of mere things, but
also with the same immediacy as a world of values, a world of
goods, a practical world.

2 Erfahrung und Urteil, §21 a, p. 96 = p. 89 of Churchill and Ameriks' English translation

(Evanston, 111.: Northwestern University Press, 1973).
3 Husserliana IV, 375.31-33.
4 Ideen, §27, Husserliana HI, 1, 58.13-19 = Kersten's English translation p. 53,1 have
changed his translation slightly.
 OBJECTS AND CONCEPTS 137

Just as Husserl never held that we first perceive sense data, or

Downloaded from https://academic.oup.com/aristoteliansupp/article-abstract/68/1/131/1774421 by guest on 25 October 2019

perspectives or appearances, which are then synthesized into
physical objects, or that we first perceive bodies and bodily move-
ments and then infer that there are persons and actions, so it would
be a grave misunderstanding of Husserl to attribute to him the view
that we first perceive objects that have merely physical properties
and then assign to them a value or a practical function. Things are
directly experienced by us as having the features, functional and
valuational as well as factual, that are of concern for us in our
natural pursuit of life.

IV
Intersubjectivity. Husserl emphasizes, early and late, that the world
we intend and thereby constitute, is not our own private world, but
an intersubjective world, common to and accessible to all of us.
Thus in the Ideas he writes:
I continually find at hand as something confronting me a spatio-
temporalreality[Wirklichkeit] to which I belong like all other
human beings who are to be found in it and who arerelatedto it as
lam.5
One of the many places where Husserl stresses the shared,
intersubjective nature of the world is §29 of the Ideas, entitled 'The
"Other" Ego-subjects and the Intersubjective Natural Surrounding
World'. He there says:
I take their surrounding world and mine Objectively as one and the
same world of which we are conscious, only in different ways
[Weise]... For all that, we come to an understanding with our fellow
human beings and together with them posit an Objective spatio-
temporalreality...6
In Husserl's later works one finds similar ideas, particularly in
the many texts that have been collected by Iso Kern in the three
volumes of the Husserliana devoted to intersubjectivity, but also
in many other works, for example in the Crisis:

5 Ideen, §30, Husserliana HI, 1, 61.15-18 = Kersten's translation, pp. 56-57, slightly
modified by me.

6 Ideen, §29, Husserliana HI, 1,60.16-26 = Kersten, pp. 55-56.

138 I—DAGFINN F0LLESDAL

Thus in general the world exists not only for isolated men but for
the community of men; and this is due to the fact that even what is

Downloaded from https://academic.oup.com/aristoteliansupp/article-abstract/68/1/131/1774421 by guest on 25 October 2019

straightforwardly perceptual is communal.
Husserl discusses in great detail empathy and the many other
varieties of intersubjective adaptation that enable us to intend a
common, intersubjective world. For these discussions I refer you
to the three volumes on intersubjectivity that I just mentioned.

V
Wesensschau. Once we acknowledge mat the impulses that reach
our sense organs do not determine uniquely what object we perceive,
it seems natural to take a further step and admit that the object may
well be an abstract entity. What object we experience depends on
what anticipations we have concerning further features of the object.
For example, standing in front of a triangularly shaped tree, our
anticipations may concern its triangularity rather than the tree or its
front side, its bark, one of its branches or leaves, etc. It does men not
matter that the tree is replaced by another thing of the same shape;
as long as the thing is triangular and hence fulfils our anticipations,
the object of the act remains the same. This characteristic, that the
object we experience may remain the same although the physical
object in front of us is replaced by a distinct physical object, differ-
entiates acts directed towards abstract objects from acts directed
towards physical objects. Objects of this abstract kind Husserl calls
eidos, or essences. He called the procedure that leads us to them,
where we disregard certain expectations and preserve only those that
pertain to the particular general feature or eidos that we are inter-
ested in, the eidetic reduction, since it leads us to the eidos.
Typically eidos, or essences, are studied in mathematics, but
Husserl suggested the possibility of studying various essences
which were in his time not studied in mathematics, for example,
some of those that nowadays are studied in topology, as well as
other general notions like 'humanity', etc.
The experience of essences Husserl calls essential insight, or
Wesensschau. There is hence nothing mysterious about Wesens-

7 Krisis, §47, Husserliana VI, 166.19-22 = Carr's translation, p.163.

 OBJECTS AND CONCEPTS 139

schau; it is on a par with ordinary perception. However, philo-

Downloaded from https://academic.oup.com/aristoteliansupp/article-abstract/68/1/131/1774421 by guest on 25 October 2019

sophers have been far too quick in passing from the premiss of
empiricism—knowledge reaches us through the senses—to the
conclusion that the only things we can know through our senses are
physical objects.
Husserl uses the phrase 'intuition' (Anschauung) for both kinds of
experience, perception of physical objects and essential insight.
Intuition, or Anschauung, is what gives us evidence. We have
negative evidence when the restraints force us to restructure our
experience. We have positive evidence when some of our anti-
cipations are fulfilled and yet no restraints are violated. We always
have many more anticipations than those that are fulfilled in a given
experience: the anticipations always go beyond our present exper-
ience and point to further experiences that may conform with the
anticipations and thereby provide more positive evidence, or go
against the anticipations and give us negative evidence. Evidence
hence comes in degrees. Husserl warns against the overtones in the
direction of perfect evidence that the word 'evident' has in German,
as in English.8
Whether we study nature or pursue mathematics or other eidetic
sciences, we are gathering evidence to eliminate wrong theories and
support better ones. We find out more about essences basically in the
same way as we find out more about physical objects: we follow up
our anticipations and check them by exploring the object further.
In mathematics and the eidetic sciences, the individual physical
objects that exemplify the eidos are of no interest to us except in so
far as they illustrate the eidos. We are merely interested in the
interconnections between the different eidos. To explore these inter-
connections we use the method of eidetic variation, a method that
is commonly used by mathematicians and that was thematized and
developed by Bernard Bolzano, who called it simply the method of
variation: We go through a number of examples, as varied as
possible, that instantiate the eidos, and we check which properties
and relations are preserved in all the examples. To use an example
that Husserl knew well: Before Bolzano mathematicians had gen-
erally taken for granted that continuous functions are differentiable,

8 Erfahrung und Urteil, §4, p. 12 = Churchill & Ameriks, p. 20. See also Ideas, §§137-138.
140 I—DAGHNN F0LLESDAL

if they had thought about the question at all; many generally shared

Downloaded from https://academic.oup.com/aristoteliansupp/article-abstract/68/1/131/1774421 by guest on 25 October 2019

assumptions are so deep-seated in us that they are hardly even
thematized. However, Bolzano constructed a simple example of a
continuous function that is nowhere differentiable. (Bolzano's result
remained unpublished and unknown, and thirty years later Husserl's
teacher Weierstrass gave another, slightly more complicated ex-
ample.)
Since the individual physical objects that instantiate an eidos are
of no interest to us when we pursue mathematics, it does not matter
whether there is any such object. The mathematician, unlike the
physicist, may hallucinate the object or, what is more common, he
may imagine the object. All that matters for him is what connections
there are between the different eidos, and this can usually be found
out more quickly and easily by imagining, rather than actually
materially constructing, a variety of objects that instantiate the
eidos and see what other eidos they instantiate.
A further reason for using thought experiments is that many
essences, for example triangularity and other geometrical essences,
are never precisely instantiated by physical things. Even if we were
to use concrete things to aid our intuition, we would therefore still
have to make use of idealization in order to arrive at the objects we
are studying in mathematics.

