ESSAYS ON REFERENCE, LANGUAGE, AND MIND

This page intentionally left blank

ESSAYS ON REFERENCE,
LANGUAGE, AND MIND
Keith Donnellan

Edited by Joseph Almog

AND

Paolo Leonardi

3
Oxford University Press, Inc., publishes works that further
Oxford Universitys objective of excellence
in research, scholarship, and education.
Oxford New York
Auckland Cape Town Dar es Salaam Hong Kong Karachi
Kuala Lumpur Madrid Melbourne Mexico City Nairobi
New Delhi Shanghai Taipei Toronto
With offices in
Argentina Austria Brazil Chile Czech Republic France Greece
Guatemala Hungary Italy Japan Poland Portugal Singapore
South Korea Switzerland Thailand Turkey Ukraine Vietnam

Copyright 2012 Oxford University Press

Published by Oxford University Press, Inc.
198 Madison Avenue, New York, New York 10016
www.oup.com
Oxford is a registered trademark of Oxford University Press
All rights reserved. No part of this publication may be reproduced,
stored in a retrieval system, or transmitted, in any form or by any means,
electronic, mechanical, photocopying, recording, or otherwise,
without the prior permission of Oxford University Press.
Library of Congress Cataloging-in-Publication Data
Donnellan, Keith.
Essays on reference, language, and mind / Keith Donnellan ; edited by
Joseph Almog and Paolo Leonardi.
p. cm.
ISBN 978-0-19-985799-9 (alk. paper)
1. Philosophy, American20th century. 2. Philosophy,
American21st century. 3. Reference (Philosophy) 4. Language
and languagesPhilosophy. 5. Philosophy of mind. I. Almog, Joseph.
II. Leonardi, Paolo. III. Title.
B945.D6251A46 2012
121'.68dc23
2012003947
1 3 5 7 9 8 6 4 2
Printed in the United States of America
on acid-free paper

CONTENTS

Editors Preface: Keith DonnellanPhilosophy by Ear

Introduction by Keith Donnellan

vii
xiii

1. Reference and Definite Descriptions

2. Putting Humpty Dumpty Together Again

31

3. Proper Names and Identifying Descriptions

49

4. Speaking of Nothing

81

5. Speaker Reference, Descriptions, and Anaphora

115

6. The Contingent A Priori and Rigid Designators

147

7. Kripke and Putnam on Natural Kind Terms

179

Subject Index
Author Index

205
209

This page intentionally left blank

EDITORS PREFACE:
Keith DonnellanPhilosophy by Ear

A few months ago, after fifteen years of retirement in northern California, surrounded by fine horses, quality wines, and English detective novels, Keith Donnellan showed up one Thursday afternoon in
the old room of the philosophy of language workshop, which he did
so much to launch in the late eighties.1 He listened patiently to a
series of celebratory speeches. But if you tracked his body language,
he seemed like an athlete in the starting blocks waiting for the sprint
to start. When it did, four cutting-edge papers, all commenting on
his work, were given by quick-witted young philosophers. A heated
discussion ensued each time, led by the tribes leaders, David Kaplan
and Tyler Burge. Four papers led to four hours of high-speed sparring. And four times over, this unassuming and shy man made a oneparagraph contribution. Fifteen years out of practice, literally out of
the woodwork, his comments were always the simplest, the sharpest, and, of course, the most intuitive of all. And though he did not
use like the rest of us the high-tech gadgets of rigid designators,
scope distinctions, and singular propositions, he spellbound
everybodyold hands who knew him well, new grads who never

vii

E S S AY S O N R E F E R E N C E

saw him before. A sort of Gandalf the white, a sage from a different
era, when direct and intuitive philosophy was still possible, indeed
a form of daily life.
My own first realization that Keith Donnellan was endowed
with special powers was twenty-four years before that Thursday,
when he managed to make me enjoy eating a fish. By then, in my late
twenties, I had reneged on all my principles of youth. But not this
one last gut feelingI hated fish. It is my belief that Keith knew Id
never ever eat fish. One day, he invited me to dinner at his place. He
was the cook. I was so involved in the conversation that I did not
attend at first to what I was eating. Slowly, it grew on me that I was
munching on something extraordinarily tasty and that this would
be my most memorable culinary experience since childhood days in
France. I went on and on celebrating the ingenuity of the food and
the cook, not knowing what I was talking abouta fishbut referring to it anyway, a vindication of the theoretical views of the cook.
As we were about to go home, at the doorway, Keith took me aside
and whispered in my ear, By the way, the thing you kept drooling
over the whole evening was a special gig of mine . . . stuffed fish.
There he left the matter, with his little shy and subtle smile. Such was
Keiths sotto voce style in human interactions, and such was his way
in philosophy.
Donnellan did see things directly. It was no accident that of all
the innovative minds who came to us with game-changing ideas
in the late sixties and early seventies, it was Donnellan who came
up with the fundamental idea of a referential use of a piece of
language.
This transformative idea emanates directly from perceptual cases.
You and I are standing in the alcohol-free party. The man in front of
us, Sir Alfred, is holding a martini glass, drinking from it plain water
and allowing his tongue to roll. Unbeknownst to us there is only one
viii

E D I T O R S P R E FA C E

martini drinker in the party, Bertrand, hidden in the kitchen, where

he abuses discreetly of the forbidden juice. Deprived of a rude use of
a pointing finger, I say to you in a whisper, The man drinking a martini is tipsy. Did I say something true? Did I say something of Sir
Alfred? Did I refer to Sir Alfred? Thus was engendered the idea of a
referential use.
The idea came to Donnellan in 1964. But it was published in
Philosophical Review, in a now classic paper called Reference and
Definite Descriptions, only in 1966. The paper, submitted to PR,
was initially . . . rejected by an external referee. (Keith was at Cornell
at the time, so an outside hand was called upon.) It was only the intervention of an internal knowing hand that saved the paper from
the absurdities of professional refereeing. And so with a little help
from discerning friends, the paper was out in 1966.
Like the negative referee, not everybody was immediately appreciative. These were heady times in the philosophy of language
and mind, and the air was filled with ideas that struck a much louder
chord.
In the late sixties and all the way to the critical year 1970, the
philosophy of language and the philosophy of mind went through
related paradigm shifts. The then-reigning paradigm, Freges theory
of language and thought, became the focus of a fundamental critique. A new theory, direct reference theory, emerged from the critique. It went on to become the new paradigm, and it is the one
informing our perspective today.
The current volume of Donnellans essays offers a unique twotiered insight into this 19661970 turn. First, we get a new perspective on the distinct contribution of one of the main game changers,
Keith Donnellan. Second, and more generally, we are offered a
newsomewhat surprisingpicture of what the direct reference
turn was all about.
ix

E S S AY S O N R E F E R E N C E

Keith Donnellans work in the essays here focuses mainly on

(i) the relation of semantic reference (and keep in mind the modifying adjective I just used to qualify reference), a touchstone notion
in the philosophy of language, and (ii) the relation of thinking about,
of primal concern in the philosophy of mind. I believe that it is fair
to say that, until recently, it was commonplace to view Donnellans
work on these two major fronts as significant but, in the end, playing
a merely supportive role in the path-breaking changes of 1966
1970. Some held the view of the reluctant referee. But even those
who saw the light thought referential uses are a minor sideshow
compared with the central stage ideas of direct reference and/or
rigid designation.
The current book of classical essays by Keith and a companion
book of essays about Keith may help us form a deeper sense of the
various (nonequivalent) innovations offered in 1966 through 1970
and, in particular, to appreciate the very distinct set of foundational
ideas proposed by Keith.2
So much concerns theoretical matters, the theory of thinking
about things and semantically referring to them. But over and above
the theoretical issues, there remains the human imprint of this perceptive and innovative thinker. The landscape of the philosophy of
logic and language is filled with glamorously brilliant thinkers whose
writing style reflects their scintillating minds. One only has to think
of Russell, Quine, and Keiths own peers in the 19661970 direct
reference turn, Kripke, Kaplan, and Putnam. As a person and
teacher, Keith shone in his own distinct subtle way. Those who knew
him by acquaintance will realize the mans style is reflected in his
crystalline papers. Those who will be having him in mind only by
way of his papers will come, upon a close rereading the papers, to
share this image of a naturally born philosopher.

E D I T O R S P R E FA C E

NOTES
1. The publication of the present volume of essays by Keith Donnellan is accompanied by the publication of a companion volume of essays about Donnellans work
(With Donnellan in Mind, OUP, 2012). The companion volume resulted from
two recent conferences about his work, in Bologna in March 2008 and at UCLA
in May 2008. The present introduction is focused solely on Donnellans own
essays. I owe thanks to comments of T. Burge, B. Herman, S. Shiffrin, J. Carriero,
D. Kaplan, T. Martin, S. Cumming, H. Wettstein, S. Coolidge, and my coeditor,
P. Leonardi. Earlier on, in Bologna, I benefited from remarks of H. Kamp and
J. Perry (for more personal remarks about Keiths philosophical style, see Perrys
contribution in the companion volume). Special thanks regarding our understanding of referential uses are due to D. Kaplan. My debt to Keith Donnellan,
philosophical and personal, is unique. He is the man who made me understand
what is the source of something I lived bydirect reference. I am also very grateful to his comments on the present introduction.
2. See fn. 1.

xi

This page intentionally left blank

INTRODUCTION
by Keith Donnellan

Collected here are several of my papers on topics in the theory of

reference. A theory of reference, as it is to be found in my work,
would explain how something becomes the referent of an expression in an utterance by a speaker on a particular occasion or becomes
the thing a speaker refers to by using an expression on an occasion.
The thing here may be a person, an object, a place, or one of the
many diverse sorts of things we talk abouta play, a political movement, a theory of structure of the atom, to mention only a few. Some
of these may in the end require special treatment in their own right.
Several of these papers can be seen as attempting to accomplish
two aims. First, I try to refute a standard reply to the question, What
determines the referent, if there is one, of definite descriptions and
proper names? I will call it the identifying descriptions view, following my use in the paper Proper Names and Identifying Descriptions. * Second, an answer to this same question is sketched out or

* I believe that the paper Proper Names and Identifying Descriptions (1970) is the earliest publication of the argument it presents for the conclusion that the referents of proper
names are not established through associated descriptions.

xiii

E S S AY S O N R E F E R E N C E

used in several papers. Simply put, I argue that in a wide range of

cases of the use of these two important kinds of singular terms, it is
not a set of descriptions that determines the referent, but rather the
intention of the speaker in using the term. The identifying descriptions view, in contrast, would hold that if definite descriptions, for
example, have referents, then the descriptive content of the description will provide the referent; it will be that which satisfied the
description uniquely. To obtain uniqueness, the uttered description
will perhaps require augmentation by certain unexpressed descriptions from the content of utterance (for example, if the overt
description is the man drinking a martini, it may need to be augmented by the unexpressed description in this room). The most
common version of the identifying descriptions view in regard to
proper names appeals to the descriptions that the speaker would or
could give as identifying the referent (for example, having used the
name Thales, the speaker might describe the referent as the
ancient Greek philosopher who held that all is water and who fell in
a well while contemplating the stars).
As against the identifying descriptions view in each of the mentioned cases, the argument I give in the two main papers on the subject (Reference and Definite Descriptions and Proper Names
and Identifying Descriptions) work by way of a variety of constructed examples in which there are different sorts of failures of the
descriptions to describe the actual referent. Either the descriptions
pick out something that is surely not being referred to, fail to pick
out the referent uniquely, or are insufficient to pick out anything at
all. And thus the descriptions are not what yield the solution to the
problem of how the referent is determined. So, having, as I thought,
countered this plausible but wrong view, I went on to adumbrate
what I took to be the most promising alternative. Looking at the
paper on reference and definite descriptions, I must confess that
xiv

INTRODUCTION

only one of the two writers mentioned there as my targets, Bertrand

Russell and Peter Strawson, could fairly be said to have held that
definite descriptions are used to refer and that the entity referred
to is identified in virtue of its having the properties mentioned in
the description (plus properties assumed from the context of
utterance). Of the two it is Peter Strawson who, in his (at the time of
my papers publication) widely read paper, On Referring, espouses
this view. He says, for example:
In the case of phrases of the form the so-and-so used referringly,
use of the together with the position of the phrase in the sentence . . . acts as a signal that a unique reference is being made and
the following noun, or noun and adjective, together with the context of utterance, shows what unique reference is being made.1

Strawson, then, is an example of one who holds an identifying descriptions view about definite descriptions. This, however, is not true
of the second target of the paper, Bertrand Russell. He needed nothing to determine the referent of any use of a definite description, for
definite descriptions on his view do not refer at all, and hence there
is nothing to identify. Only what he calls genuine or logically
proper names can perform reference. And the rarefied requirements
for using these is such that they cannot even occur in ordinary discourse. And to utterances in ordinary discourse containing definite
descriptions Russell would apply, of course, his well-known theory,
which shows them to be composed of a conjunction of quantified
sentences, the definite description having disappeared. As against
Russell, I wanted to show that while there are many uses of definite
descriptions, what I called attributive uses, to which I was (and am)
happy to apply Russells analysis, there is also an equally large, important, and quite normal use of them in which they do refer. These,
xv

E S S AY S O N R E F E R E N C E

I called, (what else?) referential uses. The constructed examples

mentioned were meant to be of this latter sort.
The papers Reference and Definite Descriptions and Proper
Names and Identifying Descriptions were chronologically my first
papers on the subject of reference (1964 and 1970) and also express
the theoretical basis for my further forays into the topic. I would like
here to add a few comments to these two base papers.
I should have supplemented the paper on definite descriptions,
I now believe, by pointing to some features of the situations involving referential definite descriptions that give reason to think that the
sentence uttered cannot be analyzed via Russells theory of descriptions, features other than the fact that the speaker seems to be
making a reference in uttering the sentence. This would have blocked
any suggestion that while the sentence uttered may be analyzed as
Russell proposes, some other feature of the situation gives rise to
our intuition that reference is going on.2 Here I will point to two
such characteristics that seem to argue against treating the sentence
according to Russells theory.
The first characteristic of the situations is that very many of them
involve what I have called indefinite definite descriptions. Howard
Wettstein3 has shown, I believe, that this fact raises a problem for the
treatment of the uttered sentences according to Russells theory, a
problem that cannot be removed by the usual maneuver of augmenting the description by features from the context. It is, he also
maintains, no problem for the position of my paper that referential
definite descriptions are on a par with names and demonstratives as
making direct reference rather than introducing a quantified form.
Let me illustrate what I take to be Wettsteins point with an example
or two.
For simplicity, I will restrict myself to simple sentences of the
form The is , and I will hope it is agreed that if an uttered senxvi

INTRODUCTION

tence on an occasion is correctly analyzable, say, as in Russells

theory of descriptions, then the speaker on that occasion can be
held to have asserted the proposition such an analysis implies. At a
largish college party, suppose a speaker has been asked by a visitor
whether the dean of the college is present, for the visitor has a need
to meet him. The speaker, looking around, sees the dean in the
crowd and now is faced with how to bring about a reference to the
person he thinks to be the dean. In many circumstances, at least
three ways may present themselves. If he knows the name of that
person, George Smith, perhaps, and has reason to think the visitor
also knows the person by that name, he may choose to use it. If the
person he picks out as being the dean is not too far distant, he may
point or use some other gesture and a demonstrative: That is the
dean of the college. Or he may describe the person, choosing a description that, if possible, will allow the visitor to correctly pick out
the person he intends. And here, or so my view has it, he will very
frequently embed the description he comes up with in the linguistic
form we call a definite description. Now, a question arises here:
Must the speaker aim at so describing his intended referent as to
apply to it uniquely? Well, if we look at practice, it will seem hardly
so: Where did I leave my purse? I think the purse is in the car.
There are I dont know how many cars in the world. Yet this
might be a satisfactory answer to someones question. At a large faculty celebratory gathering, a visitor to the college asks, Who is the
dean? I need to introduce myself to him. The speaker, glancing
around, replies, The man drinking a martini is the dean. Is the description to be delimited so as to secure uniqueness by something in
the context that the speaker relies on to be understood by the questioner? Well, what is it, and did the speaker think of it or assume it?
Most often, cases of what I would see as referential definite descriptions are quite indefinite in this matter and, if simply accepted as
xvii

E S S AY S O N R E F E R E N C E

such, would give a reason to say that Russells analysis does not fit
here, for on his analysis there is a uniqueness condition. And in
order to save his view, many have suggested that augmenting descriptions are assumed by the speaker to be supplied by context. But
usually there are, it seems, many possibilities given that could be
seen as given by the context. In the last example, shall we say that the
augmenting description is in this room? But then, what about that
man sipping a martini across the room, who isnt the dean or thought
to be by the speaker? It is not unlikely that the speaker either didnt
notice him or simply ignores him. Perhaps the augmenting description might be in front of us. But how wide a sweep and how far?
Can we say that at the time of utterance the context specifies one
possibility rather than all the others? If we cannot, there is not one
Russellian analysis to assign to the utterance but rather several.
Suppose we ask, if Russells analysis will not work in these cases,
why do we use the form we call a definite description. My suggestion is that it might seem natural because there is one and only one
thing in these cases of referential definite descriptions that the
speaker intends to refer to; it signals that the description is supposed
to refer to one definite thing, even when it is not the only thing satisfying the description in the context. Lets go back to my somewhat
artificial elucidation of the problem facing the speaker faced with
several possible ways of getting his reference across to the audience,
who chooses to describe the object of the intended reference. A partial description, which may be all he has resources to give, is very
likely to do the job of getting the listener in a position to find the
referent. The speaker simply relies on the listener to exercise
common sense: The drinks are on the table in the kitchen. The
guest goes into the kitchen. There are, in fact, three tables in the
kitchen. One is too small to contain drinks; another, large enough,
is covered with the makings of the dinner they will enjoy later; and,
xviii

INTRODUCTION

lo and behold, there is a table, not a salient table, but it holds an

assortment of snacks and drinks. He has no problems with which
table the speaker was referring to. The use of description table,
however, was useful, perhaps, in preventing him from looking in
the refrigerator or searching cupboards. So I think Wettstein was
correct: Indefinite definite descriptions provide reason to doubt the
application of Russells analysis to referential uses, and my treatment
of the cases yields a somewhat satisfactory solution to the problem
such descriptions may seem to present.
Were what I would call instances of the referential use of definite
descriptions to be analyzed in accord with Russells theory, the speaker
could be taken as asserting the proposition the theory yields. And the
truth conditions of that proposition will depend on the properties of
the denotation of the description, if there is one. The denotation
should then be treated by speaker and listener as very important. But
in fact it seems often that if the description is what I count as referential, the denotation is treated as irrelevant. So if the speaker says, The
man drinking a martini is the dean in our example, and it turns out
that the man he intended to refer to has only water in a martini glass,
the truth or falsity of the utterance is still a matter of whether that man
is the dean. Finding another man who is drinking a martini, but not
the intended reference, will not make the utterance true if he turns out
to be the dean nor make it false if he is not, even if that man is the sole
drinker of a martini. Thus another feature we would expect to hold if
a Russellian proposition is involved fails to be present.
The main thing I wish to say about the third of the essays reprinted
here, Proper Names and Identifying Descriptions, is that it was for
me a natural extension of the views on definite descriptions. Here the
descriptions involved are those supposedly internal and available to a
user of a proper name. On the view I attack, it is these that determine
or identify the referent. This explanation of how the referent of a proper
xix

E S S AY S O N R E F E R E N C E

name is determined was, I think, both widespread and not at all unintuitive. The structure of the argument in this paper paralleled that in
the previous paper on referential definite descriptions: By several constructed examples, I hoped to show that the referent of a proper name
used on a particular occasion could be identified, even though the descriptions available to the speaker were mainly false of that referent,
not uniquely identifying or identifying of the wrong thing. The way in
which the referent comes to be known, in each of these examples in the
paper, was by finding reason or evidence of the intentions of the speaker
in using the name; to what did he or she intend to refer? And the reason
or evidence is not restricted to the descriptions of the referent possibly
given by the speaker, although such descriptions, if available, may give
reason or evidence for concluding that this or that is the referent. The
positive thesis of the paper was the same as the previous one also: The
speakers intentions provide the ultimate answer.
I would like to express my gratitude to the editors, Joseph Almog
and Paolo Leonardi, for suggesting this collection, for selecting the
papers, and for seeing to its publication. Without their interest and
labor, it would not exist to be denoted or referred to.

NOTES
1. On Referring, in Anthony Flew, ed., Essays in Conceptual Analysis, London,
Macmillan, 1956, pp. 2152. The quote is from p. 45.
2. See for such a conclusion: Saul Kripke, Speakers Reference and Semantic Reference, Midwest Studies in Philosophy 2, pp. 255276. This is not the place to investigate Kripkes somewhat complex apparatus and argument. Of course, if
sufficient reasons for denying that a speakers sentence employing, in my sense, a
referential definite description is to be analyzed via Russells theory, that would
be to refute Kripkes conclusion.
3. Demonstrative Reference and Definite Descriptions, Philosophical Studies, 40,
pp. 241257.

xx

ESSAYS ON REFERENCE

This page intentionally left blank

[1]
REFERENCE AND DEFINITE
DESCRIPTIONS

I
Definite descriptions, I shall argue, have two possible functions.1
They are used to refer to what a speaker wishes to talk about, but
they are also used quite differently. Moreover, a definite description
occurring in one and the same sentence may, on different occasions
of its use, function in either way. The failure to deal with this duality
of function obscures the genuine referring use of definite descriptions. The best-known theories of definite descriptions, those of
Russell and Strawson, I shall suggest, are both guilty of this. Before
discussing this distinction in use, I will mention some features of
these theories to which it is especially relevant.
On Russells view a definite description may denote an entity:
if C is a denoting phrase [as definite descriptions are by definition], it may happen that there is one entity x (there cannot be more
than one) for which the proposition x is identical with C is
true. . . . We may then say that the entity x is the denotation of the
First published in The Philosophical Review, 1966, 75: 281304)

E S S AY S O N R E F E R E N C E

phrase C. 2 In using a definite description, then, a speaker may use

an expression which denotes some entity, but this is the only
relationship between that entity and the use of the definite description recognized by Russell. I shall argue, however, that there are two
uses of definite descriptions. The definition of denotation given by
Russell is applicable to both, but in one of these the definite description serves to do something more. I shall say that in this use the
speaker uses the definite description to refer to something, and call
this use the referential use of a definite description. Thus, if I am
right, referring is not the same as denoting and the referential use of
definite descriptions is not recognized on Russells view.
Furthermore, on Russells view the type of expression that comes
closest to performing the function of the referential use of definite
descriptions turns out, as one might suspect, to be a proper name
(in the narrow logical sense). Many of the things said about proper
names by Russell can, I think, be said about the referential use of
definite descriptions without straining senses unduly. Thus the gulf
Russell thought he saw between names and definite descriptions is
narrower than he thought.
Strawson, on the other hand, certainly does recognize a referential use of definite definitions. But what I think he did not see is that
a definite description may have a quite different rolemay be used
nonreferentially, even as it occurs in one and the same sentence.
Strawson, it is true, points out nonreferential uses of definite descriptions,3 but which use a definite description has seems to be for him a
function of the kind of sentence in which it occurs; whereas, if I am
right, there can be two possible uses of a definite description in the
same sentence. Thus, in On Referring, he says, speaking of expressions used to refer, Any expression of any of these classes [one being
that of definite descriptions] can occur as the subject of what would
traditionally be regarded as a singular subject-predicate sentence;
4

REFERENCE AND DEFINITE DESCRIPTIONS

and would, so occurring, exemplify the use I wish to discuss.4 So the

definite description in, say, the sentence The Republican candidate
for president in 1968 will be a conservative presumably exemplifies
the referential use. But if I am right, we could not say this of the sentence in isolation from some particular occasion on which it is used
to state something; and then it might or might not turn out that the
definite description has a referential use.
Strawson and Russell seem to me to make a common assumption here about the question of how definite descriptions function:
that we can ask how a definite description functions in some sentence independently of a particular occasion upon which it is used.
This assumption is not really rejected in Strawsons arguments
against Russell. Although he can sum up his position by saying,
Mentioning or referring is not something an expression does; it
is something that someone can use an expression to do,5 he means
by this to deny the radical view that a genuine referring expression
has a referent, functions to refer, independent of the context of some
use of the expression. The denial of this view, however, does not
entail that definite descriptions cannot be identified as referring expressions in a sentence unless the sentence is being used. Just as we
can speak of a function of a tool that is not at the moment performing its function, Strawsons view, I believe, allows us to speak of the
referential function of a definite description in a sentence even when
it is not being used. This, I hope to show, is a mistake.
A second assumption shared by Russells and Strawsons account
of definite descriptions is this. In many cases a person who uses a
definite description can be said (in some sense) to presuppose or
imply that something fits the description.6 If I state that the king is
on his throne, I presuppose or imply that there is a king. (At any rate,
this would be a natural thing to say for anyone who doubted that
there is a king.) Both Russell and Strawson assume that where the
5

E S S AY S O N R E F E R E N C E

presupposition or implication is false, the truth value of what the

speaker says is affected. For Russell the statement made is false; for
Strawson it has no truth value. Now 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. This is
what I shall in fact argue. It will turn out, I believe, that one or the
other of the two views, Russells or Strawsons, may be correct about
the nonreferential use of definite descriptions, but neither fits the
referential use. This is not so surprising about Russells view, since
he did not recognize this use in any case, but it is surprising about
Strawsons since the referential use is what he tries to explain and
defend. Furthermore, on Strawsons account, the result of there
being nothing which fits the description is a failure of reference.7
This too, I believe, turns out not to be true about the referential use
of definite descriptions.

II
There are some uses of definite descriptions which carry neither any
hint of a referential use nor any presupposition or implication that
something fits the description. In general, it seems, these are recognizable from the sentence frame in which the description occurs.
These uses will not interest us, but it is necessary to point them out
if only to set them aside.
An obvious example would be the sentence The present king of
France does not exist, used, say, to correct someones mistaken impression that de Gaulle is the king of France.
A more interesting example is this. Suppose someone were to ask,
Is de Gaulle the king of France? This is the natural form of words for
a person to use who is in doubt as to whether de Gaulle is king or
president of France. Given this background to the question, there
6

REFERENCE AND DEFINITE DESCRIPTIONS

seems to be no presupposition or implication that someone is the

king of France. Nor is the person attempting to refer to someone by
using the definite description. On the other hand, reverse the name
and description in the question and the speaker probably would be
thought to presuppose or imply this. Is the king of France de Gaulle?
is the natural question for one to ask who wonders whether it is de
Gaulle rather than someone else who occupies the throne of France.8
Many times, however, the use of a definite description does carry
a presupposition or implication that something fits the description.
If definite descriptions do have a referring role, it will be here. But it
is a mistake, I think, to try, as I believe both Russell and Strawson do,
to settle this matter without further ado. What is needed, I believe,
is the distinction I will now discuss.

III
I will call the two uses of definite descriptions I have in mind the
attributive use and the referential use. A speaker who uses a definite
description attributively in an assertion states something about
whoever or whatever is the so-and-so. A speaker who uses a definite
description referentially in an assertion, on the other hand, uses the
description to enable his audience to pick out whom or what he is
talking about and states something about that person or thing. In
the first case the definite description might be said to occur essentially, for the speaker wishes to assert something about whatever or
whoever fits that description; but in the referential use the definite
description is merely one tool for doing a certain jobcalling attention to a person or thingand in general any other device for doing
the same job, another description or a name, would do as well. In the
attributive use, the attribute of being the so-and-so is all important,
while it is not in the referential use.
7

E S S AY S O N R E F E R E N C E

To illustrate this distinction, in the case of a single sentence, consider the sentence, Smiths murderer is insane. Suppose first that
we come upon poor Smith foully murdered. From the brutal manner
of the killing and the fact that Smith was the most lovable person in
the world, we might exclaim, Smiths murderer is insane. I will
assume, to make it a simpler case, that in a quite ordinary sense we
do not know who murdered Smith (though this is not in the end
essential to the case). This, I shall say, is an attributive use of the definite description.
The contrast with such a use of the sentence is one of those situations in which we expect and intend our audience to realize whom
we have in mind when we speak of Smiths murderer and, most importantly, to know that it is this person about whom we are going to
say something.
For example, suppose that Jones has been charged with Smiths
murder and has been placed on trial. Imagine that there is a discussion of Joness odd behavior at his trial. We might sum up our
impression of his behavior by saying, Smiths murderer is insane.
If someone asks to whom we are referring, by using this description,
the answer here is Jones. This, I shall say, is a referential use of the
definite description.
That these two uses of the definite description in the same sentence are really quite different can perhaps best be brought out by
considering the consequences of the assumption that Smith had no
murderer (for example, he in fact committed suicide). In both situations, in using the definite description Smiths murderer, the
speaker in some sense presupposes or implies that there is a murderer. But when we hypothesize that the presupposition or implication is false, there are different results for the two uses. In both cases
we have used the predicate is insane, but in the first case, if there is
no murderer, there is no person of whom it could be correctly said
8

REFERENCE AND DEFINITE DESCRIPTIONS

that we attributed insanity to him. Such a person could be identified

(correctly) only in case someone fitted the description used. But in
the second case, where the definite description is simply a means of
identifying the person we want to talk about, it is quite possible for
the correct identification to be made even though no one fits the
description we used.9 We were speaking about Jones even though he
is not in fact Smiths murderer and, in the circumstances imagined,
it was his behavior we were commenting upon. Jones might, for example, accuse us of saying false things of him in calling him insane
and it would be no defense, I should think, that our description, the
murderer of Smith, failed to fit him.
It is, moreover, perfectly possible for our audience to know to
whom we refer, in the second situation, even though they do not
share our presupposition. A person hearing our comment in the
context imagined might know we are talking about Jones even
though he does not think Jones guilty.
Generalizing from this case, we can say, I think, that there are
two uses of sentences of the form, The is . In the first, if nothing
is the then nothing has been said to be . In the second, the fact
that nothing is the does not have this consequence.
With suitable changes the same difference in use can be formulated for uses of language other than assertions. Suppose one is at a
party and, seeing an interesting-looking person holding a martini
glass, one asks, Who is the man drinking a martini? If it should
turn out that there is only water in the glass, one has nevertheless
asked a question about a particular person, a question that it is possible for someone to answer. Contrast this with the use of the same
question by the chairman of the local Teetotalers Union. He has just
been informed that a man is drinking a martini at their annual party.
He responds by asking his informant, Who is the man drinking a
martini? In asking the question the chairman does not have some
9

E S S AY S O N R E F E R E N C E

particular person in mind about whom he asks the question; if no

one is drinking a martini, if the information is wrong, no person can
be singled out as the person about whom the question was asked.
Unlike the first case, the attribute of being the man drinking a martini is all-important, because if it is the attribute of no one, the chairmans question has no straightforward answer.
This illustrates also another difference between the referential
and the attributive use of definite descriptions. In the one case we
have asked a question about a particular person or thing even though
nothing fits the description we used; in the other this is not so. But
also in the one case our question can be answered; in the other it
cannot be. In the referential use of a definite description we may
succeed in picking out a person or thing to ask a question about
even though he or it does not really fit the description; but in the
attributive use if nothing fits the description, no straightforward
answer to the question can be given.
This further difference is also illustrated by commands or orders
containing definite descriptions. Consider the order, Bring me the
book on the table. If the book on the table is being used referentially, it is possible to fulfill the order even though there is no book
on the table. If, for example, there is a book beside the table, though
there is none on it, one might bring that book back and ask the issuer
of the order whether this is the book you meant. And it may be.
But imagine we are told that someone has laid a book on our prize
antique table, where nothing should be put. The order, Bring me
the book on the table cannot now be obeyed unless there is a book
that has been placed on the table. There is no possibility of bringing
back a book which was never on the table and having it be the one
that was meant, because there is no book that in that sense was
meant. In the one case the definite description was a device for
getting the other person to pick the right book; if he is able to pick
10

REFERENCE AND DEFINITE DESCRIPTIONS

the right book even though it does not satisfy the description, one
still succeeds in his purpose. In the other case, there is, antecedently,
no right book except one which fits the description; the attribute
of being the book on the table is essential. Not only is there no book
about which an order was issued, if there is no book on the table, but
the order itself cannot be obeyed. When a definite description is
used attributively in a command or question and nothing fits the
description, the command cannot be obeyed and the question
cannot be answered. This suggests some analogous consequence for
assertions containing definite descriptions used attributively. Perhaps the analogous result is that the assertion is neither true nor
false: this is Strawsons view of what happens when the presupposition of the use of a definite description is false. But if so, Strawsons
view works not for definite descriptions used referentially, but for
the quite different use, which I have called the attributive use.
I have tried to bring out the two uses of definite descriptions by
pointing out the different consequences of supposing that nothing
fits the description used. There are still other differences. One is
this: when a definite description is used referentially, not only is
there in some sense a presupposition or implication that someone
or something fits the description, as there is also in the attributive
use, but there is a quite different presupposition; the speaker presupposes of some particular someone or something that he or it fits
the description. In asking, for example, Who is the man drinking a
martini? where we mean to ask a question about that man over
there, we are presupposing that that man over there is drinking a
martininot just that someone is a man drinking a martini. When
we say, in a context where it is clear we are referring to Jones, Smiths
murderer is insane, we are presupposing that Jones is Smiths murderer. No such presupposition is present in the attributive use of
definite descriptions. There is, of course, the presupposition that
11

E S S AY S O N R E F E R E N C E

someone or other did the murder, but the speaker does not presuppose of someone in particularJones or Robinson, saythat he
did it. What I mean by this second kind of presupposition that
someone or something in particular fits the descriptionwhich is
present in a referential use but not in an attributive usecan perhaps be seen more clearly by considering a member of the speakers
audience who believes that Smith was not murdered at all. Now in
the case of the referential use of the description, Smiths murderer,
he could accuse the speaker of mistakenly presupposing both that
someone or other is the murderer and that also Jones is the murderer, for even though he believes Jones not to have done the deed,
he knows that the speaker was referring to Jones. But in the case of
the attributive use, he can accuse the speaker of having only the first,
less specific presupposition; he cannot pick out some person and
claim that the speaker is presupposing that that person is Smiths
murderer. Now the more particular presuppositions that we find
present in referential uses are clearly not ones we can assign to a
definite description in some particular sentence in isolation from a
context of use. In order to know that a person presupposes that
Jones is Smiths murderer in using the sentence Smiths murderer is
insane, we have to know that he is using the description referentially and also to whom he is referring. The sentence by itself does
not tell us any of this.

IV
From the way in which I set up each of the previous examples it
might be supposed that the important difference between the referential and the attributive use lies in the beliefs of the speaker. Does
he believe of some particular person or thing that he or it fits the
12

REFERENCE AND DEFINITE DESCRIPTIONS

description used? In the Smith murder example, for instance, there

was in the one case no belief as to who did the deed, whereas in the
contrasting case it was believed that Jones did it. But this is, in fact,
not an essential difference. It is possible for a definite description to
be used attributively even though the speaker (and his audience)
believes that a certain person or thing fits the description. And it is
possible for a definite description to be used referentially where the
speaker believes that nothing fits the description. It is trueand
this is why, for simplicity, I set up the examples the way I didthat
if a speaker does not believe that anything fits the description or
does not believe that he is in a position to pick out what does fit the
description, it is likely that he is not using it referentially. It is also
true that if he and his audience would pick out some particular thing
or person as fitting the description, then a use of the definite description is very likely referential. But these are only presumptions
and not entailments.
To use the Smith murder case again, suppose that Jones is on
trial for the murder and I and everyone else believe him guilty. Suppose that I comment that the murderer of Smith is insane, but instead of backing this up, as in the example previously used, by citing
Joness behavior in the dock, I go on to outline reasons for thinking
that anyone who murdered poor Smith in that particularly horrible
way must be insane. If now it turns out that Jones was not the murderer after all, but someone else was, I think I can claim to have been
right if the true murderer is after all insane. Here, I think, I would be
using the definite description attributively, even though I believe
that a particular person fits the description.
It is also possible to think of cases in which the speaker does not
believe that what he means to refer to by using the definite description fits the description, or to imagine cases in which the definite
description is used referentially even though the speaker believes
13

E S S AY S O N R E F E R E N C E

nothing fits the description. Admittedly, these cases may be parasitic

on a more normal use; nevertheless, they are sufficient to show that
such beliefs of the speaker are not decisive as to which use is made
of a definite description.
Suppose the throne is occupied by a man I firmly believe to be
not the king, but a usurper. Imagine also that his followers as firmly
believe that he is the king. Suppose I wish to see this man. I might
say to his minions, Is the king in his countinghouse? I succeed in
referring to the man I wish to refer to without myself believing that
he fits the description. It is not even necessary, moreover, to suppose that his followers believe him to be the king. If they are cynical
about the whole thing, know he is not the king, I may still succeed
in referring to the man I wish to refer to. Similarly, neither I nor the
people I speak to may suppose that anyone is the king and, finally,
each party may know that the other does not so suppose and yet the
reference may go through.

V
Both the attributive and the referential use of definite descriptions
seem to carry a presupposition or implication that there is something which fits the description. But the reasons for the existence of
the presupposition or implication are different in the two cases.
There is a presumption that a person who uses a definite description referentially believes that what he wishes to refer to fits the description. Because the purpose of using the description is to get the
audience to pick out or think of the right thing or person, one would
normally choose a description that he believes the thing or person
fits. Normally a misdescription of that to which one wants to refer
would mislead the audience. Hence, there is a presumption that the
14

REFERENCE AND DEFINITE DESCRIPTIONS

speaker believes something fits the descriptionnamely, that to

which he refers.
When a definite description is used attributively, however, there
is not the same possibility of misdescription. In the example of
Smiths murderer used attributively, there was not the possibility
of misdescribing Jones or anyone else; we were not referring to
Jones nor to anyone else by using the description. The presumption
that the speaker believes someone is Smiths murderer does not arise
here from a more specific presumption that he believes Jones or
Robinson or someone else whom he can name or identify is Smiths
murderer.
The presupposition or implication is borne by a definite description used attributively because if nothing fits the description the
linguistic purpose of the speech act will be thwarted. That is, the
speaker will not succeed in saying something true, if he makes an
assertion; he will not succeed in asking a question that can be answered, if he has asked a question; he will not succeed in issuing an
order that can be obeyed, if he has issued an order. If one states that
Smiths murderer is insane, when Smith has no murderer, and uses
the definite description nonreferentially, then one fails to say anything true. If one issues the order Bring me Smiths murderer under
similar circumstances, the order cannot be obeyed; nothing would
count as obeying it.
When the definite description is used referentially, on the other
hand, the presupposition or implication stems simply from the fact
that normally a person tries to describe correctly what he wants to
refer to because normally this is the best way to get his audience to
recognize what he is referring to. As we have seen, it is possible for
the linguistic purpose of the speech act to be accomplished in such
a case even though nothing fits the description; it is possible to say
something true or to ask a question that gets answered or to issue a
15

E S S AY S O N R E F E R E N C E

command that gets obeyed. For when the definite description is

used referentially, ones audience may succeed in seeing to what one
refers even though neither it nor anything else fits the description.

VI
The result of the last section shows something to be wrong with the
theories of both Russell and Strawson; for though they give differing accounts of the implication or presupposition involved, each
gives only one. Yet, as I have argued, the presupposition or implication is present for a quite different reason, depending upon whether
the definite description is used attributively or referentially, and exactly what presuppositions or implications are involved is also different. Moreover, neither theory seems a correct characterization
of the referential use. On Russells there is a logical entailment:
The is entails There exists one and only one . Whether or
not this is so for the attributive use, it does not seem true of the
referential use of the definite description. The implication that
something is the , as I have argued, does not amount to an entailment; it is more like a presumption based on what is usually true of
the use of a definite description to refer. In any case, of course, Russells theory does not showwhat is true of the referential use
that the implication that something is the comes from the more
specific implication that what is being referred to is the . Hence, as
a theory of definite descriptions, Russells view seems to apply, if at
all, to the attributive use only.
Russells definition of denoting (a definite description denotes
an entity if that entity fits the description uniquely) is clearly applicable to either use of definite descriptions. Thus whether or not a
definite description is used referentially or attributively, it may have
16

REFERENCE AND DEFINITE DESCRIPTIONS

a denotation. Hence, denoting and referring, as I have explicated

the latter notion, are distinct and Russells view recognizes only the
former. It seems to me, moreover, that this is a welcome result, that
denoting and referring should not be confused. If one tried to maintain that they are the same notion, one result would be that a speaker
might be referring to something without knowing it. If someone
said, for example, in 1960 before he had any idea that Mr. Goldwater
would be the Republican nominee in 1964, The Republican candidate for president in 1964 will be a conservative (perhaps on the
basis of an analysis of the views of party leaders), the definite description here would denote Mr. Goldwater. But would we wish to
say that the speaker had referred to, mentioned, or talked about Mr.
Goldwater? I feel these terms would be out of place. Yet if we identify referring and denoting, it ought to be possible for it to turn out
(after the Republican Convention) that the speaker had, unknown
to himself, referred in 1960 to Mr. Goldwater. On my view, however,
while the definite description used did denote Mr. Goldwater (using
Russells definition), the speaker used it attributively and did not
refer to Mr. Goldwater.
Turning to Strawsons theory, it was supposed to demonstrate
how definite descriptions are referential. But it goes too far in this
direction. For there are nonreferential uses of definite descriptions
also, even as they occur in one and the same sentence. I believe that
Strawsons theory involves the following propositions:
(1) If someone asserts that the is he has not made a true or
false statement if there is no .10
(2) If there is no then the speaker has failed to refer to
anything.11
(3) The reason he has said nothing true or false is that he has
failed to refer.
17

E S S AY S O N R E F E R E N C E

Each of these propositions is either false or, at best, applies to only

one of the two uses of definite descriptions.
Proposition (1) is possibly true of the attributive use. In the example in which Smiths murderer is insane was said when Smiths
body was first discovered, an attributive use of the definite description, there was no person to whom the speaker referred. If Smith
had no murderer, nothing true was said. It is quite tempting to conclude, following Strawson, that nothing true or false was said. But
where the definite description is used referentially, something true
may well have been said. It is possible that something true was said
of the person or thing referred to.12
Proposition (2) is, as we have seen, simply false. Where a definite description is used referentially it is perfectly possible to refer to
something though nothing fits the description used.
The situation with proposition (3) is a bit more complicated.
It ties together, on Strawsons view, the two strands given in
(1) and (2). As an account of why, when the presupposition is
false, nothing true or false has been stated, it clearly cannot work
for the attributive use of definite descriptions, for the reason it
supplies is that reference has failed. It does not then give the
reason why, if indeed this is so, a speaker using a definite description attributively fails to say anything true or false if nothing fits
the description. It does, however, raise a question about the referential use. Can reference fail when a definite description is used
referentially?
I do not fail to refer merely because my audience does not correctly pick out what I am referring to. I can be referring to a particular man when I use the description the man drinking a martini,
even though the people to whom I speak fail to pick out the right
person or any person at all. Nor, as we have stressed, do I fail to refer
when nothing fits the description. But perhaps I fail to refer in some
18

REFERENCE AND DEFINITE DESCRIPTIONS

extreme circumstances, when there is nothing that I am willing to

pick out as that to which I referred.
Suppose that I think I see at some distance a man walking and
ask, Is the man carrying a walking stick the professor of history?
We should perhaps distinguish four cases at this point: (a) There is a
man carrying a walking stick; I have then referred to a person and
asked a question about him that can be answered if my audience has
the information. (b) The man over there is not carrying a walking
stick, but an umbrella; I have still referred to someone and asked a
question that can be answered, though if my audience sees that it is
an umbrella and not a walking stick, they may also correct my apparently mistaken impression. (c) It is not a man at all, but a rock that
looks like one; in this case, I think I still have referred to something,
to the thing over there that happens to be a rock but that I took to be
a man. But in this case it is not clear that my question can be answered correctly. This, I think, is not because I have failed to refer,
but rather because, given the true nature of what I referred to, my
question is not appropriate. A simple No, that is not the professor of
history is at least a bit misleading if said by someone who realizes
that I mistook a rock for a person. It may, therefore, be plausible to
conclude that in such a case I have not asked a question to which
there is a straightforwardly correct answer. But if this is true, it is not
because nothing fits the description I used, but rather because what
I referred to is a rock and my question has no correct answer when
asked of a rock. (d) There is finally the case in which there is nothing
at all where I thought there was a man with a walking stick; and perhaps here we have a genuine failure to refer at all, even though the
description was used for the purpose of referring. There is no rock,
nor anything else, to which I meant to refer; it was, perhaps, a trick of
light that made me think there was a man there. I cannot say of anything, That is what I was referring to, though I now see that its not
19

E S S AY S O N R E F E R E N C E

a man carrying a walking stick. This failure of reference, however,

requires circumstances much more radical than the mere nonexistence of anything fitting the description used. It requires that there
be nothing of which it can be said, That is what he was referring to.
Now perhaps also in such cases, if the speaker has asserted something, he fails to state anything true or false if there is nothing that
can be identified as that to which he referred. But if so, the failure of
reference and truth value does not come about merely because nothing fits the description he used. So (3) may be true of some cases of
the referential use of definite descriptions; it may be true that a failure of reference results in a lack of truth value. But these cases are of
a much more extreme sort than Strawsons theory implies.
I conclude, then, that neither Russells nor Strawsons theory
represents a correct account of the use of definite descriptions
Russells because it ignores altogether the referential use, Strawsons
because it fails to make the distinction between the referential and
the attributive and mixes together truths about each (together with
some things that are false).