VI
Individuation and laws. We have noted that when we constitute the
world, we do not first constitute objects of a simple kind, for
example, sense data, and then step by step build up physical objects,
persons, etc., from there. I shall now argue that we likewise do not
start by individuating objects and events and then look for laws that
govern them and interconnect them, but that the objects and laws
emerge together, as one 'package'. 'Argue' may not be the proper
word here; I am not going to present an argument, but rather sketch
a view on individuation which does not seem implausible and which
throws light on some of the classical problems in philosophy, such
as that of induction.
The basic point is that there is an interplay between how we
individuate objects and how we conceive of the laws, or regularities
in the world. Individuation of objects of all kinds, physical objects,
 OBJECTS AND CONCEPTS 141

persons, abstract entities, events, etc., is intimately connected with

Downloaded from https://academic.oup.com/aristoteliansupp/article-abstract/68/1/131/1774421 by guest on 25 October 2019

our imposing a regularity structure upon the world. We have
already noted, with Quine and Husserl, how individuation, space
and time come as a package. This package contains much more. It
includes various regularity patterns: what we call natural laws
meshes with the way we individuate objects and events: we
individuate them in such a way that they fit tolerably in with one
another as causes and effects. As Husserl puts it in the case of
physical objects, including our own body: 'Thus if one takes away
causality, the body loses its ontic meaning as body, its identifiability
and distinguishability as a physical individual.'9 Similarly, the
regularity pattern called 'rationality' is established together with
individuation of actions, reasons, etc.10 And the structures studied
in mathematics match the schematism of space and time that we
impose upon the world.
On the linguistic level, as Quine has observed, there is also a
whole package of devices connected with individuation. It includes
the whole apparatus of singular terms, terms with divided reference,
identity, distinctness, plural endings, etc.
The way we differentiate properties and the way we group
several properties together as belonging to the same object are two
sides of our attempt to cope with the world, to create order out of
chaos and to anticipate what is to come, and to extrapolate from the
known to the unknown. The word 'tolerably' above means that we
do not succeed perfectly at this. The properties we discern in our
daily lives are dispositional, they are propensities. Each event and
object has numerous properties, and what is going to happen,
depends upon the interplay of all these properties. To be an
entrenched predicate is not to occur in exceptionless laws, but to
have predictive relevance. The properties which have emerged as
helpful in our daily lives, and which are denoted by the normal

9 Krisis, §62, Husserliana VI, 222.1-3 = Carr's translation, p. 218.

10 This is related to Davidson's point, in 'Mental Events', (1970), and many of the other
essays reprinted in Davidson, Essays on Actions and Events (Oxford: Clarendon Press,
1980), that 'there is no way of assigning beliefs to a person one by one on the basis of
his verbal behaviour, his choices, or other local signs no matter how plain and evident,
for we can make sense of particular beliefs only as they cohere with other beliefs, with
preferences, with intentions, hopes, fears, expectations and the rest'. ('Mental events',
p. 221)
142 I—DAGHNNF0LLESDAL

predicates in our language, are properties that tend to be associated 5

Downloaded from https://academic.oup.com/aristoteliansupp/article-abstract/68/1/131/1774421 by guest on 25 October 2019

with other properties and thereby enable us to extrapolate and ,
survive. The association works in two main ways: properties are
associated by belonging to the same object or event, or they are
associated as cause and effect.
In our daily lives and even more in science, there is a perpetual
mutual adjustment of objects, properties and laws. 'Science is not
exclusively empirical; it is also the application to nature, in
successive phases, of the principle of identity, the essence of our
understanding', observed Emile Meyerson in Identite et Realite}1
To illustrate his point Meyerson uses the principle of inertia, which
assimilates rectilinear and uniform motion to rest, and several other
principles of conservation. He also quotes Francis Bacon, who (sixty
years before Newton) wrote that 'Human intelligence... pretends to
find constant those things which are in flux.'12 Quine, in 'Natural
Kinds', explores these issues and observes how man has 'regrouped
things into new kinds which prove to lend themselves to many
inductions better than the old'.13 Quine also notes that in different
branches of science one might look for different similarity relations,
appropriate for the phenomena with which the particular science is
concerned, reminiscent of Felix Klein's Erlangerprogramm in
geometry.
The regularities, properties and structures cum objects that we
impose upon the world are not arbitrary. They are constrained by the
stimulations of our sensory surfaces. But they are, as we noted
earlier, not uniquely determined by these stimulations. The
stimulations I have had up to a given moment in time leave a lot of
possibilities open. As my life goes on, further stimulations eliminate
some of these possibilities. However, and here Husserl and Quine
would agree, no amount of sensory stimulation uniquely fixes the
structure we give the world. The world is experienced by us as
something that constrains us, something that lies there to be
explored. Normally, we are not even aware of ourselves as imposing

11 Emile Meyerson, Identiti et Reality, Paris, 1908. English translation by Kate

Loewenberg: Identity and Reality (London, Allen & Unwin, 1930), p. 402.
12 Francis Bacon, Novum Organon (1620), Book I, Aphorism 51. Quoted by Meyerson in
Identity and Reality, p. 286.
13 W.V. Quine, 'Natural Kinds', Ontological Relativity and Other Essays (New York:
Columbia University Press, 1969), p. 128.
 OBJECTS AND CONCEPTS 143

structure, we find the structure, do not make it. This holds for the

Downloaded from https://academic.oup.com/aristoteliansupp/article-abstract/68/1/131/1774421 by guest on 25 October 2019

abstract world of mathematics as well as for the concrete world of
physical objects.

VII

Singular terms. Reference. We will now pass to language and see

how the various features of objects are reflected in the semantics
of the special terms we use to refer to objects, the singular terms.
There are in particular three features of objects that are crucial for
reference.

First: They are the bearers of a (usually) large number of properties

and relations. Normally we know only a small number of these, but
the object is conceived of as having numerous further properties
that we do not know yet, but which are there to be explored. They
transcend our knowledge, to use Husserl's phrase.
Secondly: Objects, except mathematical ones and a few others,
change over time. One and the same object can have a property at
one time and lack it at another time. The object remains identical
through changes. Modalities come in at this point; not only are there
the actual changes, there are also possible ones, there are accidents
and there are necessities. Or, at least, so we say.
Finally: There is OUT fallibility. We may have false beliefs about
objects. We may seek to correct these beliefs, but all the while our
beliefs, true or false, are of the objects in question. A belief, or set
of beliefs, is not about whichever object happens best to satisfy
our beliefs. A semantics that just would seek to maximize our set
of true beliefs would reflect poorly the role that objects play in
epistemology.
Given that objects play an important role in our attempts to
explore and cope with the world, and given that objects have these
features, we should expect these features to be reflected in our
language. We should expect a language to have a category of
expressions that is especially designed to refer to these objects and
stay with them through all these changes that they and our beliefs
about them undergo. The expressions that do this, I called in my
144 I—DAGFINN F0LLESDAL

dissertation 'genuine singular terms'.14 When used in modal con-

Downloaded from https://academic.oup.com/aristoteliansupp/article-abstract/68/1/131/1774421 by guest on 25 October 2019

texts, these expressions are intended to relate to the same object in
all 'possible worlds' where that object occurs. In worlds where their
objects do not occur, the terms relate to nothing. Other terms—for
example, definite descriptions in most of their uses—are in many
ways much more like general terms that happen to be true of just
one object.