VII
It does not seem possible to say categorically of a definite description in a particular sentence that it is a referring expression (of
course, one could say this if he meant that it might be used to refer).
In general, whether or not a definite description is used referentially
or attributively is a function of the speakers intentions in a particular case. The murderer of Smith may be used either way in the sentence The murderer of Smith is insane. It does not appear plausible
to account for this, either, as an ambiguity in the sentence. The
grammatical structure of the sentence seems to me to be the same
20

REFERENCE AND DEFINITE DESCRIPTIONS

whether the description is used referentially or attributively: that is,

it is not syntactically ambiguous. Nor does it seem at all attractive to
suppose an ambiguity in the meaning of the words; it does not
appear to be semantically ambiguous. (Perhaps we could say that
the sentence is pragmatically ambiguous: the distinction between
roles that the description plays is a function of the speakers intentions.) These, of course, are intuitions; I do not have an argument
for these conclusions. Nevertheless, the burden of proof is surely on
the other side.
This, I think, means that the view, for example, that sentences
can be divided up into predicates, logical operators, and referring
expressions is not generally true. In the case of definite descriptions
one cannot always assign the referential function in isolation from a
particular occasion on which it is used.
There may be sentences in which a definite description can be
used only attributively or only referentially. A sentence in which it
seems that the definite description could be used only attributively
would be Point out the man who is drinking my martini. I am not
so certain that any can be found in which the definite description
can be used only referentially. Even if there are such sentences, it
does not spoil the point that there are many sentences, apparently
not ambiguous either syntactically or semantically, containing definite descriptions that can be used either way.
If it could be shown that the dual use of definite descriptions can
be accounted for by the presence of an ambiguity, there is still a
point to be made against the theories of Strawson and Russell. For
neither, so far as I can see, has anything to say about the possibility
of such an ambiguity and, in fact, neither seems compatible with
such a possibility. Russells does not recognize the possibility of the
referring use, and Strawsons, as I have tried to show in the last section, combines elements from each use into one unitary account.
21

E S S AY S O N R E F E R E N C E

Thus the view that there is an ambiguity in such sentences does not
seem any more attractive to these positions.

VIII
Using a definite description referentially, a speaker may say something true even though the description correctly applies to nothing.
The sense in which he may say something true is the sense in which
he may say something true about someone or something. This sense
is, I think, an interesting one that needs investigation. Isolating it is
one of the by-products of the distinction between the attributive
and referential uses of definite descriptions.
For one thing, it raises questions about the notion of a statement. This is brought out by considering a passage in a paper by
Leonard Linsky in which he rightly makes the point that one can
refer to someone although the definite description used does not
correctly describe the person:
. . . said of a spinster that Her husband is kind to her is neither
true nor false. But a speaker might very well be referring to someone using these words, for he may think that someone is the husband of the lady (who in fact is a spinster). Still, the statement is
neither true nor false, for it presupposes that the lady has a husband, which she has not. This last refutes Strawsons thesis that if
the presupposition of existence is not satisfied, the speaker has
failed to refer.13

There is much that is right in this passage. But because Linsky does
not make the distinction between the referential and the attributive
uses of definite descriptions, it does not represent a wholly adequate
22

REFERENCE AND DEFINITE DESCRIPTIONS

account of the situation. A perhaps minor point about this passage is

that Linsky apparently thinks it sufficient to establish that the speaker
in his example is referring to someone by using the definite description her husband, that he believe that someone is her husband. This
will only approximate the truth provided that the someone in the
description of the belief means someone in particular and is not
merely the existential quantifier, there is someone or other. For in
both the attributive and the referential use the belief that someone
or other is the husband of the lady is very likely to be present. If, for
example, the speaker has just met the lady and, noticing her cheerfulness and radiant good health, makes his remark from his conviction that these attributes are always the result of having good
husbands, he would be using the definite description attributively.
Since she has no husband, there is no one to pick out as the person
to whom he was referring. Nevertheless, the speaker believed that
someone or other was her husband. On the other hand, if the use of
her husband was simply a way of referring to a man the speaker has
just met whom he assumed to be the ladys husband, he would have
referred to that man even though neither he nor anyone else fits the
description. I think it is likely that in this passage Linsky did mean by
someone, in his description of the belief, someone in particular.
But even then, as we have seen, we have neither a sufficient nor a
necessary condition for a referential use of the definite description.
A definite description can be used attributively even when the
speaker believes that some particular thing or person fits the description, and it can be used referentially in the absence of this belief.
My main point, here, however, has to do with Linskys view that
because the presupposition is not satisfied, the statement is neither
true nor false. This seems to me possibly correct if the definite description is thought of as being used attributively (depending upon
whether we go with Strawson or Russell). But when we consider it
23

E S S AY S O N R E F E R E N C E

as used referentially, this categorical assertion is no longer clearly

correct. For the man the speaker referred to may indeed be kind to
the spinster; the speaker may have said something true about that
man. Now the difficulty is in the notion of the statement. Suppose
that we know that the lady is a spinster, but nevertheless know that
the man referred to by the speaker is kind to her. It seems to me that
we shall, on the one hand, want to hold that the speaker said something true, but be reluctant to express this by It is true that her husband is kind to her.
This shows, I think, a difficulty in speaking simply about the
statement when definite descriptions are used referentially. For the
speaker stated something, in this example, about a particular person,
and his statement, we may suppose, was true. Nevertheless, we
should not like to agree with his statement by using the sentence he
used; we should not like to identify the true statement via the speakers words. The reason for this is not so hard to find. If we say, in this
example, It is true that her husband is kind to her, we are now using
the definite description either attributively or referentially. But we
should not be subscribing to what the original speaker truly said if
we use the description attributively, for it was only in its function as
referring to a particular person that the definite description yields
the possibility of saying something true (since the lady has no husband). Our reluctance, however, to endorse the original speakers
statement by using the definite description referentially to refer to
the same person stems from quite a different consideration. For if
we too were laboring under the mistaken belief that this man was
the ladys husband, we could agree with the original speaker using
his exact words. (Moreover, it is possible, as we have seen, deliberately to use a definite description to refer to someone we believe not
to fit the description.) Hence, our reluctance to use the original
speakers words does not arise from the fact that if we did we should
24

REFERENCE AND DEFINITE DESCRIPTIONS

not succeed in stating anything true or false. It rather stems from the
fact that when a definite description is used referentially there is a
presumption that the speaker believes that what he refers to fits the
description. Since we, who know the lady to be a spinster, would not
normally want to give the impression that we believe otherwise, we
would not like to use the original speakers way of referring to the
man in question.
How then would we express agreement with the original speaker
without involving ourselves in unwanted impressions about our beliefs? The answer shows another difference between the referential
and attributive uses of definite descriptions and brings out an important point about genuine referring.
When a speaker says, The is , where the is used attributively, if there is no , we cannot correctly report the speaker as
having said of this or that person or thing that it is . But if the definite description is used referentially we can report the speaker as
having attributed to something. And we may refer to what the
speaker referred to, using whatever description or name suits our
purpose. Thus, if a speaker says, Her husband is kind to her, referring to the man he was just talking to, and if that man is Jones, we
may report him as having said of Jones that he is kind to her. If Jones
is also the president of the college, we may report the speaker as
having said of the president of the college that he is kind to her. And
finally, if we are talking to Jones, we may say, referring to the original
speaker, He said of you that you are kind to her. It does not matter
here whether or not the woman has a husband or whether, if she
does, Jones is her husband. If the original speaker referred to Jones,
he said of him that he is kind to her. Thus where the definite description is used referentially, but does not fit what was referred to, we
can report what a speaker said and agree with him by using a
description or name which does fit. In doing so we need not, it is
25

E S S AY S O N R E F E R E N C E

important to note, choose a description or name which the original

speaker would agree fits what he was referring to. That is, we can
report the speaker in the above case to have said truly of Jones that
he is kind to her even if the original speaker did not know that the
man he was referring to is named Jones or even if he thinks he is not
named Jones.
Returning to what Linsky said in the passage quoted, he
claimed that, were someone to say Her husband is kind to her,
when she has no husband, the statement would be neither true nor
false. As I have said, this is a likely view to hold if the definite description is being used attributively. But if it is being used referentially it is not clear what is meant by the statement. If we think
about what the speaker said about the person he referred to, then
there is no reason to suppose he has not said something true or
false about him, even though he is not the ladys husband. And
Linskys claim would be wrong. On the other hand, if we do not
identify the statement in this way, what is the statement that the
speaker made? To say that the statement he made was that her
husband is kind to her lands us in difficulties. For we have to decide
whether in using the definite description here in the identification
of the statement, we are using it attributively or referentially. If the
former, then we misrepresent the linguistic performance of the
speaker; if the latter, then we are ourselves referring to someone
and reporting the speaker to have said something of that person,
in which case we are back to the possibility that he did say something true or false of that person.
I am thus drawn to the conclusion that when a speaker uses a
definite description referentially he may have stated something true
or false even if nothing fits the description, and that there is not a
clear sense in which he has made a statement which is neither true
nor false.
26

REFERENCE AND DEFINITE DESCRIPTIONS

IX
I want to end by a brief examination of a picture of what a genuine
referring expression is that one might derive from Russells views.
I want to suggest that this picture is not so far wrong as one might
suppose and that strange as this may seem, some of the things we
have said about the referential use of definite descriptions are not
foreign to this picture.
Genuine proper names, in Russells sense, would refer to something without ascribing any properties to it. They would, one
might say, refer to the thing itself, not simply the thing insofar as it
falls under a certain description.14 Now this would seem to Russell
something a definite description could not do, for he assumed that
if definite descriptions were capable of referring at all, they would
refer to something only insofar as that thing satisfied the description. Not only have we seen this assumption to be false, however,
but in the last section we saw something more. We saw that when
a definite description is used referentially, a speaker can be reported as having said something of something. And in reporting
what it was of which he said something we are not restricted to the
description he used, or synonyms of it; we may ourselves refer to
it using any descriptions, names, and so forth, that will do the job.
Now this seems to give a sense in which we are concerned with the
thing itself and not just the thing under a certain description,
when we report the linguistic act of a speaker using a definite description referentially. That is, such a definite description comes
closer to performing the function of Russells proper names than
certainly he supposed.
Russell thought, I believe, that whenever we use descriptions,
as opposed to proper names, we introduce an element of generality which ought to be absent if what we are doing is referring
27

E S S AY S O N R E F E R E N C E

to some particular thing. This is clear from his analysis of sentences

containing definite descriptions. One of the conclusions we are supposed to draw from that analysis is that such sentences express what
are in reality completely general propositions: there is a and only
one such and any is . We might put this in a slightly different way.
If there is anything which might be identified as reference here, it is
reference in a very weak sensenamely, reference to whatever is the
one and only one , if there is any such. Now this is something we
might well say about the attributive use of definite descriptions, as
should be evident from the previous discussion. But this lack of particularity is absent from the referential use of definite descriptions
precisely because the description is here merely a device for getting
ones audience to pick out or think of the thing to be spoken about,
a device which may serve its function even if the description is incorrect. More importantly perhaps, in the referential use as opposed
to the attributive, there is a right thing to be picked out by the audience and its being the right thing is not simply a function of its fitting the description.

NOTES
1. I should like to thank my colleagues, John Canfield, Sydney Shoemaker, and Timothy Smiley, who read an earlier draft and gave me helpful suggestions. I also had
the benefit of the valuable and detailed comments of the referee for the paper, to
whom I wish to express my gratitude.
2. On Denoting, reprinted in Logic and Knowledge, ed. by Robert C. Marsh
(London, 1956), p. 51.
3. On Referring, reprinted in Philosophy and Ordinary Language, ed. by Charles
C. Caton (Urbana, 1963), pp. 16263.
4. Ibid., p. 162.
5. Ibid., p. 170.
6. Here and elsewhere I use the disjunction presuppose or imply to avoid taking
a stand that would side me with Russell or Strawson on the issue of what the
relationship involved is. To take a stand here would be beside my main point as
28

REFERENCE AND DEFINITE DESCRIPTIONS

well as being misleading, since later on I shall argue that the presupposition
orimplication arises in a different way depending upon the use to which the definite description is put. This last also accounts for my use of the vagueness indicator, in some sense.
7. In a footnote added to the original version of On Referring (op. cit., p. 181)
Strawson seems to imply that where the presupposition is false, we still succeed
in referring in a secondary way, which seems to mean as we could be said to
refer to fictional or make-believe things. But his view is still that we cannot refer
in such a case in the primary way. This is, I believe, wrong. For a discussion of
this modification of Strawsons view see Charles C. Caton, Strawson on Referring, Mind, LXVIII (1959), 53944.
8. This is an adaptation of an example (used for a somewhat different purpose)
given by Leonard Linsky in Reference and Referents, in Philosophy and Ordinary Language, p. 80.
9. In Reference and Referents (pp. 7475, 80), Linsky correctly points out that
one does not fail to refer simply because the description used does not in fact fit
anything (or fits more than one thing). Thus he pinpoints one of the difficulties
in Strawsons view. Here, however, I use this fact about referring to make a distinction I believe he does not draw, between two uses of definite descriptions.
I later discuss the second passage from Linskys paper.
10. In A Reply to Mr. Sellars, Philosophical Review, LXIII (1954), 21631, Strawson admits that we do not always refuse to ascribe truth to what a person says
when the definite description he uses fails to fit anything (or fits more than one
thing). To cite one of his examples, a person who said, The United States Chamber of Deputies contains representatives of two major parties, would be allowed
to have said something true even though he had used the wrong title. Strawson
thinks this does not constitute a genuine problem for his view. He thinks that
what we do in such cases, where the speakers intended reference is pretty clear,
is simply to amend his statement in accordance with his guessed intentions and
assess the amended statement for truth or falsity; we are not awarding a truth
value at all to the original statement (p. 230).
The notion of an amended statement, however, will not do. We may note,
first of all, that the sort of case Strawson has in mind could arise only when a definite description is used referentially. For the amendment is made by seeing the
speakers intended reference. But this could happen only if the speaker had an
intended reference, a particular person or thing in mind, independent of the description he used. The cases Strawson has in mind are presumably not cases of
slips of the tongue or the like; presumably they are cases in which a definite description is used because the speaker believes, though he is mistaken, that he is
describing correctly what he wants to refer to. We supposedly amend the statement by knowing to what he intends to refer. But what description is to be used
in the amended statement? In the example, perhaps, we could use the United
States Congress. But this description might be one the speaker would not

29

E S S AY S O N R E F E R E N C E

even accept as correctly describing what he wants to refer to, because he is

misinformed about the correct title. Hence, this is not a case of deciding what the
speaker meant to say as opposed to what he in fact said, for the speaker did not
mean to say the United States Congress. If this is so, then there is no bar to the
amended statement containing any description that does correctly pick out
what the speaker intended to refer to. It could be, e.g., The lower house of the
United States Congress. But this means that there is no one unique amended
statement to be assessed for truth value. And, in fact, it should now be clear that
the notion of the amended statement really plays no role anyway. For if we can
arrive at the amended statement only by first knowing to what the speaker intended to refer, we can assess the truth of what he said simply by deciding
whether what he intended to refer to has the properties he ascribed to it.
11. As noted earlier (n. 7), Strawson may allow that one has possibly referred in a
secondary way, but, if I am right, the fact that there is no does not preclude
one from having referred in the same way one does if there is a .
12. For a further discussion of the notion of saying something true of someone or
something, see sec. VIII.
13. Reference and Referents, p. 80. It should be clear that I agree with Linsky in
holding that a speaker may refer even though the presupposition of existence is
not satisfied. And I agree in thinking this an objection to Strawsons view. I think,
however, that this point, among others, can be used to define two distinct uses of
definite descriptions which, in turn, yields a more general criticism of Strawson.
So, while I develop here a point of difference, which grows out of the distinction
I want to make, I find myself in agreement with much of Linskys article.
14. Cf. The Philosophy of Logical Atomism, reprinted in Logic and Knowledge,
p. 200.

30

[2]
PUTTING HUMPTY DUMPTY
TOGETHER AGAIN

Who is master, Humpty Dumpty or Humpty Dumptys language?1

Mr. MacKay2 compares some things I said about definite descriptions and reference3 to Humpty Dumptys theory of meaning. In
doing so, I think the principle he draws from Humpty Dumptys
conversation with Alice is that the conventions of ones language
dictate the meaning of ones words, and intentions are powerless to
intervene. Since glory does not mean a nice knockdown argument in the language Humpty Dumpty speaks, he could not have
meant that by glory nor could glory have meant that in the conversation, no matter what intentions he may have had. This has some
importance beyond any application to my article, since theories of
meaning that take the intentions of speakers as primitive (for example, Grices view in his article Meaning)4 must either deny the
principle or show how they avoid running afoul of it. MacKay, of
course, uses Humpty Dumpty only for the purpose of drawing an
analogy and there is no reason to suppose that what I said about
reference commits me to any theory of meaning. Nevertheless, I will
First published in The Philosophical Review, 1968, 77: 21315.

31

E S S AY S O N R E F E R E N C E

have something to say about the significance of Humpty Dumptys

conversation with Alice and about Wittgensteins challenge to Say
Its cold here and mean Its hot here, because it is easy to make a
mistake about them very similar to one I find in MacKays criticisms
of my treatment of reference.
First, however, I want to discuss the connection between what
MacKay says about reference and my article. The purpose of the article he discusses was to make a distinction between two uses of
definite descriptions. MacKay does not deal with the distinction
itself; instead, he criticizes some things I said about reference in the
course of drawing it. Nevertheless, what he says may seem to cast
doubt on the existence of a difference between two uses of definite
descriptions, though I am not sure whether that was his purpose. In
any case, it turns out that MacKays view can be shown to presuppose the distinction. To show this will require a brief look at how
I attempted to draw it.
In the article in question, I pointed out several features of what
I called the referential use of definite descriptions not shared
with another use of them, which I called the attributive use. The
two uses can be thought of as corresponding to two possible purposes a speaker may have in using a definite description. Looking at
the referential use first, a speaker may wish to attribute a property to,
ask a question about, issue an order concerning some particular
person, object, situation, and so forth. We can view his task as that
of finding a description suitable for allowing his audience to identify
that person, object, situation, and so forth. The description chosen
need not be unique; a speaker will often use some description such
as the table, leaving it to his audience to determine from context or
from a knowledge of the speakers likely wants, beliefs, and so forth
to which table he is referring.5 The description chosen may be different for different audiences. Suppose, for example, that a speaker
32

PUTTING HUMPTY DUMPTY TOGETHER AGAIN

wants to relate an amusing thing done by Jones. To one audience he

might begin, The man I introduced to you yesterday . . . , to another
Our mutual friend . . . , to still another The man you see over
there. . . . The distinguishing characteristic of the referential use is
the existence of an entity the speaker wants to talk about and in relation to which he chooses a description as a means of referring to it.
There is a sense in which the particular description chosen is inessential, though not irrelevant, to what the speaker accomplishes, in
this case, telling something about Jones. The following is a familiar
sort of conversation in which a speaker shows his willingness to
abandon one description for another when it becomes evident that
the first is not fulfilling its purpose of identifying for the audience
the reference of the speakers remarks: The man I introduced to
you yesterday did an amusing thing. I dont recall your introducing
anyone to me. Well, perhaps I didnt, but dont you remember the
man in the loud checked suit? Oh, now I know whom you mean.
In contrast, a definite description used attributively is not in the
same way dispensable; it is essential to the purpose of the speech
act. As an illustration, suppose someone to make a wager, expressing his side of the bet in the words, The winner of the Indianapolis
500 race drove a turbine-powered car. The speaker could intend his
bet to be understood not as being a bet about, say, Parnelli Jones or
any other race driver, but about the winner whoever he may have
been. And this could be so even if he happens to have a belief about
who won the race (though, of course, he need have none). Suppose
that he believes Parnelli Jones won the race; nevertheless he will
expect to collect even if this belief is mistaken so long as whoever
won the race did drive a turbine-powered car, and he will lose if
whoever won drove an ordinary car even though Joness car was
turbine-powered. It would be inappropriate and irrelevant to the bet
for the speaker to try out a new descriptionfor example, the man
33

E S S AY S O N R E F E R E N C E

in the red car, when his belief about who won the race turns out to
be mistaken. Similarly, if there were no winner of the 500the race
having been called before the finishthe speaker could not appropriately turn to another description, as the speaker did in the previous example when it turned out that he had not introduced anyone.
Finally, if Parnelli Jones was the winner, one can still say that had he
not been the winner, then what kind of car he drove would have
been irrelevant to the bet.
The central difference between the two uses of definite descriptions that emerges is this. In connection with an attributive use, if we
want to identify some entity as what the speaker was talking about in
using the definite description, the only candidate would be something
that satisfies the description. A referential use, however, allows of another possibilitythat the entity the speaker was talking about does
not in fact satisfy the description. When the entity so identified does
fit the description, it can nevertheless be said that had it not done so,
it still would have been what the speaker was talking about. Sometimes we want to say something about an entity and try to identify it
for our audience via a description of it, thus making it a possibility that
the entity and description do not match up. At other times we want to
talk about whatever fits a certain description (uniquely) and then
there is no possibility of an entity that is at once what we wanted to
talk about but not correctly characterized by the description.
As a consequence of this difference several things seemed to me
to be true of a referential use that are not true of an attributive use.
These were discussed in some detail in the article MacKay criticizes,
so they will only be listed here. If a speaker S uses a definite description, the , referentially there will be some entity e (or, at least, the
speaker will intend that there should be) about which the following
will be true. (I assume that an assertion was made. Appropriate
changes would be required for questions, commands, and so forth.)
34

PUTTING HUMPTY DUMPTY TOGETHER AGAIN

(1) S will have referred to e whether or not e is in fact .