VIE

The tie between word and object. What I have called 'genuine
singular terms' Kripke calls 'rigid designators'. There has been a
difference in emphasis: I focused in my dissertation on the formal
arguments for such a category of terms, while Kripke has focused
on how to account for the tie between these terms and their objects,
and proposed his causal approach to reference. It is important to
keep these issues apart, for while I agree with Kripke on the need
for such a class of terms, I disagree with his causal approach.
The difference springs from a difference in our view on the role
that epistemology plays in semantics. Kripke has emphasized the
difference between the ontological issue of what a name as a matter
of fact refers to and the epistemological issue of how v/efind out
what it refers to. He has focused almost exclusively on the former
issue and has thereby been led to his causal view.
I tend to look upon the ontological and the epistemological issue
as much more closely intertwined. This is largely because language
is a social institution. What our names refer to—and not only how
we find out what they refer to—depends upon evidence that is
publicly available in situations where people learn and use
language. In various articles I have presented a view on how a term's
reference is determined through a complex interplay between many
different factors of this public kind.15

14 Referential Opacity and Modal Logic (Harvard 1961).

15 Particularly in 'Reference and sense', in Venant Cauchy, ed., Philosophy and Culture,
Proceedings of the XVIIth World Congress of Philosophy (Montreal: Editions du Beffroi,
Editions Montmorency, 1986), pp. 229—239. See also 'Meaning and Experience', in
Samuel Guttenplan, ed., Mind and Language (Oxford: Clarendon Press, 197S), pp.
25-44, and "The Status of Rationality Assumptions in Interpretation and in the
Explanation of Action', Dialectica 36 (1982), pp. 301-316.
 OBJECTS AND CONCEPTS 145

IX

Downloaded from https://academic.oup.com/aristoteliansupp/article-abstract/68/1/131/1774421 by guest on 25 October 2019

Rigidity as an ideal. It is in this interplay that we best can see what
the rigidity or genuineness, of singular terms amounts to. Rigidity
is not something that is achieved through the introduction of a
genuine singular term in our language. Stability of reference is
never guaranteed. There is always a risk that in spite of the best of
our efforts, a name comes to change its reference.
I look upon rigidity as an ideal, something like a Kantian regul-
ative idea, that prescribes the way we use language to speak about
the world. There is in our use of names and other genuine singular
terms a normative pull towards always doing our best to keep track
of the reference and keep on referring to it. Sometimes we go wrong
and it may then be unclear both what we believe and what our
beliefs are about until a new usage has been established.
All our talk about change, about causation, ethics and knowledge
and belief, as well as about the other modalities, presupposes that
we can keep our singular terms referring to the same objects. To
the extent that we fail, these notions become incoherent.
X
Quantified modal logic. All that is needed in order to interpret
quantified modal logic, is to acknowledge that there is at least one
category of singular terms—namely, the variables, or their natural-
language counterpart, the pronouns—that behave quite differently
from the general terms. As we have just noted, there are several
other kinds of expression, in addition to variables, that belong in
the former group—notably, most proper names.
There is a main dividing line in semantics between the genuine
singular terms on the one hand, and all othertermsand sentences on
the other. Genuine singular terms relate to their reference in a quite
different way from the way in which general terms and sentences
relate to their extension or truth value. Instead of assimilating these
expressions to one another and treating them all on a par with singular
terms (Frege) or general terms (Carnap), one comes to emphasize the
difference between reference and extension, and one becomes able
to distinguish two notions of opacity: referential opacity, where
substitutivity of co-referential singular terms breaks down, and
extensional opacity, where substitutivity of co-extensional general
146 I—DAGFINNF0LLESDAL

terms and sentences breaks down. For the exact definitions I refer to
my dissertation, pages 4-8, or to my article 'Quine on Modality',16

Downloaded from https://academic.oup.com/aristoteliansupp/article-abstract/68/1/131/1774421 by guest on 25 October 2019

pages 152-153. In these two places I also give proofs that only some
combinations of referential transparency or opacity and extensional
transparency or opacity are possible. What is important is that the
combination of referential transparency and extensional opacity,
which is required for quantified modal logic, is possible. This means
that quantified modal logic can be interpreted in a coherent way.
Quantified modal logic requires both referential transparency and
extensional opacity. The referential transparency part is required by
the quantifiers: whatever is true of an object is true of it regardless
of how it is referred to. The extensional opacity part reflects the
non-extensional character of the modal operators: we are not free to
substitute co-extensional general terms or sentences for one another
everywhere.
XI
De re propositional attitudes. This approach to the semantics of
singular terms makes it possible to deal with de re propositional
attitudes. I shall now argue that de re propositional attitudes are of
crucial importance for communication.
When I report, in my own words, somebody else's propositional
attitude, then I am faced with two problems. One is to put myself
in the other's place and describe the world as she or he sees it.
Another is to relate the other's perspective on the world to my own.
In the former case I am talking de dicto, in the second, de re. In the
former my singular and general terms take references and
extensions from the universe as the other sees it, and my quantifiers
range over that universe. The intelligibility of what I say depends
on the degree to which I succeed in getting clear about the other's
universe.
In the second case, where I am speaking de re, my problem is
compounded. I have to be able to see the world as the other sees it,
and at the same time relate the world, as seen from that perspective,
to the world as seen from my own perspective. I have to make sense

16 In: Donald Davidson and Jaakko Hintikka, eds., Words and Objections: Essays on the
Work ofW.V. Quine (Synthese Library, Dordrecht: Reidel, 1968), pp. 175-85.
 OBJECTS AND CONCEPTS 147

of sentences such as: "There is a person, in the world as seen by me,

Downloaded from https://academic.oup.com/aristoteliansupp/article-abstract/68/1/131/1774421 by guest on 25 October 2019

about whom Ralph believes that he is a spy.' The clause 'in the
world as seen by me' is normally regarded by the speaker as
redundant and so omitted. Hence, the problem of quantifying into
propositional attitude idioms from the outside requires that I master
two perspectives on the world, with their different individuations,
and that I am able to correlate at least some of the individuals in
one of these worlds with individuals in the other.
Given that when I am speaking derelam speaking of objects in
my universe, the universe I take to be the real one, then when I
quantify into such a context, the position(s) bound by the quantifier
must be referentially transparent. That is, any genuine singular term
in such a position must be subject to substitutivity of identity. Quine
had already seen this connection between quantification and
substitutivity in the early forties. Many have argued against it, but
in my opinion without success.
Quantification into propositional attitudes can err for two main
reasons, corresponding to the two tasks I just mentioned: I may fail
to see the world and individuate its individuals as the other does it,
or I may have correlated the other's individuals incorrectly with my
own. In either case quantification into propositional attitudes will
become murky or unintelligible. These are formidable difficulties.
However, they must to some extent be difficulties that can be over-
come: Quine has pointed out17 that some mastering of the idioms of
propositional attitudes is indispensable for learning a language and
for handing it down from generation to generation. I agree whole-
heartedly with this. For this task, however, de dicto propositional
attitudes will not suffice. It is essential for communication and for
the learning of language to master propositional attitudes de re.
Communication and translation are a matter of correlating not just
two world perspectives, but two perspectives on the same world.
It is therefore not enough to know what sentences the other
assents to. It is crucial that we are able to find out what objects are
referred to in these sentences, how the other's objects correlate with
our own. Learning a language therefore depends upon de re pro-
positional attitudes. Reciprocally, acquiring a language and using

17 In 'Promoting extensionality'. Forthcoming in Synthese, 1994.

148 I—DAGHNN F0LLESDAL

it to communicate consists to a large extent in becoming better able

Downloaded from https://academic.oup.com/aristoteliansupp/article-abstract/68/1/131/1774421 by guest on 25 October 2019

to perform the two key tasks that are required for mastering de re
propositional attitudes: getting into the other's perspective on the
world and its individuals, and correlating this with one's own.
Mastering de re propositional attitudes and mastering a language
are therefore inseparable. De re propositional attitudes are vital for
our ability to communicate.
Genuine singular terms, or rigid designators, enable one to make
sense of quantification into modal contexts. However, let us in
passing note that such terms do not name their objects on the score
of essential traits. Some proponents of quantified modal logic have
had such a view, but I do not. I have discussed this issue elsewhere18
and shall not pursue this here.
This does not mean that all is well with the modalities. There
remains a problem with the interpretation of the modal operators
themselves. As Quine has pointed out, some modalities, like those
of necessity and possibility, seem to depend upon murky meta-
physics. Others, such as knowledge and belief, obligation, per-
mission, and preference, and causality, dispositions, and probability,
are highly important notions that are main themes in philosophy. It
is fortunate, indeed, that quantification into such contexts can be
made sense of.

18 'Essentialism and reference'. In Lewis E. Hahn and Paul Arthur Schilpp, eds., The
Philosophy ofW.V. Quine. The Library of Living Philosophers (La Salle, 111.: Open Court,
1986), pp. 97-113.
 OBJECTS AND CONCEPTS

Downloaded from https://academic.oup.com/aristoteliansupp/article-abstract/68/1/131/1774421 by guest on 25 October 2019

Dagfinn F0llesdal and David Bell

II—David Bell

ABSTRACTION AND INTUITION

F rege wrote hardly any poetry at all. On at least one occasion,

however, he felt strongly enough about a certain topic to forsake
prose and bequeath us the following immemorable lines:
Abstraction's beneficial power
Works best when it is tamed and guarded.
This magic force turns quickly sour,
Its fetters loosed and reins discarded.1

In what follows I shall attempt to comment on some of the issues

raised in the opening sections of Professor F0llesdal's paper.2 In
particular I shall concentrate on some of the philosophical problems
concerning concept formation, abstraction, and intuition, as they
arise in connection with our grasp of abstract entities like numbers,
sets, concepts, 'essences' and the like; and here I shall follow
F0llesdal in focusing especially on Husseri's notion of intuition and
his doctrine of Wesenserschauung. First, however, I shall say some-
thing about abstraction and abstractionism—and here Frege's poem
is relevant; for it seems to me that in recent times a certain hostile

1 Wohltut des Abstrahierens Macht

Wenn es der Mensch bezahmt, bewacbt;
Doch furchtbar wird die Himmelskraft
Wenn sie der Fessel sich entrafft.

'Antwort auf die Ferienplauderei des Herrn Thomae', Jahresbericht der Deutschen
Mathematiker-Vereinigung, Vol.15,1906, p. 588. For a different English translation, see
G. Frege, Collected Papers on Mathematics, Logic, and Philosophy, edited by B.
McGuinness (Oxford: Basil Blackwell, 1984), p. 343.

2 D. F0llesdal, 'Individuation', Proceedings of the Aristotelian Society, Suppl. Vol. 68,

1994, pp. 131-40 above. Unfortunately only the opening sections of Professor F0llesdal's
paper were available to me for comment.
150 H—DAVID BELL

attitude to abstraction has been widely adopted in Frege's name,

Downloaded from https://academic.oup.com/aristoteliansupp/article-abstract/68/1/131/1774421 by guest on 25 October 2019

but that that attitude is in fact neither Frege's nor defensible.
Abstractionism and intuitionism have traditionally represented
opposing tendencies amongst theories whose aim is to account for
our cognitive access to abstracta of various kinds. To a first, very
rough approximation we might say that abstraction has typically
been emphasized by those attracted to a broadly constructivist
and/or empiricist approach: concepts, universals, abstract objects,
and the like are, on this view, essentially products or reflections of
our activities and procedures. Intuition, on the other hand, has
natural affinities with a more platonistic and/or rationalistic
orientation: although the existence and nature of abstract entities
are independent of our cognitive activities, on this view knowledge
of such entities is nevertheless attainable in so far as it is grounded
in a form of quasi-perceptual awareness of them, or acquaintance
with them.3 Frege has been credited with the demise of conceptual
abstractionism, and Husserl has been credited with the formulation
of a defensible version of conceptual intuitionism. Neither verdict,
I shall suggest, is sound.

I
As an account of how we come to form or acquire certain of our
concepts, abstractionism has enjoyed a long and honourable history.
It originated, perhaps, with Aristotle's use of the notion of
aphairesis, and has been subsequently endorsed, in one form or
another, by figures as diverse as Boethius, Aquinas, Gassendi,
Leibniz, Locke, Kant, Husserl, Russell, Dedekind, Cantor, and
Frege, amongst others. Now it is, of course, highly improbable that
there exists a single, determinate doctrine of abstraction held in
common by all the members of so varied a list of thinkers. And
indeed abstractionism has taken widely differing forms at different
times and in different hands. Nevertheless we can, I think, specify

3 For discussion of some of the issues alluded to here, in the context of debates concerning
the foundations of mathematics, see e.g., G. Boolos, 'Nominalist Platonism',
Philosophical Review, Vol.94, 1985, pp. 327-344; C. Chihara, Constructibility and
Mathematical Existence (Oxford: Oxford University Press, 1990); P. Kitcher, The Nature
ofMathematical Knowledge (Oxford: Oxford University Press, 1983); P. Maddy, Realism
in Mathematics (Oxford: Oxford University Press, 1990); C. Parsons, 'Mathematical
Intuition', Proceedings of the Aristotelian Society, Vol.80,1979-80, pp. 145-168.
 OBJECTS AND CONCEPTS 151

a number of rather general characteristics which together serve to

Downloaded from https://academic.oup.com/aristoteliansupp/article-abstract/68/1/131/1774421 by guest on 25 October 2019

justify calling these theories forms of 'abstractionism', and which
serve also to distinguish them from rival 'intuitionist' doctrines.
For a doctrine to count as a theory of abstraction in the sense I
require, four things are necessary: (i) the doctrine must aim to
provide an account of concept formation, for at least some of the
concepts we possess; (ii) it must assign a central role to the trans-
formation of contents of consciousness, or of items directly given
in experience; (iii) the process of transformation must take as its base
that which is (relatively or absolutely) concrete and/or particular,
yielding as a product that which is (relatively or absolutely) abstract
and/or general; and (iv) the process of transformation must involve
separating out, or selectively attending to, partial elements or
aspects comprising some complex phenomenon. The generality of
this characterization is necessary in order to encompass the wide
variety of forms that abstractionism has taken. Historically, that is,
the process of abstractive transformation has been assigned not only
inputs of different types (including sensations, impressions, per-
ceptions, representations, intentional objects, ideas, concepts, and
thoughts), but also a variety of kinds of output (including abstract
ideas, representative ideas, concepts, universals, meanings, abstract
objects, and thoughts). Moreover, the process of abstraction itself
has been taken to comprise a number of quite different procedures,
taking us from the particular to the universal, from the concrete to
the abstract, from that which is essentially nominal to that which is
predicative or adjectival, from the sensory to the intellectual, from
that which is ontologically self-subsistent to that which is neces-
sarily partial or aspectival, and so on.
Frege's views about how we generate and acquire concepts are
both complex and sophisticated. According to him there are at least
five distinct procedures by which we can attain concepts. I shall
call them (A) analysis, (B) synthesis, (C) Kantian abstraction, (D)
decomposition, and (E) transformation.
(A). Frege's context principle, according to which, famously,
'words only mean something in the context of a proposition,'4 has

4 G. Frege, Foundations of Arithmetic, trans. J. L. Austin, (Oxford: Basil Blackwell, 1974),

§62, p. 73.
152 H—DAVID BELL

a natural analogue for judgements and thoughts. As Frege put it: 'I
come by the parts of a thought by analysing the thought.'5 That is,

Downloaded from https://academic.oup.com/aristoteliansupp/article-abstract/68/1/131/1774421 by guest on 25 October 2019

'I start out from judgements and their contents, not from concepts,
[and] only allow the formation of concepts to proceed from judge-
ments.'6 In other words, 'instead of putting ajudgement together out
of an individual as subject and an already previously formed concept
as predicate, [I] do the opposite and arrive at the concept by splitting
up the content of a possible judgement.'7 According to Frege, for
example, the thought expressed by the sentence 'Sirius is bigger than
the sun' has a unique, determinate analysis: the thought comprises
the sense of the two singular terms 'Sirius' and 'the sun', along with
the sense of the incomplete function-name % is bigger than £'. Now
strictly speaking Fregean analysis is not a means of concept
formation: it is clear that Frege did not intend the process of analysis
to produce or create concepts. Analysis is, rather, a means of concept
retrieval. As Frege himself observed: 'if the expression of the
content of a possible judgement is to be analysable in [a certain] way,
it must already be itself articulated.'8 In other words, although the
analysis of thoughts and judgements into their component concepts
is a necessary step in the isolation and retrieval of those concepts,
such analysis can only reveal concepts that were in some sense there
all along.
(B). Conceptual synthesis is a process by which complex
concepts are formed from elements already isolated as a result of
analysis. If we are already in possession of a stock of concepts on
which we can draw, it is a simple matter to combine one or more
of them so as to form a new concept: the concepts fox and female,
for instance, can be combined to yield the concept vixen. Frege is
surelyrightto be dismissive of synthesis, and of the use made of it
by Boole and others, as a potential source either of genuinely new
concepts, or of logico-philosophical enlightenment. 'In this sort of
concept formation,' he wrote, 'one must assume as given a system
of concepts, or speaking metaphorically, a network of lines. These