(2) S will have said something true or false about e whether or
not e is in fact (provided that everything is in order concerning the remainder of the speech act).
(3) S, in using the to refer to e, will have presupposed or implied that e is .
(4) In reporting Ss speech act, it will be correct to say that he
stated something about e and in reporting this to use expressions to refer to e other than the or synonyms of it.6

Had the definite description been used attributively there would be

no such entity e (nor would the speaker have intended that there
should be).
The first of these characteristics is the one that bothers MacKay.
His view, I believe, can be summed up in the following four points:
(A) Referring is one way among others of making it known what
a speaker is talking about.
(B) In using a definite description to refer, one succeeds in referring only to something that fits the description used (barring, perhaps, near misses).
(C) What one is talking about, however, may fail altogether to fit
the description used.
(D) In that case, in view of (B), one does not succeed in referring
to what one is talking about, although one may succeed in
making it knowable what one is talking about.

Looking at points (C) and (D) one will notice that MacKay is
describing a use of definite descriptions for which it is possible to
distinguish an entity the speaker is talking about independently of
the description he usesan entity, that is, that may not fit the
35

E S S AY S O N R E F E R E N C E

description. Now the use of definite descriptions that I call referential provides for just such a possibility. If then it can be shown that
there are uses of definite descriptions, in fact those that I call attributive, for which no such possibility exists, MacKays view must
itself yield the very distinction I wanted to make. It will still be true,
of course, that there is an issue. If he is correct, the referential use of
definite descriptions lacks at least one of the features I thought it
had. But the distinction itself would not only remain in the face of
his criticism, it would, in fact, be presupposed by it.
In his discussion, then, MacKay uses the possibility that what a
speaker is talking about may fail to fit the description he uses. But
looking at the example of the attributive use given above, we find
that this possibility is not present. A wager is made, expressed in the
words, The winner of the Indianapolis 500 drove a turbine-powered car. The bettor collects, we imagine, if and only if the race was
won by a driver in a turbine-powered car, regardless of which driver
that may have been. Throughout his paper, MacKay uses the notion
of what the speaker was talking about. This is an extremely flexible
expression, with many uses.7 In our example, it seems natural to say
that the speaker was talking about the winner of the Indianapolis
500 race (and that the bet was a bet about the winner of this race).
But the expression X was talking about Y, as it finds application in
this example, is an intentional or referentially opaque context with
respect to Y. Suppose that Smith was a driver in the race. In making
the wager in question, could the speaker have been talking about
Smith? If Smith was not the winner of the race, then certainly there
seems to be no sense in which the speaker was talking about him
(and the bet was not a bet about Smith). Even if, at the time of the
bet, the speaker believed Smith to be the winning driver, he was not,
I think, talking about him.8 This is enough to show that in this use of
a definite description there is not the possibility that the speaker is
36

PUTTING HUMPTY DUMPTY TOGETHER AGAIN

talking about an entity that, in fact, fails to fit the description he

uses. It seems to me, moreover, that even if Smith was the winner of
the race, the speaker was not talking about him. From The speaker
was talking about the winner of the race and Smith was the winner
of the race, it does not follow, in this case, that the speaker was talking about Smith.
To state his view about reference and definite descriptions,
MacKay employed a notion of what the speaker is talking about
that is extensional and allows for the possibility that what the speaker
is talking about is an entity that fails to fit the description he used.
But this notion is applicable only to some uses of definite descriptions and these would have to be distinguished from those to which
it does not apply. Working out this distinction, however, would
simply give us another way of getting at the difference between the
referential and the attributive use of definite descriptions.
MacKay notes that in my paper many of the examples of a
speaker referring to something that does not fit the description used
are what he calls near misses. Could I have sustained my point, he
wonders, had I instead considered examples in which what the
speaker intended to refer to was very far from fitting the description?9 Before dealing with this directly, however, I want to comment
on another question raised by the subject of near misses.
In my paper I suggested that when a speaker uses a definite description, the , referentially to refer to something, e, he may have
said something true about e even if e is not or even if nothing is .
In contrast, I thought, if the definite description is used attributively, then if nothing is , the speaker cannot have said something
true (I left it open whether he would then have said something
false). The reason for this difference is that in the case of the
attributive use there is no entity e standing as the referent whether
or not it fits the description, and so there is nothing for the speaker
37

E S S AY S O N R E F E R E N C E

to have predicated something of when nothing satisfies the description. But this does not take account of near misses and so is, perhaps, too rigid in relation to our actual practice in such cases. In
one example of the attributive use in my paper, a person upon finding the body of his friend Smith exclaims, Smiths murderer is
insane. In the example, the speaker had no particular person in
mind as Smiths murderer. I then asked what the consequences
were if Smith in fact had not been murdered and I suggested that
the speaker could not have said something true. But suppose that
while Smith did die of natural causes, he had indeed been assaulted
before death and that the evidence that led the speaker to attribute
insanity to Smiths murderer is still good evidence that his assailant is insane. In a sense the speaker has scored a near miss. Are we
prepared to say flatly that the speaker did not state the truth? Since
some of my examples of referential uses of definite descriptions
also involved in some sense near misses, this sort of consideration
has led several people to wonder whether there is any difference
between the two uses along these lines.
But if I ignored the possibility of near misses in connection with
what I called attributive uses, the distinction is not really harmed.
For there are two sources of near misses, one for the attributive use
and another for the referential. A near miss occurs with an attributive use when nothing exactly fits the description used, but some
individual or other does fit a description in some sense close in
meaning to the one used. It is a quite different sort of near miss,
however, that is recognized by seeing that the particular individual
the speaker wanted to refer to has been described in a slightly inaccurate way. In the one case a near miss is scored, we might say, when
the speaker just misses some target or other; in the other when he
just misses the target he aimed at. Only in the referential use can a
speaker have missed by a mile, because only that use involves
38

PUTTING HUMPTY DUMPTY TOGETHER AGAIN

a particular entity that the description either fits neatly, just misses,
or misses wildly. Once this is seen, taking near misses into account
does not blur the distinction. If anything, it helps one to see what
the distinction is.
MacKay holds that when a speaker grossly misdescribes what he
wants to talk about, he does not succeed in referring to it. He admits,
in a footnote, that this may seem to be no more than a verbal disputewhat he calls what the speaker is talking about I call what
the speaker is referring to. If, as I have argued above, his view of the
matter yields the distinction in uses of definite descriptions I wanted
to draw, why is it not a verbal issue? It is not a question of the ordinary use of the verb to refer. (I think, in fact, considerations of that
sort may be more on my side than his.) Neither of us seems to be
analyzing the ordinary use of the term. What then are the unacceptable consequences of adopting my way of talking about such cases?
MacKay imagines someone intending to refer to a book saying,
Bring me the rock on the table. On my view this should be a perfectly clear case of referring to a book. But, he argues, If one can
refer to a book by using the rock, then one can refer to a book by
using any o.r.e. [ostensible referring expression], and so the actual
o.r.e. used becomes irrelevant. MacKays conclusion does not immediately follow. It does not follow from the fact that we can imagine some circumstances in which a speaker refers to a book by using
the definite description, the rock, that we can imagine this happening in any circumstances. But the latter would be required to
make the particular definite description used irrelevant. From what
I said in my article it is only in some circumstances that we can
imagine a speaker referring to a book by using the definite description, the rocknamely, when he really does intend to refer to
something which is in fact a book, by using the definite description, the rock. This may seem hardly any limitation at all, but that,
39

E S S AY S O N R E F E R E N C E

I believe, is the interesting mistake in MacKays criticism. I believe

the route MacKays reasoning takes is the following. In the speech
acts we are considering, it is the speakers intention to refer to
something. If he can accomplish this when he uses a definite description for that purpose regardless of the content of the description, then the content cannot matter to himthe actual o.r.e. used
becomes irrelevant. He might as well use one description as another so long as he uses them with the right intention. And this
result is absurd. This is where the analogy with Humpty Dumpty
enters. Humpty Dumpty thinks that he invests a word with meaning simply by using it with the intention that it shall mean that. But
then it seems that it should be irrelevant to him what word he uses.
I think this line of reasoning contains a mistake. No such consequence really follows from my way of talking about referring or
even from Humpty Dumptys theory of meaning, whatever other
defects there may be in that. Because it has some independent interest, I will begin with Humpty Dumptys theory.
Glory does not mean a nice knockdown argument in the language Humpty Dumpty speaks. Does it follow that glory could
not have meant that in the conversation with Alice? Or that Humpty
Dumpty did not mean that by glory? It is not my purpose here to
defend a theory of meaning based on the intentions of speakers. My
limited purpose is to suggest that no disastrous consequences follow
from the position that what a speaker refers to in a referential use of
a definite description is determined by his intentions. (I hope it is
obvious that adopting this viewpoint on reference does not commit
one to any such theory of meaning.) Nevertheless, the mistake I believe Mr. MacKay commits in his criticism of what I had to say about
reference seems one that it would be easy to commit in thinking
about Humpty Dumptys conversation. If so, it may be of interest to
look at it in that connection. In outline, what I want to show is that
40

PUTTING HUMPTY DUMPTY TOGETHER AGAIN

there is an explanation of the absurdity we find in the exchange with

Alice that is compatible with theories of meaning based on the intentions of speakers. Similarly, the inability to respond to Wittgensteins challenge to Say Its cold here and mean Its hot here can
be explained without subscribing to a particular view such as that
meaning can arise only in the context of an established practice.
Finally, I will try to show that the availability of this sort of explanation means that speaking of reference in the way I did does not have
the consequences MacKay believes it does.
Humpty Dumpty believes himself to be master of the meaning
of his words in that if he intends a word to mean such and such then
it will mean that when he uses it. MacKay, in making use of what he
considers an absurd theory of meaning for the purpose of comparing it to what I said about reference, seems to find no difficulty in the
possibility of Humpty Dumpty forming the intention to mean by a
word what it does not standardly mean, while at the same time
having no reason at all to suppose that his audience, Alice, will understand him. Forming the intention to mean a nice knockdown
argument by the word glory in such circumstances may seem as
easy as falling off a wall. Similarly, MacKay finds nothing odd about
the possibility that a person should simply intend to refer to a book
by the words the rock. What goes wrong, he seems to think, is that
the intention cannot be fulfilled. After all, one might say, what could
be easier than forming an intention? And if intentions were sufficient, then a speaker could mean anything by any word at any time
or refer to anything using any definite description at any time. Or so
it may seem.
The fact about intentions that I want to stress is that they are essentially connected with expectations.10 Ask someone to flap his arms
with the intention of flying. In response he can certainly wave his
arms up and down, just as one can easily on command say the words
41

E S S AY S O N R E F E R E N C E

Its cold here. But this is not to do it with the intention of flying. Nor
does it seem to me that a normal adult in normal circumstances can
flap his arms and in doing so really have that intention. Perhaps one
can, by a stretch of the imagination, conceive of someone (a child,
say, who has seen birds flying) doing this. But such a personthe
child, for examplewould have expectations not shared with us.
Similarly, one cannot say entirely out of the blue, Its cold here and
mean Its hot here, but not, I think, because whatever ones intentions the words will not get invested with that meaning. Rather, we
can explain this by the impossibility of having the right intention in
such circumstances. To the next person who comes in the room I say,
Its cold here. I have no expectations, any more than Humpty
Dumpty did about Alice, that the person will construe my words in a
novel way. Could I really intend that cold should mean hot? Or
would my performance not be so much arm-flapping?
In the analysis of meaning given by Grice,11 a speaker means
something by an utterance when he has a certain complex kind of
intention involving recognition on the part of his audience of his
intention. And what the speaker means is determined by the content of that intention. Whether he can form that intention, however,
may depend upon what expectations he has about his audience and
their ability to grasp his intention. It does not follow, then, from this
analysis that speakers might, out of the blue, mean anything at all by
any utterance. And the existence of an established practice may be
usually required for speakers to have the right expectations.
Had Humpty Dumpty prefaced his comment to Alice with an
explicit stipulation of what he would mean by the word glory, the
episode could not have been used by Lewis Carroll to raise a problem about meaning. By stating that he would mean a nice knockdown argument by glory, Humpty Dumpty could expect Alice to
understand him correctly when he said, Theres glory for you.
42

PUTTING HUMPTY DUMPTY TOGETHER AGAIN

I think the episode would also lose its interest if there were in the
situation something implicit to make it possible for Humpty
Dumpty to expect Alice to grasp his intention. If I were to end this
reply to MacKay with the sentence Theres glory for you I would
be guilty of arrogance and, no doubt, of overestimating the strength
of what I have said, but given the background I do not think I could
be accused of saying something unintelligible. I would be understood, and would I not have meant by glory a nice knockdown
argument? Would not glory in that last sentence mean that? It is
at least not obvious that it would not. But, for all I know, no one
(except possibly Humpty Dumpty) has ever meant that by glory,
and there is not an established practice or convention backing it up.
What is strange about Humpty Dumptys conversation may not be
so much the theory of meaning that he seems to have, but rather his
wanting us to believe that, without any assumption that Alice might
understand him, he really did have that intention about the word
glory. I cannot credit him with that intention any more than I could
credit a seemingly rational adult with the intention to fly when I see
him flapping his arms up and down.
Returning now to reference, given the way I talked about the referential use of definite descriptions, one can imagine circumstances in which
someone refers to a book by using the words the rock. But it does not
follow that, for example, I can now refer to a book by saying to the next
person to come into the room, Please bring me the rock. To think that
it does involves the view just now discussed, that there is no difficulty in
forming any intention whatever. The reason I cannot say that to the next
person I see and refer to a book is the same as the reason I cannot now say
to that person, Its cold here and mean Its hot here. I do not have the
right expectations about my audience. In the same way I cannot now flap
my arms intending to fly. What we have to imagine are situations in which
a person does really intend to refer to a book in saying the rock, and
43

E S S AY S O N R E F E R E N C E

these are not so easy to come by. Most of the examples in my paper of
speakers referring to things that do not fit the description used had the
speaker believing that what he wanted to refer to did fit the description.
Thus, if by a trick of light what is a book looks to me like a rock, I am inclined to say that I would have referred to a book in saying Bring me the
rock. I would, of course, have reason to expect this description to enable
my audience to pick out what I was referring to (even though in this case
I may be wrong about it). As an extreme case I also thought that one
could refer to something knowing that it did not fit the description used,
but again when one had certain expectations about the audience. To
change the example slightly, suppose there is a rock on my shelf that has
been carefully carved to resemble a book and that I know the person I am
speaking to cannot recognize it for what it is. I say to him, Bring me the
book with the blue binding. It seems at least plausible to say that I referred to a rock. One might put the matter this way. Saying the book
while merely thinking the rock does not constitute referring to the
rock, but neither does it constitute having the intention to refer to the
rock. What MacKay has ignored is the fact that the intention to refer to
something in using a definite description is a complex intention involving
expectations regarding ones audience. When a speaker uses a definite
description referentially he intends his audience to take the description
as characterizing what it is he wants to talk about. In so doing he hopes
that they will successfully recognize what that is. The content of the description is clearly relevant to this intention and not something the
speaker can ignore. Now if we choose to characterize the situation in
which the speaker genuinely has such an intention with respect to an
object, his audience, and a description, but where the description chosen
fails to fit the object, as a situation in which the speaker referred to the
object, no dire consequences ensue.12
The role of the content of a definite description used to refer
comes out in another way. Suppose that two friends have a private
44

PUTTING HUMPTY DUMPTY TOGETHER AGAIN

joke about a certain book; reading it is such heavy going that they
call it the rock. One of them might say to the other, Bring me the
rock and in doing so refer to that book. But this would not be a definite description used referentially. The expression the rock is being
used rather like a name. The standard meaning of rock has no more
than historical connection with its use here in referring to something. The two friends might even have forgotten how it came about
that they call this book the rock. In the same way it may be possible to imagine some private joke or other background that makes it
suitable for a speaker using the phrase the square root of two to
refer to a book. But it seems to me impossible to imagine a speaker
to use this phrase in its standard meaning as a definite description
used referentially. The reason is that I cannot conceive of the set of
beliefs that would allow a speaker to suppose that this description
was a means to fulfilling the intention of making it known to his
audience what he is talking about via the content of the description.
I can imagine a speaker mistaking a book for a rock or believing that
his audience will, but I cannot imagine someone mistaking a book
for the square root of two.
I have hoped to make two points in reply to MacKay. The first is
that the distinction between two uses of definite descriptions can be
drawn as easily in the terminology MacKay prefers as in that I used.
Secondly, the dire consequences of adopting my way of talking
about referring that MacKay foresees really do not materialize.

NOTES
1. I am indebted to Professor David Sachs for reading an earlier draft and making
several valuable suggestions for improvement.
2. Alfred F. MacKay, Mr. Donnellan and Humpty Dumpty on Referring, Philosophical Review, this issue, pp. 197202.

45

E S S AY S O N R E F E R E N C E

3. Reference and Definite Descriptions, Philosophical Review, LXXV (1966),

281304.
4. Philosophical Review, LXVI (1957), 37788.
5. Strawson thought Russells theory of definite descriptions would have difficulty
accounting for this fact, since on Russells view a statement such as The table is
covered with books would be literally false so long as there is more than one
table in the world. Without considering the two uses of definite descriptions, the
reply that one is inclined to make on Russells behalf is that in the loose way of
everyday speech the context is relied upon to supply further qualifications on the
description to make it unique. This seems a plausible reply when considering attributive uses. Someone says, The next President will be a dove on Vietnam, and
context easily supplies the implicit of the United States. But where one has a
very indefinite definite description, with many things in the world satisfying the
actual description, the reply is not so plausible. These are commonly, I believe,
referential uses. A speaker wants to refer to some object and uses an indefinite
definite description. Asked to make his description more precise, he may have to
think about how best to do it. Several further descriptions may come to mind, not
all of which are actually correct. Which, then, shall we say is the full but implicit
one? Once we see the function of a referential description, however, we need not
suppose that there is any one description recoverable from the speech act that is
supposed uniquely to apply to the object referred to. The audience may through
the partial description and various clues and cues know to what the speaker refers
without being in possession of a description that uniquely fits it and which was
implicit all along in the speakers speech act. We can, I would suggest, view Russells theory as a theory of attributive definite descriptions against which examples
drawn from referential uses are not relevant.
6. This last feature is not independent of the first three. It is implicit in my use of e.
It should be noted that I allow here (and strictly should throughout) for the possibility of a referential use without a referent. This would occur when no entity
can be correctly identified as what the speaker meant to be talking about,
although the speaker intended that there should be. This failure of reference will
not come about simply because nothing is or because more than one thing is
and, moreover, could happen when there is a unique . Thus it lends no support
to MacKays view. Exactly when it will happen could be specified only after an
analysis of the criteria for deciding what entity it is to which a speaker is referring,
an analysis I have not tried to give here or in the original paper.
7. For this reason it would be rather unhelpful to try to give an analysis of referring
utilizing this expression. For example, in one use of the expression, a speaker can
be talking about something although he uses no expression to refer to it. (The
Secretary of State was talking about Vietnam when he said, We must honor our
commitment.) The real job of analysis would come when we tried to specify the
sense of what the speaker was talking about we meant.

46

PUTTING HUMPTY DUMPTY TOGETHER AGAIN

8. The speaker could lose his bet even if Smith drove a turbine-powered car and he
could win it even if Smith did not.
9. Not all of my examples could be considered near missese.g., the example of
misdescribing Jones as Smiths murderer (p. 286) when he is innocent, or of
referring to a rock as the man carrying a walking stick (p. 296).
10. The exact connection is probably complicated and I have, I think, oversimplified
it in what follows. What we can do with a certain intention not only depends
upon expectations, but also upon the possibility of other means of accomplishing the same end and upon incentives. A man in the water from a sinking ship
might move his arms with the intention of swimming a hundred miles to shore,
if that is the only hope, even though he has no rational expectation of doing it.
But is it open to an ordinary man at the beach to strike out with that intention?
11. In Meaning, op. cit.
12. The situation might be summed up through an analogy. Imagine the following
game of describing. A player chooses some object and attempts to describe it.
His purpose is to get a description that allows the other players to recognize the
object he has chosen. Sometimes, however, a players description will fail to characterize the object correctly. Given enough imagination we can even envisage
the situation in which the object is a book and the description is a rock. When
a player grossly misses in this way, should we say that he has not described the
object? Or should we say that he has described it, but incorrectly? This seems to
be a verbal issue. If we opt for the second way, however, we will not be forced to
say that the description a player uses is irrelevant just because even when he
chooses the wrong one, we still say that he has described the object. (We can also
imagine a game analogous to the attributive use of definite descriptions. Here a
player does not first choose an object to describe. Rather he gives a description
with the aim of giving one that uniquely fits some object or other.)

47

This page intentionally left blank

[3]
PROPER NAMES AND IDENTIFYING
DESCRIPTIONS

I
There is an extremely plausible principle about proper names that
many philosophers up to the present have either assumed or argued
for.* I will call it the principle of identifying descriptions. One
illustration of it is in this passage from Strawsons Individuals:
. . . it is no good using a name for a particular unless one knows
who or what is referred to by the use of the name. A name is
worthless without a backing of descriptions which can be produced on demand to explain the application.1

The backing of descriptions Strawson speaks of supposedly functions as the criterion for identifying the referent of a name, if it has
one, or, alternatively, for deciding that there is no referent. If I say,
for example, Homer is my favorite poet, then, roughly speaking,
the descriptions I could supply in answer to the question, Who is
First published in Synthese, 1970, 21: 33558.

49

E S S AY S O N R E F E R E N C E

Homer? provide the backing of descriptions. And these in turn

either pick out a single individual as the referent of the name (as it
occurs in my utterance) in virtue of his fitting these descriptions or
make it true that there is no referentthat Homer did not exist.
While this initial statement of the principle needs refinement
and the acknowledgement of variants, it seems at first sight almost
indisputable that some such principle governs the referential function of proper names. Must not a user of a proper name know to
whom or what he is referring? And what can this knowledge consist
in if not the ability to describe the referent uniquely?
Nevertheless, I believe the principle to be false. In the first sections
of the paper I will state the principle more precisely and fill in some of
the details of how it would have to operate. The exercise of trying to
make it more precise and giving various needed qualifications is
enough, I think, to rob it of some of its initial attractiveness. I will then,
however, meet it head-on by means of counter-examples. I will argue
that (a) a proper name may have a referent even though the conditions
laid down by the principle are not satisfied and (b) where the conditions are satisfied, the object that ought to be the referent according to
the principle need not be the true referent. In the course of this I will
suggest certain positive things about how the referent of a name is determined, though these will not amount to an alternative principle.

II
What I call the principle of identifying descriptions should not be
thought of as expressing the thesis that proper names have a sense
(or meaning or connotation). (That thesis, I think, suffers in any
case from vagueness about what is to count as showing that an expression has a sense.) Anyone who holds that proper names have a
sense almost certainly subscribes to the principle, but the converse
50

PROPER NAMES AND IDENTIFYING DESCRIPTIONS

is doubtful. In his influential paper, Proper Names,2 John Searle

begins with the question, Do proper names have senses? and he
ends by saying that in a sense they do and in a sense they do not.
Searle, however, though he would not without heavy qualification
ascribe senses to proper names, is one of the prime examples of a
philosopher who defends the principle I have in mind. In this he is
in company with Frege who would have no reluctance in talking
about the sense of a proper name.
The simplest application of the principle, to be sure, can be found
in the view of someone such as Russell who holds that proper names
are concealed definite descriptions. Russell says, . . . the name Romulus is not really a name [that is, in the narrow logical sense] but a sort
of truncated description. It stands for a person who did such-andsuch things, who killed Remus, and founded Rome, and so on.3 And
again, When I say, e.g., Homer existed, I am meaning by Homer
some description, say the author of the Homeric Poems . . .4 Russell
associates with the use of a name some definite description for which
the name is a simple substitutethe same proposition would be expressed by a sentence containing the name as by the sentence formed
from it by substituting the associated description for the name.
This tight connection between proper names and definite descriptions was rightly challenged by Searle in Proper Names. Yet
Searle still retains the backing of descriptions and these serve, as
they would also for Russell, as criteria for identifying the referent,
albeit in a looser and more complicated manner:
Suppose we ask the users of the name Aristotle to state what
they regard as certain essential and established facts about him.
Their answers would be a set of uniquely referring descriptive
statements. Now what I am arguing is that the descriptive force
of This is Aristotle is to assert that a sufficient but so far unspecified number of these statements are true of this object.5
51

E S S AY S O N R E F E R E N C E

Without doubt this departs significantly from Russells simplistic

view. It allows for (what surely we should allow for) the possibility,
for example, of discovering that Aristotle was not the teacher of Alexander the Great without having to deny Aristotles existence,
which would be impossible on Russells view if that description was
part of the associated description for our use of Aristotle. Only a
sufficient number of the things we believe about Aristotle need be
true of some individual for him to be Aristotle.
But the flexibility introduced is limited. Vague and indeterminate as we may leave the notion of sufficient number, behind our
use of a name a set of descriptions still operates to determine the
referent. The formulation of the principle of identifying descriptions I shall give will allow both for Searles looser and Russells
tighter connection between names and descriptions.
I should like to make one more general comment about the
issue I am concerned with. The importance of the principle in
question is not confined to a narrow issue about how proper names
refer. It also has a bearing on the general problem of reference. For
proper names constitute something like a test case for theories of
reference. A peculiar feature of the situation is that two classical
but opposing paradigms for referring expressions can both lead
one to adopt the same theory about proper names. The model
referring expression has been for many philosophers of language, I believe, a definite description (used attributively in the
terminology I used elsewhere).6 An object is referred to in virtue
of possessing uniquely the properties mentioned in the definite
description. It is not hard to see how this standard leads to adopting the principle of identifying descriptions for proper names.
Proper names are referring expressions, yet on the surface fail to
exhibit any descriptive content. Given definite descriptions as the
paradigm, one is forced to look under the surface (which amounts
52

PROPER NAMES AND IDENTIFYING DESCRIPTIONS

to looking into the user(s) of the name) for the backing of descriptions that must be there.
The major alternative to a definite description as the paradigm
of a referring expression is represented by Russells and Wittgensteins (in the Tractatus) notion of a name in the narrow logical
sense. Ordinary names, of course, are not names at all in this sense;
they cannot meet the austere requirements of referring in some
mysterious, unanalyzable and absolutely direct way to their referents. And given this notion of genuine names, Russell adduces very
good reasons why no such ordinary name as Homer or Aristotle
can be a genuine name. But some account has to be given of how
ordinary names function. Russell saw no alternative but to treat
them as concealed definite descriptions, what they name, if anything, being whatever is denoted by the concealed description.
(Had he thought of Searles perhaps more sophisticated view, there
seems no reason why he should not have adopted that for ordinary
proper names.)
Strangely enough, then, two antagonistic models of what a genuine referring expression is like lead their proponents to the principle of identifying descriptions. Demonstrating that that principle is
mistaken would not irrevocably discredit either model, but it would,
I think, take away much of the motivation for adopting either. Ordinary proper names may not have as much claim to being genuine
referring expressions as Russells names in the strict logical sense
(could we but understand what those are and discover some of
them), but as against definite descriptions it is hard to see how they
could come out second best. If their mode of functioning, however,
is not captured by the principle of identifying descriptions, if, that
is, they do not name in much the same way a definite description
denotes,7 then can definite descriptions possibly be model referring
expressions?
53

E S S AY S O N R E F E R E N C E

And on the other side, if ordinary proper names are neither

names in the strict logical sense, as they surely are not, nor concealed descriptions, then some other relationship will have to be
recognized as holding between some singular expressions and what
they stand for. In that case, much of the reason for supposing that
there are such things as names in the strict logical sense will be
gone. For it is clear from Russells writings, at least, that these are
introduced in part because he felt that definite descriptions not
really being referring expressions (but only denoting expressions),
some other sort of expression must serve the purpose of allowing us
to talk directly about things in the world. If (ordinary) proper names
do not function via the relationship of denoting nor through whatever relationship Russells names are supposed to enjoy, then perhaps the way they do function represents the alternative Russell was
seeking.8

III
The principle of identifying descriptions is a two-stage thesis, the
second stage depending upon the first. It states, in the first place,
that (with some qualifications to be noted later) the user(s) of a
proper name must be able to supply a set of, as I shall call them,
non-question-begging descriptions in answer to the question, To
whom (or what) does the name refer? The important qualifier,
non-question-begging, I will explain later.9 I will call these descriptions that speakers supposedly must be able to supply the set of
identifying descriptions.
Secondly, the principle states that the referent of a proper
name (as used by a speaker in some particular utterance), if there is
one, is that object that uniquely fits a sufficient number of the
54

PROPER NAMES AND IDENTIFYING DESCRIPTIONS

descriptions in the set of identifying descriptions. As a corollary,

when no entity (or more than one) satisfies this condition, the name
has no referent and a negative existential statement expressible by a
sentence of the form N does not exist (where N is the name in
question) will be true.
I have tried to state the principle so as to make it possible for alternative positions still to embody it. I should like to show that we
ought not to accept any of the versions of it to be found in the literature. Thus, for reasons that will emerge, I leave it open in the first part
whether the set of identifying descriptions is to be formed from what
each speaker can supply or from what speakers collectively supply. In
the second part, the sufficient number of descriptions that an object
must satisfy to be the referent might be all of them, as in Russells
view, or some indeterminate number of them, as in Searles.
The counter-examples I later give are directed against the second
part of the principle; they are designed to show that even if the
user(s) of a name must be able to supply a set of identifying descriptions, as laid down by the first part, these descriptions do not provide necessary and sufficient conditions for what shall count as the
referent. But the first part of the principle is not without difficulties.
To strengthen my case against the principle I want first to point out
some of these while formulating some of the needed qualifications
to the principle as I have just stated it.

IV
There are two views on the source of the set of identifying descriptions that supposedly must back up the use of a proper name.
We find in Russell and Frege10 the idea that different speakers
who use the same name in an otherwise identical propositional
55

E S S AY S O N R E F E R E N C E

context will most likely not express the same proposition (or
thought, in Freges terminology). This happens because very probably they do not associate with the name the same set of descriptions. The propositions might have different truth-values, because
the speakers, with different sets of identifying descriptions, may be
referring to different things.11 Russell and Frege, in other words,
look to the individual speaker for the set of identifying
descriptions.
In contrast, Searle tells us that the set of identifying descriptions
is formed from the descriptions users of the name give of what they
refer to. And Strawson, in discussing this question,12 imagines a situation in which a name is used by a group in which each member
knows some distinguishing fact or facts, not necessarily the same
ones, about Socrates, facts which each is prepared to cite to indicate
whom he now means or understands, by Socrates. He then suggests that we form a composite description incorporating the most
frequently mentioned facts and continues, Now it would be too
much to say that the success of term-introduction within the group
by means of the name requires that there should exist just one person
of whom all the propositions in the composite description are true.
But it would not be too much to say that it requires that there should
exist one and only one person of whom some reasonable proportion
of these propositions is true.13 Given this difference of opinion, I allowed for alternatives in the statement of the principle.
Both means of determining the set of identifying descriptions
contain difficulties. To take the Russell-Frege view first, it seems to
me, though evidently not to them, absurd to suppose that a beginning student of philosophy, who has learned a few things about
Aristotle, and his teacher, who knows a great deal, express different propositions when each says Aristotle was the teacher of
Alexander. Even if this can be swallowed, there are very unpleasant
56

PROPER NAMES AND IDENTIFYING DESCRIPTIONS

consequences. Given the second part of the principle of identifying

descriptions the student and teacher possess different criteria for
identifying Aristotle and even for establishing his existence. For the
student Aristotle would be a person satisfying (substantially) some
fairly small number of descriptions; for the scholar of philosophy a
much larger number would determine the existence and identity of
Aristotle. This means that if each affirms Aristotles existence there is
the theoretical possibility, at least, that one is correct and the other
wrong. Yet suppose that the smaller supply of descriptions available
to the student turns out generally to be incorrect (we can imagine
him to be unfortunate enough to have been told mostly things about
Aristotle that historians of Greek philosophy are mistaken about).
Would he really be in error in saying that Aristotle existed? Should
we say to him, if we uncover the errors, Your Aristotle doesnt exist,
though Professor Smiths does.?
Worse still, suppose that the few things the student has learned
about Aristotle are not only not true of the individual his teacher
refers to, but turn out substantially to be true of, say, Plato. He has
been told, perhaps, that Aristotle wrote the Metaphysics when, in
fact, Plato wrote it and Aristotle cribbed it, etc. Should we say that
he has all along been referring to Plato, though his teacher, for whom
these few descriptions are not the only source of criteria for what
the referent is, continues to refer to Aristotle? The principle of identifying descriptions seems to lead to that result when interpreted in
this way.
The more liberal view that utilizes descriptions suppliable by users
of the name, in the plural, is not in much better shape. In the first place,
what group of speakers is to form the reference set from which the
composite description is to be drawn? Searle speaks of properties
commonly attributed to Aristotle. Commonly attributed by whom?
By contemporary speakers? One thing seems certain: the speakers in
57

E S S AY S O N R E F E R E N C E

question cannot be all those who have ever used the name Aristotle
to refer to Aristotle. Aside from the appearance, at least, of circularity,
none of us would likely ever be in a position to know what properties
that group would attribute to Aristotle. Childhood friends of Aristotle,
who did not follow his subsequent career, would have a quite different
set of descriptions of him from ours. I doubt that we shall ever know
what those were. Using this total class of those who have ever spoken
of Aristotle is a practical impossibility and can hardly form the basis
for our use. (It would also seem to do violence to the motivation
behind the principle of identifying descriptionsthat users of a name
should be able to supply criteria for identifying the referent.)
On the other hand, to limit the group of speakers whose descriptions will generate the composite description to, say, those at a particular time yields consequences similar to those of the Russell-Frege
view. Different times and ages might have different beliefs about
Aristotle. And in conjunction with the second part of the principle of
identifying descriptions it would be possible that the affirmation that
Aristotle existed should have different truth-values from one time to
another. Or, because of the particular beliefs they held, we could
imagine that the people of one age, unknown to any of us, referred to
Plato when they used the name Aristotle. On the Frege-Russell view
any two people using the same sentence containing the name
Aristotle and believing that they are referring to the same person,
etc., very likely do not express the same proposition. The more liberal view only expands this possibility to different groups of people.

V
The first part of the principle of identifying descriptions tells us that
users of a name must be in a position to supply a set of identifying
58

PROPER NAMES AND IDENTIFYING DESCRIPTIONS

descriptions. (For the sake of argument I will at times allow that this
is so, although what positive remarks I make will imply that there is
no necessity involved.) How are we to understand this? Strawson
says, . . . When I speak of preparedness to substitute a description
for a name, this requirement must not be taken too literally. It is not
required that people be very ready articulators of what they know.14
I think he is surely right to allow us this latitude. Small children and
even adults often use names without literally being able to describe
the referent in sufficient detail to guarantee unique identification.
I imagine the reason philosophers who have discussed proper
names so often use historical figures such as Aristotle, Homer, etc. is
just that these names are introduced into our vocabulary via descriptions of facts about their bearers and most of us are prepared to
give something like uniquely denoting descriptions. But it is less
clear that we are ready to describe our friends, people we have met
here and there, or even public figures of our times whose images
have not yet been crystalized into a few memorable attributes. At
the very least it would be an effort to insure that a description of
someone we know fairly well and whose name we use often is both
accurate and unique. The first part of the principle, then, seems to
require of us a high level of abilityunless what counts as having
the ability is very broad indeed. (Even though it is hardly like being
able to describe the referent, the ability to point to the referent is
usually included as if it were simply a variant.)
Construe it as broadly as you will, is there really a requirement
that the user of a name be able to identify by description (or even by
pointing) what the name refers to? The following example, which
anticipates a bit some later results, may cast doubt on this. Suppose
a child is gotten up from sleep at a party and introduced to someone
as Tom, who then says a few words to the child. Later the child
says to his parents, Tom is a nice man. The only thing he can say
59

E S S AY S O N R E F E R E N C E

about Tom is that Tom was at a party. Moreover, he is unable to

recognize anyone as Tom on subsequent occasions. His parents
give lots of parties and they have numerous friends named Tom.
The case could be built up, I think, so that nothing the child possesses in the way of descriptions, dispositions to recognize, serves to
pick out in the standard way anybody uniquely. That is, we cannot
go by the denotation of his descriptions nor whom he points to, if
anyone, etc. Does this mean that there is no person to whom he was
referring? It seems to me that his parents might perfectly well conjecture about the matter and come up with a reasonable argument
showing that the child was talking about this person rather than
that. For example, they might reason as follows: Hes met several
people named Tom at recent parties, but only Tom Brown did
something that might make him say, Tom is a nice man. Of course,
Tom Brown isnt nice and he was just indulging in his usual sarcasm
when he told him, You have a nice pair of parents, but the sarcasm
wouldnt have registered.15
If this is a reasonable example, it seems the question of what a
speaker referred to by using a name is not foreclosed by his inability
to describe or even to recognize or point to the referent. The reasoning of the parents in this example is not aimed at finding out what
descriptions the child could give, if only he were able to articulate
them. I used a child in the example to sharpen the picture of someone with no descriptions or other means of identifying the referent
uniquely; but adults also sometimes conjecture about other adults
concerning what person they were referring to in using a name. Is it
beyond doubt that in such instances the inquiry must ultimately be
concerned with what descriptions the user of the name could
supply? The examples later on will challenge this, yet even now examples such as the one I have given seem to me to make the requirement that every use of a name have behind it a backing of descriptions
60

PROPER NAMES AND IDENTIFYING DESCRIPTIONS

highly suspicious (even without relying on what appears to me

beyond question, that no one has yet given a clear account of what
the ability to describe a referent amounts to).

VI
Before turning to counter-examples one more preliminary issue
should be settled. In stating the principle of identifying descriptions, I inserted the condition that the descriptions that back up
the use of a name should not be question-begging. The qualification has vital significance because there are certain descriptions
that a user of a name (providing he can articulate them) could
always provide and which would always denote the referent of the
name uniquely (providing there is one). No argument could be devised to show that the referent of a name need not be denoted by
these descriptions. At the same time anyone who subscribes to the
principle of identifying descriptions would hardly have these descriptions in mind or want to rely on them in defense of the principle. Some examples of what I shall count as question-begging
are the following:
(a) the entity I had in mind
(b) the entity I referred to
(c) the entity I believe to be the author of the Metaphysics.

I think it is clear about (a) and (b) and only a little less so about (c)
that if descriptions such as these are included in the backing of descriptions the principle would become uninteresting.
Strawson, in fact, explicitly excludes descriptions such as (a):
[the speaker] cannot, for himself, distinguish the particular which
61

E S S AY S O N R E F E R E N C E

he has in mind by the fact that it is the one he has in mind. So,
there must be some description he could give, which need not be
the description he does give, which applies uniquely to the one he
has in mind and does not include the phrase, the one he has in
mind. 16 Although Strawson mentions a particular description, it
is certain that he would exclude from consideration similar ones.
In particular, (b) above surely would not count for him. The point
of the backing of descriptions is to explain how an object gets
referred to by a proper name. Descriptions that fit the referent
simply in virtue of the fact that the speaker did, in fact, refer to it
or had it in mind as the object he meant to refer to are questionbegging in answer to the question, who (or what) did you refer
to? in the same way that What I have in my hand would be question-begging in answer to the question, What are you holding in
your hand?
It is only a little bit less obvious that descriptions of the form,
the object I believe to be , such as (c) above, must likewise be excluded from the set of identifying descriptions.
Call descriptions such as the author of the Metaphysics primary descriptions; call those such as the man I believed to be the
author of the Metaphysics secondary descriptions. Suppose that all
primary descriptions the user(s) of a name can supply are false of
everything. The backing of secondary descriptions would be useless in the same way that the object I had in mind would be. For if
I cannot rely on my primary descriptions to pick out uniquely
what I refer to, trying to identify the referent via a description of
the form the one I believed to be (though it is not) would
amount to no more than trying to identify the object I had in mind
when I held that belief.
In what follows, then, I will count what I have called secondary
descriptions as question-begging.
62

PROPER NAMES AND IDENTIFYING DESCRIPTIONS

VII
In the next sections I construct counter-examples to the principle of
identifying descriptions. To do this I must show that there are possible situations in which the referent of a name does not satisfy the
conditions the principle lays down or situations in which an entity
satisfying those conditions is not the referent. The principle tells us
that the referent of a name, if there is one, is that entity that fits some
sufficient number of a certain set of descriptions, namely the set
suppliable by the user(s) of the name. It is important to note that in
denying this, one need not deny that there are some constraints on
what the referent of a name may besome description which it
must fit. But this is only to allow that there may be a backing of descriptions that serve as necessary conditions, while the principle
tells us that such a backing of descriptions also serves as sufficient
conditions.
Thus, I should want to argue, for example, that theoretically Aristotle might turn out to be a person who did not write the Metaphysics, was not the teacher of Alexander, etc.; that is to say, a person who
does fit a sufficient number of the descriptions we, as users of the
name, would now supply. But I need not argue that even theoretically he could turn out to be, say, a fishmonger living in Hoboken or
Platos dog (although in incautious moments I am inclined to believe in even this outlandish theoretical possibility). If anyone wants
to maintain that our use of the name is such that being a human
being or not living in modern times, etc. are necessary for being the
referent of the name, I have no objection here to offer against a
backing of descriptions in that weaker sense. Such an attenuated
backing would not uniquely identify the referent.
A word about the nature of the counter-examples is required,
because they will undoubtedly seem artificial and possibly taken
63

E S S AY S O N R E F E R E N C E

on their own not wholly convincing. Their artificiality is in part

forced on me by the fact that I want to question not only the
simple view of, say, Russell that sees a name as a simple substitute for a description, but also the looser and vaguer view of
Searle and Strawson. The latter, however, uses the notion of an
ill-defined sufficient number of descriptions. Since the notion
of sufficient is ill-defined, it is necessary to invent examples in
which, for instance, the referent of a name fits no description
which is both unique to it and available to the speaker (other
than question-begging descriptions). Otherwise, a defender of
the view might take refuge in those descriptions. To make sure
that there are no remaining contaminating descriptions, the examples have to be fairly extreme ones in which the user(s) of a
name are radically deceived about the properties of what they
are talking about.
But if these pure examples are in order in everything except
their artificiality, then the fact that I do not tell more true-to-life
stories should not be an objection. For however vague sufficient
number is left, one thing is certain: the Searle-Strawson view
cannot be that the referent of a name is any entity that fits uniquely
any one of the descriptions suppliable by the user(s) of the name.
The whole purpose of this variant (as opposed to the stronger
Russell view) is to allow that we could discover, e.g., that Aristotle
did not teach Alexander without having to deny Aristotles existence or that someone else was the teacher of Alexander. But if any
one of the descriptions in the set of identifying descriptions counts
always as sufficient, there will be an overwhelming number of
cases in which there cannot be a unique referent for a name we
useall those instances in which we ascribe to the referent two or
more properties which in fact are unique properties of more than
one person.
64

PROPER NAMES AND IDENTIFYING DESCRIPTIONS

VIII
The first counter-example is the most artificial (but perhaps the
most pure). It is a situation in which a speaker uses a name to refer
to something though what is referred to is not picked out uniquely
by the descriptions available to the speaker. As well, there is something the speakers descriptions denote uniquely, but that is not the
referent.
Imagine the following circumstances: Perhaps in an experiment
by psychologists interested in perception a subject is seated before a
screen of uniform color and large enough to entirely fill his visual
field. On the screen are painted two squares of identical size and
color, one directly above the other. The subject knows nothing of
the history of the squareswhether one was painted before the
other, etc. Nor does he know anything about their future. He is
asked to give names to the squares17 and to say on what basis he assigns the names. With one complication to be noted later, it seems
that the only way in which he can distinguish the squares through
description is by their relative positions. So he might respond that
he will call the top square alpha and the bottom square beta.
The catch in the example is this: unknown to the subject, he has
been fitted with spectacles that invert his visual field. Thus, the
square he sees as apparently on top is really on the bottom and vice
versa. Having now two names to work with we can imagine the subject using one of them to say something about one of the squares.
Suppose he comes to believe (whether erroneously or not doesnt
matter) that one of the squares has changed color. He might report,
Alpha is now a different color. But which square is he referring to?
He would describe alpha as the square on top. And if this is the only
uniquely identifying description at his command then according to
the principle I am attacking, he would have referred to the square
65

E S S AY S O N R E F E R E N C E

that is on top. But given our knowledge of the presence and effect of
the inverting spectacles and the ignorance of the subject about that,
it seems clear that we should take him as referring to, not the square
on top, but the one that seems to him erroneously to be on topthe
one on the bottom. We know why he describes alpha the way he
does; we expect changes in the square on the bottom to elicit from
him reports of changes in alpha, etc. I think it would be altogether
right to say that although he does not know it, he is talking about the
square on the bottom even though he would describe it as the square
on top. If this is right, we seem to have a case in which the speakers
descriptions of what he is referring to when he uses a name do not
yield the true referent so long as we stick to what is denoted by the
descriptions he gives. The referent is something different and the
thing actually denoted is not the referent.
This counter-example to the principle of identifying descriptions depends upon the supposition that the subjects only description that could serve to pick out the referent uniquely is the one in
terms of relative position. But it must be admitted that I have so far
neglected a description of alpha that he could supply, that is not
question-begging, and that would in fact uniquely identify alpha despite the operation of the glasses. The subject could describe alpha
as, the square that appears to me to be on top. We must take appears here in its phenomenological sense. If that appears to me to
be on top means that I believe to be on top we would have a question-begging description. But in its phenomenological sense, alpha
is the one that appears to him to be on top and, indeed, it is just because the square on the bottom is the one that appears to him to be
on top that it is the referent of alpha.
There is more than one way to modify the example in order to
take care of this objection to it, but an easy way is by having the
subject use the name alpha a bit later having forgotten how alpha
66

PROPER NAMES AND IDENTIFYING DESCRIPTIONS

appeared to him, but recalling the position he took it really to have.