5 G. Frege, Posthumous Writings, ed H. Hermes et al., transl. P. Long and R. White

(Oxford: Basil Blackwell, 1979), p. 253.
6 Ibid., p. 17.
7 Ibid.
8 Ibid.
 OBJECTS AND CONCEPTS 153

really already contain the new concepts: all one has to do is to use

Downloaded from https://academic.oup.com/aristoteliansupp/article-abstract/68/1/131/1774421 by guest on 25 October 2019

the lines that are already there to demarcate complete surface areas
in a new way. It is the fact that attention is principally given to this
sort of formation of new concepts from old... which is surely
responsible for the impression one easily gets in logic that for all
our to-ing and fro-ing we never really leave the same spot.'9 In
sharp contrast, Frege's other methods of concept formation are
intended to introduce concepts that are genuinely new.
(C). Abstractionism is most often taken to be, in Dummett's
words, 'the faulty theory of concept-formation due to the British
Empiricists,'10 according to which general concepts are acquired
by a process of singling out in attention some feature given in direct
experience, while ignoring the other features simultaneously
given.11 Although this theory is usually associated with the name
of John Locke, the association is in fact unfortunate; for the account
which Locke himself provides of abstraction is in many ways too
idiosyncratic, and indeed too problematic, for it to serve as a useful
paradigm or prototype. I shall therefore prefer to talk of 'Kantian'
rather than 'Lockean' abstraction. In the Jdsche Logic of 1800,
Kant says, concerning the logical origin of concepts:
The logical acts of the understanding, through which concepts are
generated as to their form, are:
1. comparison of representations among one another in relation
to the unity of consciousness
2. reflection as to how various representations can be conceived
in one consciousness; and finally
3. abstraction of everything else in which the given represent-
ations differ.
To make concepts out of representations one must thus be able to
compare, to reflect, and to abstract.... I see, for example, a spruce,

9 Ibid., p. 15.
10 M.A.E. Dummett, Frege: Philosophy of Language (London: Duckworth; 2nd edition,
1981), p. xlii. The description is, however, misleading in a number of ways:
abstractionism did not, for example, originate with the British empiricists, many of whom
(Berkeley, Hume, Mill etc.) in fact rejected it; and the theory has not yet, to my
knowledge, been proved faulty.
11 The characterization comes from Geach's influential chapter entitled 'Abstractionism',
in P. T. Geach, Mental Acts, (London: Routledge & Kegan Paul, 1957), p. 18.
154 H—DAVID BELL

a willow, and a linden. By first comparing these objects with one

Downloaded from https://academic.oup.com/aristoteliansupp/article-abstract/68/1/131/1774421 by guest on 25 October 2019

another I note that they are different from one another in regard to
the trunk, the branches, the leaves, etc.; but next I reflect on that
which they have in common among themselves,... and I abstract
from the quantity, the figure, etc., of these; thus I acquire a concept
of a tree.12
Anachronistically, perhaps, but for the sake of simplicity, I shall
take Kantian abstraction to comprise not merely the third element
identified here, but the whole tripartite process of which it is a part.
Frege's hostility to abstractionism of this kind is almost uni-
versally taken to have been implacable,13 and it is indeed true that
he reserved some of his most scathing criticism and abusive rhetoric
for the abstractionist views advanced by contemporary math-
ematicians. Nevertheless, far from denying what he called
'abstraction's beneficial power', throughout his writings Frege
asserted that many of our concepts are acquired as a result of
abstraction from sensory experience.14 Indeed this admission forms
a crucial part of the motivation for the strategy adopted in the
Grundlagen as a whole. The central question addressed in that work
is: 'How, then, are numbers to be given to us, if we can have no
presentation (Vorstellung) or intuition (Anschauung) of them?'15
Frege's attitude is clearly that if we could have sensory present-
ations, images, impressions, or intuitions of numbers, then our \
grasp of the concept of a number could be unproblematically
explained as a result of abstraction. It is precisely because this is i
impossible, given the lack of an intuitive base to which abstraction
could be applied, that Frege is forced to embark on the invest- '
igations that comprise the Grundlagen—the major aim of which is ;
to provide an account of the concept of number which, as he says, *
'owes nothing to intuition.' \
Dummett acknowledges that in the Grundlagen and other early 1
works Frege allowed that abstraction is a viable means of concept •!

12 I. Kant, 'The Jasche Logic', in Lectures on Logic, trans, and ed. M.J. Young (Cambridge: 1
Cambridge University Press, 1992), p. 592 1
13 For a dissenting opinion, however, see G.P. Baker and P.M.S. Hacker, Frege: Logical J
Excavations, (Oxford: Basil Blackwell, 1984), pp. 57-59. i
14 See, e.g., G. Frege, Foundations of Arithmetic, §§34, 44, 45, 48; Collected Papers, pp. -|
3
231, 254, 343; and Posthumous Writings, p. 71.
15 Foundations of Arithmetic, §62, p. 73. •'
 OBJECTS AND CONCEPTS 155

formation, but he claims that Frege came eventually to see the error

Downloaded from https://academic.oup.com/aristoteliansupp/article-abstract/68/1/131/1774421 by guest on 25 October 2019

of his ways, and that in his review of Husserl's Philosophy of
Arithmetic (1894) 'he rejected abstraction altogether.'16 In fact,
however, it is far from clear that Frege's review of Husserl is as
hostile to abstractionism as Dummett claims.17 Two things are,
however, clear. One is that Frege endorsed abstraction as late as
1906. In the same passage that contains his poem, for instance, he
wrote: 'We would do well to be very careful about abstracting. But
we should also not forget its beneficial effects!' The other is that
Frege's attacks on abstractionism are always directed at broadly the
same target, namely the claim that abstract objects can be created
by psychological processes. In this regard he is especially merciless
in his criticism of any view according to which logical objects like
numbers or sets can be brought into being and assigned objective
properties merely as a result of something like selective inattention.
But Kantian abstraction was never intended—indeed, it is mani-
festly incoherent—as a method of creating independently existing
abstract objects, platonistically construed. It is precisely the onto-
logical independence ascribed to pure abstract objects by platonic
realists like Frege that precludes the possibility that we could do
anything whatsoever to generate, destroy, or alter any such object.
Kantian abstraction is a means by which we form concepts, not a
procedure for creating abstract objects of the kind in which Frege
believed; and the objections he levelled against the latter procedure
simply leave the former intact.
This, of course, is just as it should be; for it is surely clear that
we can and do create concepts by Kantian abstraction. We do, in
other words, notice that a number of diverse phenomena have one
or more general characteristics in common—characteristics for

16 M.A.E. Dummett, Frege: Philosophy of Mathematics, (London: Duckworth, 1991), p.

85.