Of course in our example as presented the subject would have no
reason to suppose that there might be a discrepancy between the
actual position of alpha and what position it appeared to him to
have and so long as he remembered it as being the one on top, he
would presumably say that that was also the way it appeared to him.
What is needed is something to make him doubt that his recollection of what position he took alpha to have is an accurate guide to
how it appeared to him.
Suppose then that our subject is an old hand at experiments of
this sort and knows that inverting lenses are sometimes put into the
spectacles he wears. Erroneously he believes he has a method of detecting when this happens. He goes through the experiment as previously described but with the mistaken belief that his spectacles
have not been tampered with and that the squares have the position
they appear to him to have. Later on he makes some statement such
as, Alpha changed color at one point. But while he remembers his
judgment that alpha was the top square (and has absolute confidence in it), he cannot remember how alpha appeared to him at the
time nor whether he had based his judgment on the assumption
that his visual field was inverted or not. The subjects set of identifying descriptions thus no longer contains the appearance description
and only the erroneous description of alpha as being the square on
top remains as a uniquely identifying description.

IX
If the preceding counter-example was persuasive, then it will also
suggest something positive. Its moral might be put this way: When
a person describes something, as when he describes what he is
67

E S S AY S O N R E F E R E N C E

referring to, we are not limited to looking for something that fits his
descriptions uniquely (or fits them better than anything else). We
can also ask ourselves, What thing would he judge to fit those descriptions, even if it does not really do so? That question will utilize
his descriptions, but will not be decided on the rigid basis of what is
denoted, if anything, uniquely by them. In this particular example
the influence of inverting spectacles was a deciding factor. We had to
know both how he described the referent and, what he did not know,
that the spectacles would influence his descriptions in a certain way.
The role of his set of identifying descriptions in determining the
referent of his use of a name is not that which the principle of identifying descriptions gives it. It had its part, but the question asked
about it was different: What do these descriptions denote uniquely
(or best)? vs. Why should he describe the referent in that way?
The next counter-example18 provides a somewhat different insight into how proper names function.
A student meets a man he takes to be the famous philosopher,
J. L. Aston-Martin. Previously, the student has read some of the philosophers works and so has at his command descriptions such as,
the author of Other Bodies and the leading expounder of the
theory of egocentric pluralism. The meeting takes place at a party
and the student engages the man in a somewhat lengthy conversation, much of it given over, it turns out, to trying to name cities over
100,000 in population in descending order of altitude above sea
level. In fact, however, although the student never suspects it, the
man at the party is not the famous philosopher, but someone who
leads the student to have that impression. (We can even imagine
that by coincidence he has the same name.)
Imagine, then, a subsequent conversation with his friends in
which the student relates what happened at the party. He might
begin by saying, Last night I met J. L. Aston-Martin and talked to
68

PROPER NAMES AND IDENTIFYING DESCRIPTIONS

him for almost an hour. To whom does he refer at this point?

I strongly believe the answer should be, to the famous philosopher,
and not, to the man he met at the party. What the student says is
simply false; a friend in the know would be justified in replying
that he did not meet J. L. Aston-Martin, but someone who had the
same name and was no more a philosopher than Milton Berle.
Suppose, however, that the audience contains no such doubting
Thomases, and that the rest of the party was of sufficient interest to
generate several more stories about what went on. The student
might use the name J. L. Aston-Martin, as it were, incidently. For
example: . . . and then Robinson tripped over Aston-Martins feet
and fell flat on his face or I was almost the last to leaveonly Aston-Martin and Robinson, who was still out cold, were left.
In these subsequent utterances to whom was the speaker referring in using the name, Aston-Martin? My inclination is to say that
here it was to the man he met at the party and not to the famous
philosopher. Perhaps the difference lies in the fact that in the initial
utterance the speakers remark would only have a point if he was referring to the famous philosopher, while in the later utterances it is
more natural to take him to be referring to the man at the party,
since what happened there is the whole point.19
If in such examples as this there are two references made (or even
if there is a strong inclination to say that there are) this is something
unaccounted for by the principle of identifying descriptions.
To see this we need only ask what the students set of identifying
descriptions consists in each time he uses the name, first when he
claims to have met Aston-Martin and later when he recounts events
at the party that incidently involve the man he met there. In both
cases the set of identifying descriptions would be the same. It
will include, first of all, those descriptions of Aston-Martin he
would have given prior to the partythe author of certain works,
69

E S S AY S O N R E F E R E N C E

propounder of certain doctrines, etc. In addition, it would now contain various descriptions derived from meeting the spurious famous
man at the partythe man who played the game about cities, whose
feet Robinson tripped over, etc.
The full set of descriptions, available to him when he later talks
about the party, would be the same whether he was asked, Who is
Aston-Martin? at the outset when he claims to have met AstonMartin at the party or later on when the name occurs in recounting
other events involving the man met at the party. We may say that the
referent changes during the course of his conversation, but the
speaker would not. And his full account, i.e. all the descriptions at
his command, of who it is he refers to would remain the same. It
would contain, for example, both the author of Other Bodies
and the man I talked to at the party about cities.
This result, however, is inconsistent with the principle of identifying descriptions. On that principle, the same set of identifying descriptions can determine at most one referent. But in this example
we seem to have two referents and only one set of identifying
descriptions.
We extracted from the first counter-example the idea that the
question we should ask is, What would the user(s) of the name describe in this way? rather than, What (substantially) fits the descriptions they give? Though these questions may usually have the
same answer, the counter-example showed that they need not.
The present example, however, shows that even this distinction
is not enough. It would do no good to ask about his set of identifying descriptions, Who would the speaker describe that way? In
the example the same set of identifying descriptions is related to
two different referents. It seems then that the ultimate question is
rather, What would the speaker describe in this way on this occasion? where describe in this way does not refer to his set of iden70

PROPER NAMES AND IDENTIFYING DESCRIPTIONS

tifying descriptions, but to the predicate he ascribes to the referent;

e.g., in the example, we might ask on one occasion, Who would he
claim to have met at the party? on another, Who would he want us
to believe Jones tripped over at the party? And although his answer,
gleaned from his set of identifying descriptions, would be the same
in either case, we may have reason to answer differently to each
question.

X
It is instructive to look at the use of proper names in historical contexts if only to see why so many philosophers who discuss proper
names appeal to examples of it. In general, our use of proper names
for persons in history (and also those we are not personally acquainted with) is parasitic on uses of the names by other peoplein
conversation, written records, etc. Insofar as we possess a set of identifying descriptions in these cases they come from things said about
the presumed referent by other people. My answer to the question,
Who was Thales? would probably derive from what I learned from
my teachers or from histories of philosophy. Frequently, as in this
example, ones identifying descriptions trace back through many
levels of parasitic derivation. Descriptions of Thales we might give go
back to what was said, using that name, by Aristotle and Herodotus.
And, if Thales existed, the trail would not end there.
The history behind the use of a name may not be known to the
individual using it. I may have forgotten the sources from whence
I got my descriptions of Thales. Even a whole culture could lose this
history. A people with an oral tradition in which names of past
heroes figure would probably not be able to trace the history back to
original sources. Yet, for all that, they may be telling of the exploits
71

E S S AY S O N R E F E R E N C E

of real men in the past and they may possess knowledge of them and
their deeds.
Yet, in such cases the history is of central importance to the
question of whether a name in a particular use has a referent and, if
so, what it is. The words of others, in conversation, books and documents can, like the inverting spectacles in a previous example, distort our view of what we are naming. But at the same time it can, to
one who knows the facts, provide the means of uncovering the referent, if there is one.
The role of this history leading up to a present use of a name has
almost always been neglected by those who accept the principle of
identifying descriptions. The sort of description generally mentioned as helping to pick out, say, Thales, is such as the Greek philosopher who held that all is water. Nothing is made of the fact that
such descriptions are given by us derivatively. We might be pardoned if we supposed that the referent of Thales is whatever ancient Greek happens to fit such descriptions uniquely, even if he
should turn out to have been a hermit living so remotely that he and
his doctrines have no historical connection with us at all.
But this seems clearly wrong. Suppose that Aristotle and Herodotus were either making up the story or were referring to someone
who neither did the things they said he did nor held the doctrines
they attributed to him. Suppose further, however, that fortuitously
their descriptions fitted uniquely someone they had never heard
about and who was not referred to by any authors known to us. Such
a person, even if he was the only ancient to hold that all is water, to
fall in a well while contemplating the stars, etc., is not our Thales.
Or, to take the other possible outcome according to the principle
of identifying descriptions, suppose no one to have held the ridiculous doctrine that all is water, but that Aristotle and Herodotus were
referring to a real persona real person who was not a philosopher,
72

PROPER NAMES AND IDENTIFYING DESCRIPTIONS

but a well-digger with a reputation for saying wise things and who
once exclaimed, I wish everything were water so I wouldnt have to
dig these damned wells. What is the situation then regarding our
histories of philosophy? Have they mentioned a non-existent person
or have they mentioned someone who existed but who did not have
the properties they attribute to him? My inclination is to say the
latter. Yet ignoring the history of these uses of the name Thales, the
principle of identifying descriptions would tell us that Thales did not
exist. But then to whom were Aristotle and Herodotus referring?
Surely we cannot conclude, to no one. It seems to me to make sense
that we should discover that Thales was after all a well-digger and
that Aristotle and Herodotus were deceived about what he did. That
would not make sense, however, if we are forced to conclude in such
a case that he did not exist. That is, if we neglect the fact that there is
a history behind our use of the name Thales or Aristotle and concentrate only upon the descriptions we would supply about their life,
their works and deeds, it is possible that our descriptions are substantially wrong without the consequence being that we have not
been referring to any existent person.
It is significant that descriptions of the form N was referred to
by A should assume central importance in the case of uses of names
that are parasitic on their use by others. Not only does the principle
of identifying descriptions, as it has usually been defended, fail to
prepare us for the special role of one type of description, but we now
see that there is a quite ordinary sense in which a person might be
ignorant of the nature of the entity he has referred to in using a
name. While I do not want to classify descriptions of this form as
question-begging in the way in which the entity I have in mind is
question-begging, it seems nevertheless natural to say that in knowing only that Thales was a man referred to by Aristotle and Herodotus, Im not in a position to describe the man Thales; that is, there is,
73

E S S AY S O N R E F E R E N C E

I think, an ordinary use of describe in which to say only the man

referred to by Aristotle and Herodotus is not yet to describe Thales.
So it seems that we could be in the position of having referred to
someone in using the name Thales, the same person in fact referred
to by Aristotle and Herodotus, although we are not in the position
of being able to describe him correctly.
Nevertheless, so long as the user of a name can fall back on such
a description as the person referred to by Aristotle, the principle of
identifying descriptions may be salvaged even if at expense of having
to elevate one type of description to special status. But it is not at all
clear that such descriptions will in general be available to the user of
a name or that without them the failure of his other descriptions to
identify the referent uniquely must mean that the name has no referent. In the case of individual people there are surely many who
would, for example, identify Thales as the pre-Socratic philosopher
who held that all is water, but who do not know that he was referred
to by Aristotle and Herodotus. And in fact they may not know even
the immediate sources of their use of the name; that, for example,
Thales was referred to by Mr. Jones, their freshman philosophy instructor. In case Thales was in fact the pre-Socratic philosopher with
that doctrine, such people surely know something about Thales and,
in using the name, they have referred to him. But if, in fact, the attribution of this view to Thales is wrong and they are left without
any descriptions that uniquely fit Thales, I do not believe it follows
that they have not referred to anyone or that (in their use of the
name) Thales did not exist. To be sure, they may have available to
them some such description of Thales as, The one who is commonly believed to have been a pre-Socratic philosopher who held
that all is water. But even this may not be true. Everyone may have
come to believe that Thales did not have that doctrine. One could
continue along these lines, I think, to deny an individual any identi74

PROPER NAMES AND IDENTIFYING DESCRIPTIONS

fying descriptions, even of the form The one referred to by

so-and-so that will serve uniquely to pick out Thales, without
the consequence that he has not referred to anyone.

XI
The previous examples have concentrated on individuals and the set
of descriptions they could supply. But I think there is no reason to
suppose that, with a bit more stretching of the imagination, the same
results could not be gotten for the whole of some group in which a
name is used. Thus, those who would form the set of identifying
descriptions from a collective effort at description seem no better
off to me.
Thus, we could imagine a future time, for example, when the
plays we attribute to Shakespeare are available and it is believed that
Shakespeare was their author, but little else is known about him
perhaps only that he was an actor in Elizabethan timesand, in
particular, nothing about the documentation we rely upon in attributing the plays to him has survived. As we now view it, the people of
this future generation would be correct in saying that Shakespeare
wrote Hamlet. But suppose in fact the Baconian hypothesis is correctFrancis Bacon wrote those plays. What should an omniscient
being who sees the whole history of the affair conclude about one of
these future beings saying that Shakespeare wrote Hamlet? (Surely
not that as they use Shakespeare it refers to BaconBacon was not
an actor and they may know a great deal about Bacon, enough to
insure that he could not have been an actor.) It seems to me that the
correct conclusion should be that (perhaps because we did not pay
enough attention to the cryptologists who claim to find this message in the plays) we and they have made a mistakewe both
75

E S S AY S O N R E F E R E N C E

believe that Shakespeare wrote the plays, though it was rather Bacon
and not Shakespeare who is the Bard.

XII
As I have admitted, my counter-examples are necessarily somewhat
artificial because of the vagueness of the position I want to attack.
Yet, it seems to me that even artificial examples are sufficient because I take the principle of identifying descriptions to be a doctrine
about how reference via proper names must take place. If these examples show that there are other possibilities for identifying the referent, they do their job. It is the idea that only a backing of descriptions
identifying the referent by its fitting them (or some sufficient
number of them) could serve to connect an object with a name that
I question.
On the positive side my view is that what we should substitute
for the question, What is the referent? is What would the speaker
be attributing that predicate to on this occasion? Thus, in an early
example, the parents of a child ask, Who would he say was a nice
man at a party of ours? when the child has said, Tom was a nice
man. How we answer such questions I do not have a general theory
about. It seems clear to me that in some way the referent must be
historically, or, we might say, causally connected to the speech act.
But I do not see my way clear to saying exactly how in general that
connection goes. Perhaps there is no exact theory.
The shift of question, however, seems to be important. One can
explain why the principle of identifying descriptions has seemed
so plausible, for example, while denying its validity. If a speaker says
a is , where a is a name, and we ask, To what would he on this
occasion attribute the predicate ? asking him for descriptions
76

PROPER NAMES AND IDENTIFYING DESCRIPTIONS

would normally be the best strategy for finding out. Generally we

know numbers of correct and even uniquely identifying descriptions of the referent of names we use. So others would naturally first
rely on these and look for what best fits them.
To illustrate this, we can imagine the following games: In the
first a player gives a set of descriptions and the other players try to
find the object in the room that best fits them. This is analogous to
the role of the set of identifying descriptions in the principle I object
to. In the other game the player picks out some object in the room,
tries to give descriptions that characterize it uniquely and the other
players attempt to discover what object he described. In the second
game the problem set for the other players (the audience in the analogue) is to find out what is being described, not what best fits the
descriptions. Insofar as descriptions enter into a determination of
what the referent of a name is, I suggest that the second game is a
better analogy. In that game, on the normal assumption that people
are unlikely to be badly mistaken about the properties of an object
they are describing, the other players would usually first look for an
object best fitting the descriptions given. But that need not always
be the best tactics. They may notice or conjecture that the circumstances are such that the describer has unintentionally mis-described
the object, the circumstances being such as distortions in his perception, erroneous beliefs he is known to hold, etc.
One final point: I earlier questioned whether we can really
expect that there must be a backing of descriptions behind the use
of a proper name. Insofar as I offer an alternative to the principle of
identifying descriptions, it has the merit of not requiring such a
backing. If a speaker says a is , where a is a name, the question of
what he referred to does not hinge on what he can supply in the way
of descriptionsthough what descriptions he does give, if any, can
constitute an important datum. It may be possible to answer the
77

E S S AY S O N R E F E R E N C E

question, To what would he on this occasion attribute the predicate is ? without any backing of descriptions.

NOTES
* I am indebted to students and colleagues for comments and suggestions, in particular Professor John Perry and Mr. Theodore Budlong. I believe also that some
departure from the traditional alternatives in theories about reference and proper
names is in the air and that views along some of the lines I take in this paper
I may share with others, although the view I attack is still the dominant one.
I believe that Saul Kripke has a very similar position, at least insofar as denial of
the prevalent theories go. And, indeed, I think I may owe one of my counter-examples to him through a secondhand source (although I did not understand the
relevance until much later). David Kaplans paper, Quantifying In, Synthese 19
(1969) 178214 , also seems to me to be in the same vein, though I am not sure
I agree with a variety of details and the main purpose of the paper is not to mount
an assault on theories of proper names.
1. P. F. Strawson, Individuals, Methuen & Co. Ltd., London, 1959, p. 20.
2. Mind 67 (1958) 166173.
3. Lectures on Logical Atomism in Logic and Knowledge (ed. by Robert C. Marsh),
George Allen & Unwin Ltd., London, 1956, p. 243.
4. Ibid., p. 252.
5. Proper Names, op. cit., p. 171.
6. In Reference and Definite Descriptions, The Philosophical Review 75 (1966)
pp. 281304, and Putting Humpty Dumpty Together Again, The Philosophical
Review 77 (1968) 20315.
7. I assume here Russells definition of denoting, which I think makes it a well-defined relation and ought always to be kept in mind in discussions of reference so
that other relations may be compared with it: An entity X is denoted by a definite description, the , just in case X uniquely possesses the property designated by .
8. Although I do not have space to develop it, my account of proper names in this
paper seems to me to make what I called referential definite descriptions (as
discussed in Reference and Definite Descriptions, op. cit.) a close relative of
proper names.
9. Below, Section VI.
10. E.g., in The Thought: A Logical Inquiry (translated by A. M. and Marcelle
Quinton), Mind (1956) 289311. Also in P. F. Strawson (ed.), Philosophical
Logic, Oxford Readings in Philosophy, Oxford University Press, Oxford, 1967,
pp. 1738.

78

PROPER NAMES AND IDENTIFYING DESCRIPTIONS

11. That is to say, if what they refer to is a function of the set of identifying descriptions each possesses. In that case there would be the logical possibility of each
speakers set picking out different objects, each possessing the properties one
speaker would attribute to the referent, but not those the other would.
12. Individuals, op. cit., pp. 19192.
13. Loc. cit.
14. Ibid., p. 182, footnote 1.
15. The last part of the remark is there simply to indicate that the parents need not
even consider what the child says to be true; not only does the child not have a
backing of descriptions, but the predicate in the sentence he uses need not
apply. This connects up with the position suggested later in the paper.
16. Individuals, op. cit., p. 182.
17. In the example as presented I have the subject of the experiment introduce the
names. Nothing hinges on this. The experimenters could just as well use the
names and give the subjects identifying descriptions. Nor is there any importance in the fact that the example contains people, the experimenters, in the
know. For all that, everyone concerned might have the inverting spectacles on
that I introduce.
18. The idea behind this example originated with me from a conversation with
Rogers Albritton in 1966 and may derive from Saul Kripke, who has, I believe,
a view about proper names not dissimilar to the one in this paper.
19. For the purpose of keeping the example within limits, I compress the two uses
of the name, that I claim refer, unknown to the speaker, to two different people,
into one conversation. I have sometimes, however, found it useful to make the
case stronger intuitively by supposing that the person met at the party, for example, who is not the famous philosopher, becomes a longer term acquaintance
of the speaker (who continues under the illusion that he is the famous man). In
subsequent conversation, perhaps months or years later and after his friends
have met the bogus philosopher, his use of the name is even more clearly a reference to the man he met at the party and whom he continues to see. Yet if he
claimed to know, as in my example, J. L. Aston-Martin, in circumstances where
it is clear that the point of the remark has to do with claiming to know a famous
man, I still think we would suppose him to have referred to Aston-Martin, the
famous philosopher, and not to man he met at the party, who later is one of his
close acquaintances.