17 In his review of Husserl, Frege writes: 'Suppose, e.g., a black and a white cat are sitting
side by side before us. We do not attend to their colour, and they become colourless—but
they still sit side by side. We pay no attention to their posture: they are no longer sitting....
We no longer attend to the place, and they cease to occupy one—but they continue to be
separate. We have thus, perhaps, attained from each of them a general concept of a cat.'
(Collected Papers, pp. 197-198). Dummett's strongly anti-abstractionist reading
depends, however, upon his contentious claim that these words—and especially the last
sentence—are intended entirely ironically, that is, as a categorical denial that one could
acquire a concept in any such way.
156 D—DAVID BELL

which we have, as yet, no specific concept or term. We then coin a

Downloaded from https://academic.oup.com/aristoteliansupp/article-abstract/68/1/131/1774421 by guest on 25 October 2019

term or invent a concept that is intended to capture and express
precisely the common features we have noticed. Now, doubtless,
on many occasions the new concept originates from a straight-
forward process of synthesis: existing concepts are concatenated so
as to provide an explicit, stipulative definition of the new concept.
('I shall call a quark any particle whose charge is a multiple of one
third of the charge of the electron'). But this is not always the case.
On other occasions a new concept may be introduced to pick out
an otherwise diverse family of phenomena, on the basis of observed
similarities, in the absence of any explicit, conceptual definition.
Amongst the clearest examples of this are the concepts we intro-
duce to capture aspects of style or fashion in architecture, music,
film and the like. Impressionism, punk, Tudor, gothic, artdeco, and
film noir are amongst the concepts which I, at least, have acquired
by Kantian abstraction.
(D). Unlike analysis, which merely makes explicit the concepts
that are already present as the parts of a particular judgement,
decomposition™ requires the discernment of a common pattern or
structure, shared by a number of different thoughts. For present
purposes the most significant difference between analysis and
decomposition is this: grasping a component concept of the kind
revealed by the analysis of a given thought is a necessary condition
of grasping that thought as a whole, whereas grasp of a concept
resultingfromdecomposition of thoughts is not a component part
of those thoughts, and is not necessarily required for a grasp of those
thoughts. An example of Dummett's makes this clear. The thought
expressed by the sentence
(1) 13 is greater than 1, and for all n, if n divides 13, then either
n = 13, or n = 1,
has a unique analysis into its component conceptual parts. That
thought is a compound thought comprising the conjunction of two
independent thoughts,
(2) that 13 is greater than 1,

18 I use this term in the sense introduced by Dummett. See, e. g., The Interpretation ofFrege Is
Philosophy, pp. 263ff.
 OBJECTS AND CONCEPTS 157

and

Downloaded from https://academic.oup.com/aristoteliansupp/article-abstract/68/1/131/1774421 by guest on 25 October 2019

(3) that for all n, if n divides 13, then either n = 13, or n = 1.
Both these thoughts have, in turn, a determinate structure. Thought
(2), for instance, comprises the sense of the two singular terms '13'
and T , and the sense of the two place function-name % is greater
than £'. These are genuine components of the thought expressed by
(1), in that one could not have that thought if one lacked the concept
of 13, say, or of one thing's being greater than another. In contrast
to such analytically identified components, decomposition enables
us to discern a pattern or structure that may be possessed by a
number of different thoughts. One such pattern can be generated
by excising the term '13' wherever it occurs in sentence (1), leaving
the incomplete expression:
(4) ^ is greater than 1, and for all n, if n divides %, then either
n = £, or/i = 1.
Grasp of the thoughts expressed by sentences exemplifying this
pattern can then be used as the basis for the introduction of a new
concept by stipulating that (4) shall express the same sense as
(5) £ is prime.
Decomposition, in other words, can introduce genuinely new con-
cepts; for possession of the concept prime number is not a com-
ponent part of, and is in no way required by, grasp of the thought
expressed by sentence (1). In a case of this kind, according to Frege,
'we see that there's no question of using the boundary lines of
concepts we already have, to form the boundaries of new ones.
Rather, totally new boundary lines are drawn by such definitions
—and these are the scientifically fruitful ones.'19
Decomposition is both a valid and a valuable method of concept
formation. It is also, interestingly, a classically abstractive pro-
cedure in which a concept is generated from a class of concrete
phenomena of which we have direct awareness, via the isolation of
a common characteristic, as a result of selective attention to partial
elements or aspects present in the members of that class. In this case
the members of which we have direct awareness are complete

19 Posthumous Writings, p. 46.

158 D—DAVID BELL

thoughts, and the bringing to attention of a new pattern is secured

Downloaded from https://academic.oup.com/aristoteliansupp/article-abstract/68/1/131/1774421 by guest on 25 October 2019

by a transformation that requires the excision of certain component
elements. All the conditions specified earlier (p. 151) for a doctrine
to count as a theory of abstraction are fulfilled by Fregean decom-
position.
(E). Transformation is the term I have adopted for the procedure
introduced by Frege by which 'a judgement-content [involving an
equivalence relation] can be transformed into an identity.' Frege's
words are well known:
The judgement 'line a is parallel to line b\ or, using symbols,
a lib,
can be taken as an identity. If we do this, we obtain the concept of
direction, and say: 'the direction of line a is identical with the
direction of line b.' Thus we replace the symbol // by the more
generic symbol =, through removing what is specific in the content
of the former and dividing it between a and b. We carve up the
content in a way different from the original way, and this yields us
a new concept.... The concept of direction is only discovered at all
as a result of a process of intellectual activity which takes its start
from intuition.2"
As with Kantian abstraction, so here, we need to distinguish care-
fully between two claims that can be made on behalf of such a
procedure. On the one hand, that is, transformation can be construed
as a means of specifying or generating new concepts, whereas on
the other hand, in the guise of so-called 'definition by abstraction',
the process has been assigned the function of introducing or
individuating abstract objects.21 Frege himself insists, rightly, that
while the former use is both unobjectionable and valuable, trans-
formation by itself is nevertheless quite incapable of fulfilling the
latter function, that is, of providing determinate identity conditions
for abstract objects.

20 Foundations of Arithmetic, §64, pp. 74-75.

21 See, for example, G. Cantor, Contributions to the Founding of the Theory ofTransfinite
Numbers (La Salle: Open Court, 1915), p. 86; R. Dedekind, 'Was sind und was sollen
die Zahlen?', in Essays on the Theory of Numbers (La Salle: Open Court, 1901), p. 68;
G. Peano, 'Le definizione perastrazione', Mathesis societa italiana di mathematica, Vol.
7, 1915, passim; B. Russell, The Principles of Mathematics (London: George Allen &
Unwin, second edition, 1937), pp. 219-220.
 OBJECTS AND CONCEPTS 159

As a method for introducing new concepts, transformation

Downloaded from https://academic.oup.com/aristoteliansupp/article-abstract/68/1/131/1774421 by guest on 25 October 2019

functions as follows. In the simplest (first-level) case, we are given
as already fully understood, a thought expressed by a sentence of
the form 'aRb', where a and b are concrete objects, and R is a two
place, transitive, reflexive, symmetrical relation. The same thought
can then be expressed by an identity statement, the terms of which
are first-level functional expressions whose arguments are the
concrete objects a and b: '/(a) = f(b)'. The thought that
(6) a is parallel to b
can equally be expressed by saying
(7) the direction of a is identical with the direction of b.
Second-level transformation is directly analogous: an equivalence
relation between (first-level) concepts is transformed into an
identity the terms of which are second-level functional expressions
that take the first-level concepts as arguments. Thus the thought
expressed by the sentence "There are just as many Fs as Gs' can also
be expressed by the sentence 'The number of Fs is identical with the
number of Gs'. If we assume that the thought expressed by the first
sentence of each pair is fully understood, and that the second
sentence of each pair expresses the same thought as the first, then
clearly transformation is a coherent and useful method of concept
acquisition. Moreover, like Kantian abstraction and decomposition,
it too is a species of abstraction, in that it meets the four conditions
specified earlier (p. 151 above). Frege was therefore right to call it
'aprocess of intellectual activity which takes its start from intuition',
and which 'yields us a new concept'.
As far as concerns transformation as a putative means for the
introduction or individuation of abstract objects, two points should
be stressed.22 In the first place, if abstractive transformation is
viewed as a procedure for the identification of some particular