79

This page intentionally left blank

[4]
SPEAKING OF NOTHING

Russell tells us1 in On Denoting to test our logical theories by

their capacity for dealing with puzzles.2 In this paper I raise the
question of how a theory of reference, one of recent origin, might
handle one of the major puzzles Russell mentioned. The theory of
reference that I have in mindand one I subscribe toI will call
the historical explanation theory. (It, or ones similar to it in important respects, has also been called the causal theory. For various
reasons, I prefer a different title.)3
Among a number of puzzles mentioned by Russell, two stand
out as more important than the others. One is the well-known problem of identity statements with which Frege begins his article, On
Sense and Reference,4 the question of how a statement of the form
a is identical to b, when true, can differ in cognitive value from a
corresponding statement of the apparently trivial form, a is identical to a. The second puzzle is the topic of this paper. In a large
number of situations speakers apparently refer to the nonexistent.
The most obvious example of this is, perhaps, the use of singular
terms in negative existence statementsfor example, The discoverer of the philosophers stone does not exist or Robin Hood did
not exist. The problem is, of course, well known and ancient in
First published in The Philosophical Review, 1974, 83: 331.
81

E S S AY S O N R E F E R E N C E

origin: such statements seem to refer to something only to say about

it that it does not exist. How can one say something about what does
not exist? For a few philosophers, to be sure, these questions have
led to attempts to provide the referent. But in general such attempts
have been met with suspicion. Russell certainly thought it a merit of
his theory of definite descriptions (and his fully developed views on
singular expressions) that such apparent references to the nonexistent were explained without having to entertain the idea of referents
of singular terms that are nonexistent.
Where the singular terms involved are definite descriptions, On
Denoting provided a solution to the two puzzles mentioned that
was at once a breakthrough in the treatment of these expressions
and satisfying in the coherent explanation it gave. Russells fully developed theory of singular terms, perhaps best represented in Lectures on Logical Atomism,5 extends the proposed solution to
ordinary proper names, for these turn out to be concealed definite
descriptions. The view of On Denoting now could be made to
cover most of the uses of singular terms in language as we actually
speak it and, moreover, seemed to meet the test of solving the various puzzles about reference. But the fully developed view also introduced a category of singular expressions that were acknowledged to
be rarely, if ever, found in everyday speechwhat Russell called
names in the strict logical sense or genuine names.
Genuine names and the motivation for giving pride of place to
such exotic singular terms have special interest for the historical explanation theory, because while its treatment of ordinary singular
expressions is radically different from Russells it has some similarities to his characterization of genuine names.
The question posed, then, is how the historical explanation theory
of reference can handle the puzzle that Russells view has no difficulty
with, the problem of apparent reference to the nonexistent.
82

SPEAKING OF NOTHING

I cannot in this paper plead the full case for the historical explanation theory, though I shall try to give its main features; so it
may be best to consider it an exercise in the hypothetical: if the
theory is correct what follows concerning apparent reference to
the nonexistent?6

I. THREE KINDS OF APPARENT REFERENCE

TO THE NONEXISTENT
We need to keep distinct three situations in which apparent reference to the nonexistent occurs. The differences are important in
their own right, but I need to call attention to them because one
kind of situation will be excluded from consideration in this paper.
I will, in the first place, distinguish what I will call discourse
about fiction from discourse about actuality; and, secondly,
within the latter category, the use of predicative statements from
the use of existence statements.7 What is to be excluded from consideration here is an account of discourse about fiction. (This is not,
of course, to say that such an account is not in the end needed.)
Under discourse about fiction I mean to include those occasions on which it is a presupposition of the discourse that fictional,
mythological, legendary (and so forth) persons, places, or things are
under discussion. I believe, for example, that said with the right intention, the following sentences would express true propositions:
The Green Hornets car was called Black Beauty, Snow White
lived with seven dwarves, and To reach the underworld, one must
first cross the River Styx. (By the right intention I mean that the
speaker wishes to be taken as talking about fiction, mythology, or
legend.) At the same time I also believe it is true that neither the
Green Hornet, his car, Snow White, nor the River Styx exists or ever
83

E S S AY S O N R E F E R E N C E

has existed. These two beliefs, however, are entirely consistent. And
therein lies the puzzle: how can there be true propositions that apparently involve predicating properties of what does not exist?
Discourse about actuality carries the presupposition that the
speaker is talking about people, places, or things that occur in the
history of our world. A puzzle arises when the speaker is unfortunate enough to use a singular expression, intending to attribute a
property to something, but fails, in his use of that expression, to
refer to anything. This very likely occurred, for example, some years
ago following the publication of The Horn Papers,8 that purported to
contain the diary of one Jacob Horn and that would, if genuine, have
shed light on the colonial history of Washington County, Pennsylvania. Many people believed them to be genuine, but, on the evidence, it seems likely that they are not and that Jacob Horn did not
exist. There must have been many believers, however, who made
statements using the name Jacob Horn with the intention of predicating various properties of a historical figure. For example, someone might well have said, Jacob Horn wrote about Augusta Town
and now we know where it was located. It would have been some
sort of inconsistencyexactly what kind is another questionfor
such a speaker then to affirm the nonexistence of Jacob Horn. This
contrasts with discourse about fictionthere one can, for example,
consistently deny the existence of Snow White while also stating
that she enraptured a prince.9
The puzzle about predicative statements, as I shall call them, in
discourse about actuality with a singular expression and no referent
is more subtle philosophically than the puzzle about fictional discourse. There is not the same possibility of stating something true.
Nor can the speaker with consistency acknowledge the nonexistence of what he speaks about. To see how statements such as those
made by believers in the authenticity of The Horn Papers can puzzle
84

SPEAKING OF NOTHING

a philosopher requires the ability to see a difficulty in how one can

even speak and be understood when using a singular expression with
no referent.
The difference between discourse about fiction and discourse about reality, it is important to keep in mind, is a matter
of presuppositions about the intent of the speech act. It is not
that in the one fictional characters are involved and in the other
real people, places, and things. A not too well-informed person might have taken (at least the first part of ) the movie
Doctor Strangelove for a documentary. His statement, Doctor
Strangelove, the top military scientist in the United States, is a
psychopath, would then be a bit of discourse about reality, even
though Doctor Strangelove is, in fact, fictional. On the other
hand, this very same sentence used by someone having seen the
whole movie would probably be a comment on the movie, a bit
of discourse about fiction.
While I will need often to consider predicative statements about
actuality, the problem I want to concentrate on concerns existence
statementsthose that have either the form S does not exist or
the form S exists, where S is a singular expression. Negative existence statements, unlike predicative statements, are true when there
is no referent for the singular expression. If I speak the truth in
saying, Jacob Horn does not exist, I would be apparently referring
to what does not exist. But even more paradoxically, the truth of
what I say depends directly upon the nonexistence of a referent for
Jacob Horn. Moreover, this is discourse about reality; I do not,
clearly, intend to talk about a fictional character. Negative existence
statements, of all those mentioned, bring apparent reference to the
nonexistent into sharpest focus.
It is of some importance to mention the difference between denying the existence of something altogether and denying its present
85

E S S AY S O N R E F E R E N C E

existence or its existence at some point or during some period in

time. To begin with we certainly want to distinguish between
(1) Napoleon no longer exists,

and
(2) Napoleon does not and never did exist.

The first statement is both true and not an apparent reference to

the nonexistent in the sense we want. (1) contains a reference to
Napoleon in the same way that Socrates was snub-nosed contains a reference to Socrates. (1) should, it seems, be put into the
class of predicative statements, despite the fact that existence is
involved. On the other hand, (2) is a paradigm of the kind of
statement that generates the problem of apparent reference to the
nonexistent.
What shall we say, however, about a statement such as
(3) Santa Claus does not exist?

Often, I believe, it expresses a statement of the same form as (2), an

absolute denial of existence, not confined to any one period in time.
But suppose, for example, that someone is unsure whether Jacob
Horn ever did exist, but is certain that he does not now exist. He
might express this by saying, Jacob Horn does not exist. So, perhaps, sentences of the surface form of (3) are ambiguous. But what
is the other meaning that they might have? The sentence given, in
the imagined circumstances, seems to me to be equivalent to
(4) Jacob Horn does not now exist.
86

SPEAKING OF NOTHING

This is neither the absolute denial of Jacob Horns existence nor the
predicative assertion that Jacob Horn no longer exists. (4), I believe,
amounts to the disjunction of the two: Either Jacob Horn does not
and never did exist or he did exist and does no longer. In which
case, the dichotomy illustrated by (1) and (2) is still maintained. In
what follows, however, it is the absolute denial of existence that will
be of concern and any examples of the form N does not exist
should be construed in that way.

II. A THEORY GONE WRONG WITH

INTERESTING MOTIVESRUSSELL
Russells theory of singular terms holds interest for the historical explanation theory, not only because of obvious oppositions on some
key issuesseveral more recent discussions would serve that purpose10but also because certain problems and issues that evidently
motivated features of Russells theory that are nowadays generally
ignored or thought obviously wrong are brought to the fore once
again by the historical explanation theory. I believe that much of
Russells theory has been accepted by many philosophers with the
thought that there was a certain excrescence that could be ignored.
Russells views on ordinary singular terms, definite descriptions,
proper names in ordinary language have wide acceptance; his addition of genuine names to the ranks has generally been ignored as
so much metaphysical meandering. I think there is no doubt that
genuine names, as Russell characterized them, have no place in a
correct theory of reference. But from the first, in On Denoting,
Russell contrasted his account of those singular terms for which his
theory provided a way out of puzzles about reference with another
kind of singular term, a genuine name, for which he seemed to feel
87

E S S AY S O N R E F E R E N C E

there was a theoretical need. But, of course, genuine names, if they

were to be included in the general theory, could not reintroduce the
same puzzles. This, I think, accounts for some of the peculiar properties attributed to genuine names: for example, the distinction
between knowledge by acquaintance and knowledge by description that gave the result that we could only genuinely name something we are acquainted with in a very strong sense seems to have
been introduced in part to make it impossible to assert negative existence statements using genuine names.
The reason this has interest for the historical explanation theory
is that Russells contrast, the radical difference between most singular terms in ordinary language and genuine names, is that the
former have descriptive content and the latter do not. Given his
view of singular terms with descriptive content, the puzzles about
reference yield easily for them. He felt, however, some need to have
singular expressions, nonetheless, that do not function in accord
with his analysis of definite descriptions and ordinary proper names.
The historical explanation theory denies that, for at least many uses
of ordinary singular expressions, Russells view is correct. In particular, it denies that ordinary proper names always have descriptive
content. The question is, does this mean that perhaps ordinary singular expressions may fulfill the function that Russell thought only
genuine names, with all their peculiarities, could? And, if so, how
can the historical explanation view deal with the puzzles about
reference?
What was the motivation for introducing genuine names?
Russell often talks in ways that can seem nonsensicalthat, for example, when a definite description such as the author of Waverly is
involved, the denotation of the definite description, Scott in this
case, is not a constituent of the proposition expressed. The implied contrast is that if Scott is a genuine name and were there in
88

SPEAKING OF NOTHING

place of the definite description the author of Waverly then Scott

would be a constituent. But it certainly sounds queer at first glance
to find a flesh-and-blood person in a proposition!
Russells analysis of statements containing definite descriptions
and, by extension, ordinary proper names, shows, he believed, that
such statements are not really about, do not really mention, the denotation of the description or the referent of the name. Russell emphasizes this again and again. Genuine names, on the other hand, can
somehow perform the feat of really mentioning a particular individual. To try to put much weight on such terms as about would
lead us, I think, into a morass. What it is for a statement to be about
an individual, if that requires any attempt to define aboutness, is a
question better avoided if we are ever to get on with the problem.
(After all, Russell himself recognized a well-defined relationship
that a statement containing a definite description can have to some
particular individualits denotation. It would be a delicate task to
show either that in no sense of about is such a statement about the
denotation of the definite description or that there is some clear
sense of about in which it is not.)
But I believe we can say something useful about the reasons
Russell had for talking in this way. On his theory of definite descriptions the singular expression, the definite description, is really a
device that introduces quantifiers and converts what might seem at
first sight a simple proposition about an individual into a general
proposition. The is expresses the same proposition as There
is a and there is at most one and all s are s; and the latter
clearly would express a general proposition about the world. Ordinary proper names, of course, function on his view in the same way,
since they are in reality concealed definite descriptions. Now if we
contrast these singular expressions with ones, if there are any, that
do not introduce quantifiers, that when put as the subject of a simple
89

E S S AY S O N R E F E R E N C E

subject-predicate sentence do not make the sentence express a general proposition, then I think there is a strong temptation to say that
only the second kind of singular term can be used to really mention
an individual.
Russell clearly believed that there must be the possibility, at
least, of singular terms that do not introduce quantifiers; that seems
in large part to be his reason for believing in genuine names.
Whether or not there is some argument that shows the necessity of
such singular terms, I believe that prior to theory the natural view is
that they occur often in ordinary speech. So if one says, for example,
Socrates is snub-nosed, the natural view seems to me to be that the
singular expression Socrates is simply a device used by the speaker
to pick out what he wants to talk about while the rest of the sentence
expresses what property he wishes to attribute to that individual.
This can be made somewhat more precise by saying, first, that the
natural view is that in using such simple sentences containing singular terms we are not saying something general about the world
that is, not saying something that would be correctly analyzed with
the aid of quantifiers; and, second, that in such cases the speaker
could, in all probability, have said the same thing, expressed the
same proposition, with the aid of other and different singular expressions, so long as they are being used to refer to the same individual. To illustrate the latter point with a different example: if, at
the same moment in time, one person were to say, Smith is happy,
a second You are happy, a third My son is happy, and a fourth
I am happy, and if in each case the singular expression refers to the
same person, then all four have expressed the same proposition,
have agreed with each other.
What I see as the natural pre-theoretical view might be captured
as a certain way of representing what proposition is expressed. For
example, the sentence Socrates is snub-nosed might be repre90

SPEAKING OF NOTHING

sented as an ordered pair consisting of Socratesthe actual man, of

course, not his nameand the predicate (or property, perhaps),
being snub-nosed. (More complicated sentences, involving relations and more than one singular expression of this sort would be
represented as ordered triplets, and so forth.) Now if someone were
to say to Socrates, You are snub-nosed, or Socrates were to say
about himself, I am snub-nosed, the proposition expressed would,
in each case, be represented by the same ordered pairpropositional identity, given the same predicate, would be a function simply
of what individual is referred to.
This way of representing propositions would, I think, meet with
at least provisional approval by Russell, but only if it were restricted
to those propositions expressed by statements containing genuine
names. We might even say that the manner of representation gives a
respectable sense in which an individual might be a constituent of a
proposition. But my examples of statements for which this representation was suggested would, on Russells view, be incorrect just
because they involve singular terms from ordinary language. For
Russell, they would be examples of sentences that express complex
general propositions and, whatever our view of the nature of propositions, I do not think we would want propositional identity for general propositions to be a function of the individuals that happen to
make the propositions true or false.
Russell pays the price, I believe, of giving up the natural view of
many uses of ordinary singular terms, a price he is willing to pay
chiefly, perhaps, because he thus can dissolve puzzles about reference.
The special properties of genuine names, on the other hand, are supposed to rescue them. The natural view, on the other hand, seems to
generate Russells budget of puzzles, in particular the one which is the
concern of this paper. If I say, Socrates is snub-nosed, the proposition I express is represented as containing Socrates. If I say, instead,
91

E S S AY S O N R E F E R E N C E

Jacob Horn does not exist, the natural view seems to lead to the
unwonted conclusion that even if what I say is true, Jacob Horn,
though nonexistent, must have some reality. Else what proposition
am I expressing? The natural view thus seems to land us with the
Meinongian population explosion.
Russell, of course, avoids this problem easily. Since the proper
name Jacob Horn would, for him, be a concealed definite description, to say Jacob Horn does not exist is not to refer to some individual in order to say something about him, but merely to assert that
a particular class of things, perhaps the class of writers of diaries
about certain events in early Pennsylvania history, is either empty or
contains more than one member. (So a singular nonexistence statement of this kind is on all fours with statements such as There are
no flying horses or There is more than one living ex-President. It
does not mention a particular individual any more than these do.)
The issue has importance for the historical explanation view because it denies that many singular terms in ordinary language, in
particular proper names, are concealed descriptions of the sort that
Russell had in mind. Homer, for example, is not a concealed description such as the author of the Homeric Poems, to use Russells own example. The question is, does the historical explanation
view, if correct, support what I have called the natural view? In the
next section this question will be considered.

III. THE HISTORICAL EXPLANATION VIEW:

NEGATIVE ASPECT
I now want to begin to lay out the bare bones of the theory of reference I want to discuss. As I have said, I will not here argue for its
correctness nor will I try to fill in all the gaps.
92

SPEAKING OF NOTHING

Russell and the majority of philosophers in contemporary

times who have discussed (ordinary) proper names have held
that by one mechanism or another they are surrogates for descriptions. For Russell, as I have mentioned, they are simply abbreviations for definite descriptions; for othersfor example,
Searle11they are correlated with a set of descriptions and what
one is saying in, say, a simple subject-predicate sentence employing a proper name is that whatever best fits these descriptions has
whatever property is designated by the predicate. The descriptions, both on Russells view and on the looser view of Searle and
others, which the proper name masks, are thought of as obtained
from the people who use themroughly speaking, by what they
would answer to the question, To whom (what) are you referring? This view of ordinary proper names embodies what I have
called the principle of identifying descriptions.12 The theory of
reference I am concerned with holds that the principle of identifying descriptions is false.
What this means, to give an example, is that, supposing you
could obtain from me a set of descriptions of who it is that I believe
myself to refer to when I say, Socrates was snub-nosedperhaps
such things as the mentor of Plato, the inventor of the Socratic
method, the philosopher who drank the hemlock, and so forth
it is theoretically possible that I am referring to something about
which no substantial number of these descriptions is true or that
although there is something that fits these descriptions to whatever
extent is required by the particular variation of the principle, that is
not in fact the referent of the name I used.
On this theory, then, ordinary proper names are like Russells
genuine names at least insofar as they do not conceal descriptions in
the way he thought. This is, I think, a virtue of the theory. As David
Kaplan has remarked, there was always something implausible
93

E S S AY S O N R E F E R E N C E

about the idea that a referent of a proper name is determined by the

currently associated descriptions.

IV. THE HISTORICAL EXPLANATION

THEORY: POSITIVE ASPECT
The first tenet of the theory of reference I have been describing was
negativethe view that proper names must have a backing of descriptions that serves to pick out their referents is false. The second
tenet is positive, but more tentative. How is the referent of a proper
name, then, to be determined? On Russells view and variants on it,
the answer to this question would be simple: the referent is that
which fits the associated descriptions best, where best may be defined differently by different writers. As I see it, one of the main reasons a backing of descriptions for proper names is so attractive is
that it furnishes a simple way of ascertaining what a speaker is saying,
and of determining whether what he says is true or false (given that
we are dealing only with assertions). We find, so to speak, that thing
in the world which uniquely fits the descriptions and then see
whether or not it has the properties ascribed to it. If proper names
do not have a backing of descriptions, how do we decide whether or
not, when someone says, for example, Russell wrote On Denoting, he has said something true or false?
Putting existence statements aside, when a speaker says something of the form, N is , where N is a name and a predicate,
we can say that in general the truth conditions will have the following form. What the speaker has said will be true if and only if
(a) there is some entity related in the appropriate way to his use of
N in this sentencethat is, he has referred to some entity, and
(b) that entity has the property designated by . (I say in general
94

SPEAKING OF NOTHING

because there are difficulties for any theory of reference about uses
of names for fictional characters, formal objects such as numbers,
and so forth.) The question is, what is the appropriate relation
mentioned in condition (a)? How, that is, does an entity have to be
related to the speakers use of the name N to be its referent? The
principle of identifying descriptions, were it only true, has a simple
answer to this: the entity must have (uniquely) the properties or
some sufficient number of the properties designated by the backing of descriptions for this use of the name N. Roughly speaking,
and on the most usual view, it will be the entity that answers to the
descriptions the speaker would (ideally) give in answer to the question, To whom are you referring?
But even without the arguments that, I believe, show the principle of identifying descriptions not only false, but implausible, putting the matter in this general way is somewhat liberating. It shows
that what we need is some relation between the speech act involving
the name N and an object in the worldthe right one, of course
but the relation supplied by the principle of identifying descriptions
is now only a candidate for that office.
But if the principle of identifying descriptions is false, what then
is the appropriate relation between an act of using a name and some
object such that the name was used to refer to that object? The
theory of reference I want to discuss has not as yet, so far as I know,
been developed in such a way as to give a completely detailed answer.
Yet there are positive things that can be said and enough, I believe,
both to contrast it with the principle of identifying descriptions and
to give us something like an answer to the original question: how
will it handle apparent reference to the nonexistent in such statements as Santa Claus does not exist?
The main idea is that when a speaker uses a name intending
to refer to an individual and predicate something of it, successful
95

E S S AY S O N R E F E R E N C E

reference will occur when there is an individual that enters into the
historically correct explanation of who it is that the speaker intended
to predicate something of. That individual will then be the referent
and the statement made will be true or false depending upon
whether it has the property designated by the predicate. This statement of the positive thesis leaves a lot to be desired in the way of
precision, yet with some clarifying remarks I think it has more content than might at first sight be supposed.
Suppose someone says, Socrates was snub-nosed, and we ask
to whom he is referring. The central idea is that this calls for a historical explanation; we search not for an individual who might best
fit the speakers descriptions of the individual to whom he takes
himself to be referring (though his descriptions are usually important data), but rather for an individual historically related to his use
of the name Socrates on this occasion. It might be that an omniscient observer of history would see an individual related to an
author of dialogues, that one of the central characters of these dialogues was modeled upon that individual, that these dialogues have
been handed down and that the speaker has read translations of
them, that the speakers now predicating snub-nosedness of something is explained by his having read those translations. This is the
sort of account that I have in mind by a historical explanation.
Several comments are in order here. First, it is not necessary, of
course, that the individual in question be snub-nosed; obviously the
speaker may have asserted something false about the referent of the
name Socrates. Second, if we take the set of descriptions the speaker
could give were we to ask him to whom he was referring, the historical explanation as seen by our omniscient observer may pick out an
individual as the referent of the name Socrates even though that
individual is not correctly described by the speakers attempt at
identification. For example, the speaker may believe that Socrates
96