22 I shall be very brief. For more on this topic, see, e.g., D. Bell, 'Epistemology and Abstract
Objects', Proceedings of the Aristotelian Society, Suppl. Vol. 53, 1979, pp. 135-152;
M.A.E. Dummett, Frege: Philosophy of Language (London: Duckworth, 1973), ch. 14;
B. Hale, Abstract Objects (Oxford: Basil Blackwell, 1987), pp. 36ff.; P. Lorenzen,
'Equality and Abstraction', Ratio, Vol.4, 1962, pp. 85-92; P. Simons, 'What is
Abstraction & What is it Good for?', in Physicalism in Mathematics, ed. A.D. Irvine
(Kluwer, 1900), pp. 17-40; C. Wright, Frege's Conception of Numbers as Objects
(Aberdeen: Aberdeen University Press, 1983), pp. 107-117.
160 H—DAVID BELL

specified class of abstract objects, then it is open to precisely the

Downloaded from https://academic.oup.com/aristoteliansupp/article-abstract/68/1/131/1774421 by guest on 25 October 2019

objections formulated by Frege. The contextual definition provided
by the move from sentence (6) to sentence (7), for example, cannot
be taken to licence reference to, or quantification over, abstract
objects called directions. As Frege observed: 'If we were to try
saying q is a direction if it is introduced by means of the definition
set out above, then we should be treating the way in which the object
q is introduced as a property of q, which it is not.'23 The most that
transformation can warrant is talk of directions-presented-as-such,
and there are, of course, no such things.
The second point to note is this. So-called 'definitions by
abstraction' have been taken to provide a general technique by
appeal to which we can make sense of the claim that abstract objects
are possible objects of knowledge. Application of this technique
provides a form of explicit, non-contextual definition according to
which an abstract object is identified with an equivalence class. That
is, in response to the objection that the transformation of the thought
expressed by (6) into that expressed by (7) is capable only of
introducing non-entities called directions-presented-as-such, the
procedure is then supplemented by a further step: directions are
identified with equivalence classes of lines. A direction, accord-
ingly, is a class of lines determined by the relation of parallelism.
Explicit 'definitions by abstraction' of this kind are, of course,
technically unimpeachable. But they can throw no light on philo-
sophical problems concerning the nature of our grasp of abstract
objects in general; for they function by stipulatively identifying
abstract objects with classes. In the context of epistemological
worries about our cognitive access to abstract objects in general, the
stipulative element is objectionable, the identification of certain
abstract objects with other entities of the same kind is question-
begging, and, indeed, the entire procedure is a case of obscurium
per obscurius: classes are epistemically more problematic than the
objects (directions, tones, shapes, ages, colours, etc.) typically
defined by appeal to them; for class abstraction is a second-level
operation applied to further abstracta, namely functions.

23 Foundations of Arithmetic, §67, p. 78.

 OBJECTS AND CONCEPTS 161

At this point, on the basis of the foregoing survey of Frege's

Downloaded from https://academic.oup.com/aristoteliansupp/article-abstract/68/1/131/1774421 by guest on 25 October 2019

treatment of concept formation, we have grounds for concluding
provisionally that abstractionism can take a number of widely
differing forms; that each can make a contribution to our under-
standing of how we acquire and form our concepts; that Frege's
authority should not be cited as grounds for hostility to abstract-
ionism; but mat, as Frege himself stressed, in none of its forms can
abstraction be employed intelligibly as a warrant for the introduction
or individuation of abstract objects, platonistically construed.

II
In the early sections of his paper, one of Professor F0llesdal's
primary aims is to demonstrate that there is 'nothing mysterious
about Wesensschau, it is on a par with ordinary perception'
(pp.138-9). According toF0llesdal, Wesensschau is Husserl's term
for the intuitive awareness we have of universals:
[In certain acts] the universal itself is given to us; we do not think
of it merely in significative fashion, as when we merely understand
general names, but we apprehend it, behold it. Talk of an intuition,
and, more precisely, of perception... is in this case, therefore,
well-justified.24
Now according to Charles Parsons, 'the principle mark of this
[intuition-based] conception is an analogy between sense per-
ception as a cognitive relation to the physical world, and "something
like perception" giving a similar relation to mathematical objects,
and perhaps other abstract entities.'25 And at first sight Husserl's
conception certainly appears to conform to this pattern. Both in the
Logical Investigations (quoted above) and in Ideas Husserl seems
to claim that general concepts are given to us, direcdy and im-
mediately, in a way that is closely analogous to the way in which we
perceive individual objects. In Ideas, for example, he writes: 'The
essence (Eidos) is an object of a new type. Just as the datum of

24 E. Husserl, Logical Investigations, trans. J.N. Findlay (London: Routledge & Kegan Paul,
1970), Vol. II, Investigation VI, §52, p. 800. The italics are Husserl's.

25 C. Parsons, 'Mathematical Intuition', p. 145.1 have tacitly followed Parsons, not only in
distinguishing two kinds of intuition—namely (nominal) intuition o/abstract objects, as
against (prepositional) intuition that something is the case—but also in concentrating
more or less exclusively on intuitions of the former sort.
162 n—DAVID BELL

individual or empirical intuition is an individual object, so the datum

of essential intuition is a pure essence.'26 Indeed, Husserl goes

Downloaded from https://academic.oup.com/aristoteliansupp/article-abstract/68/1/131/1774421 by guest on 25 October 2019

beyond claiming to detect a mere analogical relationship. 'Here we
have no mere superficial analogy', he insists, 'but a radical com-
munity of nature. Essential intuition is, precisely, an intuition, just
as the eidetic object is an object.'21 These remarks, and others like
them, seem explicit enough. And yet, I shall suggest, appearances
are somewhat misleading here—for there is a sense in which Husserl
was an abstractionist at heart.
In the Logical Investigations (1900), Husserl provides a four-part
phenomenological analysis of the route that takes us from concrete,
sensory experiences to an intellectual grasp of a universal (or
Species) as such.
(i). At the 'lowest level of possible intuition', an act of sensory
perception directly presents an intentional object that is a real,
complex, concrete individual. The terminology here is Husserl's,
however, and is apt to mislead. The intentional object (a unicorn,
say, or a cat) is an individual object, not a universal or categorial one:
it is not the sort of thing of which it makes sense to say that it has
instances. It is real insofar as it is spatio-temporal (unlike 'irreal' or
'ideal' entities which have neither temporal nor spatial attributes).
It is concrete in the specific sense with which Husserl endows this
term: it is an ontologically self-subsistent whole—in contrast to
those phenomena that depend for their existence on participation in
some encompassing whole of which they are essentially partial
aspects or elements. And the intentional object is complex in that it
comprises a number of distinguishable parts, both dependent and
independent. If the intentional object is real (i.e. possesses spatial
and/or temporal attributes) then its component parts will likewise
be real, but they will not necessarily be concrete (i.e. self-
subsistent); indeed many of the individual elements distinguishable
within an intentional object will lack ontological independence.
(ii). If the intentional object presented directly in an act of sensory
perception is complex, then we can turn our attention to its com-
ponent parts and aspects. If I see a cat, for instance, then I can focus

26 E. Husserl, Ideas: General Introduction to Pure Phenomenology, trans. W.R Boyce

Gibson (London: George Allen & Unwin, 1931), § 3, p. 55.

27 Ibid. I have altered Husserl's italics.

 OBJECTS AND CONCEPTS 163

my attention on its tail, its head, its right front leg, its left ear, and so

Downloaded from https://academic.oup.com/aristoteliansupp/article-abstract/68/1/131/1774421 by guest on 25 October 2019

forth. In so doing, these items become new intentional objects of
new perceptual acts. And, as before, these intentional objects are
themselves real, complex, concrete individuals. I can, however,
focus my attention on items of a quite different sort: I can attend to
the cat's shape, size, colour, posture, or movement. When I look at
the cat's colour, say, then that particular colour becomes the
intentional object of a new perceptual act. This intentional object,
however, is not concrete, according to Husserl, but abstract: it is an
ontologically dependent aspect of a larger whole. The cat's colour
is, on the other hand, a real individual: it is real because it is
spatio-temporal, and it is individual because it is the particular,
unshareable colour possessed uniquely by a certain cat. Husserl calls
such dependent, partial, abstract aspects moments?*
We become aware of moments as a result of an abstractive
procedure—that is, by focusing attention on partial aspects of some
concrete whole, while ignoring other such aspects, in such a way
as to transform our awareness of a concrete object into an awareness
of an abstract aspect thereof.
(iii). 'In the theory of abstraction since Locke,' Husserl writes,
'the problem of abstraction in the sense of an emphatic pointing to
"abstract contents" has been mixed up with the problem of
abstraction in the sense of concept-formation.'29 Naturally Husserl
himself is careful to keep these tsyo kinds of abstraction clearly
distinct. Step (ii) takes us from intuitions of concrete individuals to
intuitions of abstract individuals; step (iii), on the other hand, takes
us from intuitions of abstract individuals to a grasp of a general
concept or Species:
Naturally I do not here mean 'abstraction' merely in the sense of a
setting-in-relief of some non-independent moment in a sensible
object, but Ideational Abstraction, where no such non-independent
moment, but its Idea, its Universal, is brought to consciousness.

28 Others have called them tropes, concrete universals, and abstract particulars. For more
on Husserl's formal ontology, see e.g., D. Bell, Husserl (London: Routledge, 1990), pp.
93-114.
29 Logical Investigations, Investigation II, §40, p. 426.
30 Ibid., Investigation VI, §52, p. 800.
164 n—DAVID BELL

'We directly apprehend the Specific unity redness', for instance,

Downloaded from https://academic.oup.com/aristoteliansupp/article-abstract/68/1/131/1774421 by guest on 25 October 2019

if we focus attention on the red moments present in a number of
objects, and if we then 'perform a peculiar act, whose intention is
directed to the "Idea" or "universal".'31 Unfortunately, Husserl tells
us virtually nothing about the precise nature of this 'peculiar act.'
But in so far as it clearly involves abstracting, say, the universal
redness from a number of chromatically similar red moments
presented in intuition, the process is most naturally taken to be a
form of transformation. From a noticed similarity,
(8) Moment a is chromatically similar to moment b,
we can move to
(9) The redness of a = the redness of b.
In this process we acquire the concept of a shareable, common
feature that can have indefinitely many moments of red as its
instances. This is perhaps what Husserl is attempting to say when,
for example, he writes that there are certain acts
which give form to acts of straightforward... intuition, and
transform them into new presentations of objects. These latter
presentations, as opposed to the acts on which they are founded, set
up for us a peculiarly modified objectivity: the original objects are
now seen in certain interpretative and connective forms which are
our categorial forms in the second, objective sense.32
Now one might well think that the story would end at just this
point, with the provision of an account of how we can acquire an
intellectual grasp of the sense of a general term or concept-word.
The most idiosyncratic feature of Husserl's treatment, however, is
his emphatic denial that step (iii) can provide us with a full,
authentic (eigentlich) knowledge of a universal. Step (iii) does
indeed yield a grasp of 'a semantic essence'—we acquire thereby
a knowledge of the sense of the word 'redness', for instance—but
for Husserl this falls far short of constituting genuine knowledge of
the Species redness itself. Acquisition of the latter requires a fourth
and final step, namely:

31 Ibid., Investigation II, §42, p. 432.

32 Ibid., Investigation VI, §61, p. 819.

 OBJECTS AND CONCEPTS 165

(iv). We must experience a Wesenserschauung, that is, an

Downloaded from https://academic.oup.com/aristoteliansupp/article-abstract/68/1/131/1774421 by guest on 25 October 2019

immediate intuition of the universal itself. We must 'behold' the
universal, which must be present to us 'in person'. It seems, in other
words, as though abstraction and transformation can at best provide
us with indirect, second-class knowledge of a universal—know-
ledge by description, we might say. Fully authentic knowledge, on
the other hand, would seem to require genuine acquaintance with
the thing itself.
There are a number of respects in which this doctrine is, to say
the least, problematic.
In the first place, Wesenserschauungen are intended by Husserl
to be experiences to which appeal is made within the context of an
exclusively phenomenological theory—a theory, that is, whose data
are restricted ultimately to items of which we have direct experience.
In this respect one problem with eidetic intuition is simply that there
is no such kind of experience. Or at least, I can report that in my own
case I have had phenomenologically accessible experiences of red
objects (step (i)); I have had such experiences of the individual red
moments present in objects (step (ii)); I have, I think, a good grasp
of the concept red (step (iii)); but I have never beheld the pure eidetic
object redness itself.33
A second puzzling feature of Husserl's doctrine of Wesens-
erschauung is its lack of theoretical motivation: there is simply no
philosophical problem to which it can claim to be the solution. In
this respect, intuition would seem to be, as Wittgenstein claimed,
'an unnecessary shuffle.'34 In particular, if our concern is with the
nature of concepts—how we form, acquire, express, communicate,
and apply them, for example—then appeal to quasi-perceptual
intuitions of abstract entities is surely quite redundant. Though I
cannot defend the claim here, I would suggest that no valid ex-
planatory function remains to be performed by eidetic intuition,
once we have acknowledged the procedures introduced earlier

33 I am not, it seems, alone in this. See, e.g., the analogous objections to Godelian
intuitionism, in P. Kitcher, The Nature of Mathematical Knowledge, pp. 60-61; C.
Parsons, 'Mathematical Intuition', p. 148; P. Maddy, Realism in Mathematics, pp. 71,
75-80.

34 L. Wittgenstein, Philosophical Investigations, trans. G.E.M. Anscombe (Oxford: Basil

Blackwell, 1958), §213.
166 H—DAVID BELL

under the names of analysis, synthesis, abstraction, decomposition,

Downloaded from https://academic.oup.com/aristoteliansupp/article-abstract/68/1/131/1774421 by guest on 25 October 2019

and transformation.
In addition to being phenomenologically suspect and philosoph-
ically redundant, Husserl's theory of eidetic intuition is object-
ionable, thirdly, on the grounds that it is intrinsically misconceived.
The end product, yielded as a result of step (iv), Husserl emphasizes,
is a direct acquaintance with an object (a universal, Species, Idea,
essence, eidos, or concept): 'The eidetic object is indeed an object.
Essential intuition is the consciousness of something, of an 'object',
a thing towards which its glance is directed.'35 But in so far as
Husserl's aim is to provide an account of how we form and grasp
general concepts, this model is entirely inappropriate: conceptual
understanding does not consist in, and is not brought about by,
simply looking at an object—and here it matters not at all how
idiosyncratic we make the looking, or the object, or both. To possess
a concept is, at the very least, to have mastery of various skills and
techniques which mere acquaintance with an object is powerless to
explain.36
These, then, are some of the charges which, prima facie at least,
advocates of a Husserlian approach would need to address before
concluding, with F0llesdal, that there is 'nothing mysterious' about
direct intuition as a valid mode of access to that which is general or
abstract in our experience. Indeed, I have tried to suggest to the
contrary that the modes of access available to us are in fact always
indirect, and comprise a number of procedures which, their
differences notwithstanding, fully deserve to be called forms of
abstraction?7

35 Ideas, §3, p. 55.

36 Another plausible line of objection, to the effect that Husserl's theory of eidetic intuition
is misconceived, would focus on the nature of the objects intuited, rather than, as here,
on the nature of the cognitive relation it is claimed that we have to such objects. The
coherence of the very notion of a general object has been criticised, e.g., in M.A.E.
Dummett, Frege: Philosophy of Language, pp. 173-179,257-260.
37 My thanks go to colleagues in Sheffield, and especially to George Botterill, Peter
Carruthers, Stephen Makin, David Owens, Tom Pink, Peter Smith, and Bob Stern, for
helpful discussions and comments. I am grateful, also, to the British Academy for the
Research Readership during the tenure of which mis paper was written.