SPEAKING OF NOTHING

that is, the person he refers towas a philosopher who invented the
Socratic method. But it is clearly imaginable that our omniscient
observer sees that while the author of the dialogues did intend one
of the characters to be taken as a portrayal of a real person, he modestly attributed to him a method that was his own brain child. And,
in general, it would be possible to have the historical connection
with no end to mistaken descriptions in the head of the speaker. The
descriptions the speaker gives, however, may play an important role,
though not the one given to them by the principle of identifying
descriptions. The omniscient observer may see, for example, that
the reason the speaker believes himself to be referring to someone
who invented a certain philosophical method is that his present use
of the name Socrates is connected with his having read certain
translations of these dialogues. Or, to take a slightly different case,
he may see that his descriptions come from a faulty memory of
those dialogues, and so forth. The question for the omniscient observer is What individual, if any, would the speaker describe in this
way even if perhaps mistakenly?
I have used the notion of an omniscient observer of history and,
of course, we ordinary people cannot be expected to know in detail
the history behind the uses of names by those with whom we converse. Nor do we often make the sort of historical inquiries which
would reveal those details. We often assume, for example, that if another speakers descriptions of the referent of a name he has used
more or less jibe with descriptions we would give of a person, place,
or thing that we believe ourselves to know about, then he is referring
to that. Also, for example, the context of the use of a name may lead
us to assume without question that the speaker refers to someone
with whom we are both acquainted. But the historical explanation
theory need not deny this or be troubled by it. All it needs to hold is
that the final test for reference is the kind of historical connection
97

E S S AY S O N R E F E R E N C E

I have described, that the customary assumptions and use of indicators are in the end dependent upon being fairly reliable guides to the
existence of such a connection.
What the historical explanation theory must attempt to establish is that when there is an absence of historical connection between an individual and the use of a name by a speaker, then, be
the speakers descriptions ever so correct about a certain individual, that individual is not the referent; and, on the other hand, that
a certain historical connection between the use of a name and an
individual can make the individual the referent even though the
speakers descriptions would not by themselves single out the individual. This job must be accomplished by building up examples
in which these two points are made obvious. We might, for instance, try to show that the historical connection is necessary by
constructing a situation in which, for instance, one person begins
by assuming that another is referring to a friend of his, perhaps
because the descriptions seem accurate, the context is appropriate, and so forth, and who then discovers that it is practically impossible for the speaker to have been acquainted with or otherwise
related to his friend. In such an event, surely confidence that the
speaker was referring to the friend would be shaken despite the
apparent accuracy of description or appropriateness of context.
But, as I have said, I cannot here undertake the full defense of the
historical explanation theory.
There are, however, two further points of clarification that ought
to be mentioned here. It should be obvious that I have only provided an example of what counts as a historical explanation rather
than a formula for obtaining the referent of a particular use of a
name. Even in the illustration several individuals entered into the
account, only one of which was the referent. Of the individuals who
are in some way or other part of the historical explanation of a use of
98

SPEAKING OF NOTHING

a name, which is the referent? What kind of theory is this if it does

not give us the means to make this determination?
In defense against this charge that the theory is excessively
vague, it is helpful, I think, to compare it with another philosophical
theory about a quite different problem. The causal theory of perception can be taken as holding that an observer, O, perceives an object,
M, only if M causes O to have sense impressions. The theory seems
to me to have content and to be important, whether or not it is correct. For one thing, if true it means that certain other theories are
mistaken. But the theory as stated does not, obviously, allow us to
say which among the various causal factors involved in an observer
having sense impressions is the thing he perceives; nor does it tell us
which ways of causing sense impressions are relevant. Possibly no
philosophical analysis can determine this, although in any particular
case we may be able to say that this is or is not the right sort of causal
connection. Analogously, the historical explanation theory lacks
this sort of specificity. But for all that, if it is true, certain other theories, in particular the identifying descriptions theory, will be wrong
and the theory does tell us something of importance.
Because there have sometimes been misunderstandings about
this, I think I should point out that the history to which the historical explanation theory alludes is not the history of the use of a name.
It is not the history of the use of, say, the name Socrates that is
important. Socrates may not have been, as far as theory goes, called
Socrates; corruption of names is just as possible as corruption of
information. (The history of such a corruption, however, might enter
into the historical explanation.) Nor, I think, should the theory be
construed as holding that the historical connections end with some
original dubbing of the referent. It may be that people, places, and
things usually receive names by some such ceremony and that we
generally use names (or corruptions of them) as a result of such a
99

E S S AY S O N R E F E R E N C E

ceremony, but it is not a theoretical necessity that names enter our

linguistic transactions in this way.13
What the historical explanation does, then, is to provide the relationship between the use of a referring expression and the referent
which the principle of identifying descriptions presupposes could
be provided only by some measure of correct descriptions of the
referent known to the speaker. I think there are counter-examples to
the principle of identifying descriptions14 and, of course, if there are
that defeats it straight off. Still a plausible, if not clearly correct, alternative theory in this case also acts as an objection. For one of the
principle reasons that many philosophers have for adopting the
principle of identifying descriptions is that they cannot see how
there could be an appropriate relation otherwise that would pick out
the referent of (as the main example) a proper name.
I have, in describing this theory of reference, talked about a historical explanation. I hope it is obvious that historical is being
used in the broadest sense possible; that all of what I have said could
just as well be applied to cases in which one refers, by use of a name,
say, to someone still extant, to someone who has just gone out of the
room, or to someone presently in ones company. The historical explanation, in other words, can involve as brief an interval of time as
one pleases.

V. A SOLUTION TO THE PUZZLE REJECTED

My problem, then, is to show how such a theory of reference can
deal with simple existence statements expressed by the use of a
proper name, the difficulty being that on this theory, proper names
do not have a backing of descriptions and, in general, they function
to refer via what I have called a historical connection with some
100

SPEAKING OF NOTHING

individual. But a true negative existence statement expressed by

using a name involves a name with no referent and the corresponding positive existence statement, if false, will also. But in other contexts, when a name is used and there is a failure of reference, then no
proposition has been expressedcertainly no true proposition. If a
child says, Santa Claus will come tonight, he cannot have spoken
the truth, although, for various reasons, I think it better to say that
he has not even expressed a proposition.15
One apparently possible solution to the problem must be rejected. Russell and others, as we have seen, thought of (ordinary)
proper names as concealed definite descriptions; he held a version, that is, of the principle of identifying descriptions. Existential statements involving ordinary proper names were therefore
no problem for himthey were really existential statements involving definite descriptions and could be analyzed in accordance
with his theory of definite descriptions. The suggestion I want to
look at is that while our theory tells us that names in predicative
statements do not obey the principle of identifying descriptions
and are not concealed definite descriptions, existential statements
may represent a special case. Thus, so this suggestion would run,
Santa Claus in Santa Claus will come tonight is not a concealed definite description, but is one in the special context of
Santa Claus does not exist and Santa Claus exists. This would,
of course, immediately solve our problem, but unfortunately it is
not a solution that our theory can accept. The difficulty is not that
names would be treated as functioning differently in different
contexts; in fact, as will become evident, my own view is that they
do behave differently in existence statements. Rather, the trouble
is that any theory that rejects the principle of identifying descriptions for predicative statements must also reject it for existence
statements.
101

E S S AY S O N R E F E R E N C E

To simplify matters, let us restrict ourselves to Russells version

of the principle of identifying descriptions in which a name simply
stands in place of some definite description. If we adopt the principle for existence statements involving names, this will come to
saying that, for example, Socrates did not exist means the same
thing as (expresses the same proposition as) some other sentence
formed from this by replacing Socrates by a definite description
perhaps, say, the sentence, The Greek philosopher who was convicted of corrupting the youth and drank hemlock did not exist.
But, now, on any view we must, I think, accept the following:
(E) That Socrates did not exist entails that it is not true that
Socrates was snub-nosed.

Our theory tells us that the second occurrence of Socrates in (E)

is not a concealed definite description. But then neither can the first
occurrence be one. For if we take some definite description such as
the one suggested as what the first occurrence of Socrates stands
for, rejection of the principle of identifying descriptions for the
second occurrence means that it could be true that Socrates was
snub-nosed even though no unique individual existed who satisfied
that description. That is to say, if Socrates in Socrates did not
exist is a concealed definite description, but is not in Socrates was
snub-nosed, then the antecedent of (E) could be true while the
consequent is false. Since we want to accept the entailment expressed by (E) our theory cannot treat Socrates as a concealed
description in existential statements.
This solution not being open to us, we cannot on the other hand
go to the opposite extreme and handle existential statements
involving ordinary proper names in the way Russell did for what
he called names in the strict logical sense. There simply are no
102

SPEAKING OF NOTHING

meaningful existential statements involving these genuine names

and so the problem does not arise about how to deal with them.
But, of course, we cannot countenance this about ordinary proper
names, for it does make sense to say, Homer existed or Santa
Claus does not exist.

VI. TRUTH CONDITIONS AND BLOCKS

What we need to do first is see what, on our theory of reference, the
truth conditions are going to look like for existence statements involving names. In predicative statements, such as Homer was a
great poet, if everything goes well, there will be some individual
related to this use of Homer historically, as I have put it, and the
statement will be true if that individual had the property expressed
by the predicate and false otherwise. This, of course, cannot be so
for a negative existence statement such as Homer did not exist.
This statement would be true, in fact, just in case there is a failure of
reference, not in the statement itself, but in other possible or actual
predicative statements involving the name. That is, if there is no individual related historically in the right way to the use of Homer
in, say, the statement Homer was a great poet, no individual whose
possession or nonpossession of poetic genius makes this true or
false, then we can truly state that Homer did not exist.
Initially then the question comes to this: What, on our theory,
constitutes a failure of reference in a predicative statement involving
a proper name? (As we shall see there is more to the matter than
just this.) Since the positive part of our theory, the part that attempts
to say what successful reference to an individual consists in, has
been, perhaps because of the nature of things, left more suggestive
than in a rigorously formulated state, it cannot be hoped that we
103

E S S AY S O N R E F E R E N C E

shall do much better with failure of reference. But we can say some
things of a non-trivial nature.
Suppose a child who believed in Santa Claus now learns the
truth, the truth which he expresses by saying, Santa Claus does not
exist. He comes to learn this, as usual, from cynical older children;
what has he learned? Our account is that he has learned that when
in the past he believed something, for example, which he would
have expressed by saying, Santa Claus comes tonight, and would
have thought himself in saying this to be referring to someone, the
historical explanation of this belief does not involve any individual
who could count as the referent of Santa Claus; rather it ends in a
story given to him by his parents, a story told to him as factual. I do
not mean, of course, that the child would or could express the
knowledge he has in his new state of disillusionment in this fashionthat would require him to know the correct account of reference. But if we are approaching the correct theory, then this is how
we can state what he has discovered.
When the historical explanation of the use of a name (with the
intention to refer) ends in this way with events that preclude any
referent being identified, I will call it a block in the history. In this
example, the block is the introduction of the name into the childs
speech via a fiction told to him as reality by his parents. Blocks occur
in other ways. For example, children often invent imaginary companions whom they themselves come to speak of as actual. The
block in such a case would occur at the point at which a name for the
unreal companion gets introduced by the child himself via his mistaken belief that there is a companion to name. A somewhat different example would be this: suppose the Homeric poems were not
written by one person, but were a patchwork of the writings of many
people, combined, perhaps, with fragments from an oral tradition.
Suppose, further, that at some point in time an ancient scholar for
104

SPEAKING OF NOTHING

whatever reasonshe might have seen a name attached to some

written version of the poems and supposed it to be the name of the
authorattributed the poems to a single person he called Homer.
If this were the historical explanation of our saying, for example,
Homer wrote the Iliad, then the block occurs at the point at which
this scholar enters the picture.
On theories that subscribe to the principle of identifying descriptions, examples of failure of reference such as occur in this last
example would be treated as a failure to satisfy a uniqueness condition. The reason that Homer would not have existed given these circumstances is that no single individual satisfies the descriptions we
associate with Homer (or satisfies a sufficient number, according
to certain views). But according to our theory this is not the reason
for failure of reference; it is rather that the history of our use of the
name, a history with which we may not be familiar, does not end in
the right way. One way to see that the opposing account, though
plausible, is wrong is to think of the possibility of someone existing
who does satisfy the descriptions we might supply of the referent of
a name we use, but who has no historical connection with us whatsoever. Suppose, for example, that contrary to what we adults believe we know, there is, in fact, a man with a long white beard and a
belly like a bowl full of jelly who comes down chimneys on Christmas night to leave gifts (the ones whose labels are missing about
which parents worry because they dont know to what aunt the child
should write a thank-you note). We must, of course, imagine that it
is absolutely fortuitous that our descriptions of Santa Claus happen
to fit so accurately this jolly creature. In that case I do not think that
he is Santa Claus. The fact that the story of Santa Claus, told to children as fact, is historically an invention constitutes a block even if
the story happens to contain only descriptions that accurately fit
some person.
105

E S S AY S O N R E F E R E N C E

VII. A RULE FOR NEGATIVE EXISTENCE

STATEMENTS
Using the technical, but admittedly not well-defined, notion of a
block, we can now sketch the way the historical explanation theory
may treat negative existence statements involving names. A similar
treatment could then be given for positive existence statements.
I will suggest a rule, using the notion of a block, that purports to
give the truth conditions for negative existence statements containing a name. This rule, however, does not provide an analysis of such
statements; it does not tell us what such statements mean or what
proposition they express. This means that in this case we are divorcing truth conditions from meaning.
With the deletion of some qualifications that would be needed
to make it strictly correct, the rule can be expressed as follows:
(R) If N is a proper name that has been used in predicative statements with the intention to refer to some individual, then N does
not exist is true if and only if the history of those uses ends in
a block.

The rule as stated obviously requires some modifications. For one

thing we would need some way of distinguishing, for example, the
denial of the existence of Aristotle the philosopher, from Aristotle
the ship magnate. To accomplish this we must do two things: first,
find a means of collecting together the uses of Aristotle in predicative statements that were, so to speak, attempts to refer to the philosopher, separating them from a similar collection of uses of the
name that were attempts to refer to the ship magnate, and do this
without, of course, assuming that any of these uses succeeds in referring. Second, we must be able to relate a particular negative
106

SPEAKING OF NOTHING

existence statement using the name Aristotle to one such collection rather than any other.
The way of amending Rule (R) that seems to me in keeping with
the historical explanation theory and to accomplish these tasks is
this. Certain uses of the name Aristotle in predicative statements
will have similar histories, histories that will distinguish them from
other uses of the name. Each use of the name will, of course, have its
own historical explanation, but these may, at a certain point, join up.
So, in tracing back several uses of the name Aristotle by me and
several uses by you, we may find a common root in certain ancient
writings and documents, while other uses of the name by me or by
you may have nothing in common with the history of the first set of
uses. It is possible that the histories may join at what I have called a
block. Another possibility, however, is that although different uses
of the name end in different blocks, these blocks are themselves historically connected. This might occur, for example, for the use by
different children of the name Santa Claus. I have suggested that
the block in this example occurs where the parents tell the children
a fiction as if it were fact. The block, however, would be a different
one for each child. Still the blocks themselves are historically related
in an obvious way since the parents deception is rooted in a common
tradition.
Still another possible source of difficulty with Rule (R) as stated
is that it makes use of prior instances of the name in predicative
statements. Is it possible meaningfully to assert N does not exist
when N has never been used in predicative statements (about actuality)? If it is, then Rule (R) would have to be amended in some way,
perhaps by talking of potential or possible uses. But at the moment
I am not sure how this would go and I will not attempt it.
Even without worrying about the vagueness of the idea of a
block, Rule (R) may look unexciting, but its consequences are
107

E S S AY S O N R E F E R E N C E

interesting. In the first place its form is completely antithetical to the

principle of identifying descriptions, for it has nothing to do with
whether an individual of a certain description existed or not. Second,
it does not involve our theory of reference in any difficulties: there is
the connection with the notion of historical explanation and so it
ties in neatly with the positive aspects of the view, but it has no Meinongian implications, no overpopulation with entities whose existence is being denied. This result is bought, to be sure, at the price of
making a name function differently in existence statements as opposed to predicative statements. But, as I have said, I think that this
is not an unintuitive result.
While the above are important consequences of (R), what interests me about (R) is that it gives the truth conditions for statements
that assert that some individual does not exist in terms of a linguistic
failurethe failure of a name to refer on account of a block. And
it should occur to one that there may be something wrong with this.
How, it might be asked, can Homers existence or nonexistence be a
matter of a fact about language, a fact about the name Homer?
One is reminded, at this point, of a similar problem connected with
the other puzzle about reference mentioned at the beginning of this
paper. In On Sense and Reference, immediately after propounding
the puzzle about identity statements, Frege mentions a solution that
he had formerly thought correct, but which he now repudiates just
because it seems to involve turning identity statements, which apparently express facts about the world, into statements about a particular language.
Rule (R), insofar as it is supposed to express truth conditions for
negative existence statements of a certain kind, seems objectionable
for the same reasons. The crux of the problem in both cases seems to
be this. We are inclined to say that the propositions expressed by us
as The Evening Star is identical with the Morning Star and Homer
108

SPEAKING OF NOTHING

did not exist can be the very same propositions that someone else
may express using entirely different names. Therefore, how can we
give a rule, such as (R), which makes the truth conditions of what we
say depend upon facts about particular names?
The child who has become disillusioned expresses his newfound knowledge by saying Santa Claus doesnt exist. A Frenchspeaking child, with a similar history of being deceived by adults,
might express his discovery by saying, Pre Nol nexiste pas. Although the names are different, I believe we should want to say
that the two children have learned the same fact and, on that account, that they have expressed the same proposition. Yet if we
apply Rule (R) to each case it seems that the truth conditions must
be different; they involve a block in the history of the use of the
name Santa Claus for the English-speaking child and a block for
the French-speaking child in the history of the use of the different
name, Pre Nol.
Perhaps we can see the problem more clearly by looking for a
moment at predicative statements. If we consider a simple (grammatically) subject-predicate statement, such as Socrates is bald,
and think of this as divided into its referring element, Socrates,
and its predicative element, is bald, then if a certain change in the
predicative elementfor example, from is bald to is shortresults in a change in the truth conditions for the statement, we want
to say that the result expresses a different proposition. In general
only interchange of synonymous predicates will maintain the same
truth conditions and the same proposition. If referring expressions
such as Socrates were concealed descriptionsthat is, introduced
predicate elements into a statementthen the same could be said
about them: substituting a different referring expression, unless it
happened to conceal the same or synonymous descriptions as the
one it is substituted for, would shift both the truth conditions and
109

E S S AY S O N R E F E R E N C E

the proposition expressed (and, in fact, this is the heart of Freges

way of avoiding his puzzle about identity statements).
But our theory of reference denies that referring expressions
such as Socrates conceal descriptions or introduce predicate elements. If we keep the predicative element the same and substitute a
different referring expressionsay, Plato for Socratesthen
whether or not we have the same proposition expressed depends
solely upon whether or not the same thing is referred to. And this in
turn depends upon whether the historical explanation of the use of
these two expressions traces back to the same individual. If you say
Henry is bald and I say George is bald we express the same proposition if the person you referred to by using the name Henry and
I by using the name George are the same person. But what you say
is true if and only if the person you referred tothat is, the person
historically connectedwhen you used the name Henry has the
property of being bald; whereas what I say is true if and only if what
I referred to by using the name George has the property of being
bald. The truth conditions are different because they must be stated
in terms of what is referred to by different expressions, in the one
case my use of the name George and in the other your use of the
name Henry. Yet we may express the same proposition.
So with predicative statements involving proper names, given
the same predicate, sameness or difference of propositions comes
down to sameness or difference of the referent of the names. It
seems that if we try to state the truth conditions for a particular use
of such a statement, we are not going to arrive at what we should like
to call the proposition expressed. But although we thus are separating truth conditions from propositions expressed, the latter notion
is still a fairly clear concept. It seems, however, that we cannot in the
same way preserve a clear notion of what proposition is expressed
for existence statements involving proper names.
110

SPEAKING OF NOTHING

Our problem arose because we wanted on the one hand to make

it possible that one child saying Santa Claus does not exist may
express the same proposition as another who says Pre Nol nexiste
pas. But, on the other hand, our explanation of the truth conditions
for such statements in Rule (R) made them different for the two
cases. We have seen, however, that if the historical explanation
theory is correct, a difference in truth conditions without a shift in
proposition expressed can occur in any case with predicative statements. This can occur when there is a difference in names used
without a change in referent. So this seems to be a general feature of
the theorys treatment of names. When we turn to negative existence statements and Rule (R), however, we cannot give as a criterion for propositional identity sameness of referent. For, of course,
if true, the name in such a statement has no referent.
What we would like, still continuing with the example, is a
reason for saying that both children express the same proposition
that is at once in line with our theory and intuitively satisfying.
I want to suggest that we may find such a reason once more by using
the idea of a historical connection, that, in our example, it is the
blocks in the historical explanation of the use respectively of the
names Santa Claus and Pre Nol that are themselves historically connected. Once again, I do not have the resources to spell out
a general principle for what this historical connection must be, any
more than I did with the notion of a block itself. Yet in the example
before us, and others one can think of, our inclination to say that
people using different names express the same negative existence
proposition seems to be a matter of historical connection between
the blocks involved. In our example, it seems to me that the reason
we think both children express the same proposition is that the
story of Santa Claus and the story of Pre Nol, the stories passed on
to the two children as if they were factual, have a common root. And
111

E S S AY S O N R E F E R E N C E

if there were not this common history, I think we should rather hold
that the two children believed similar, perhaps, but not identical
falsehoods, for example, when the one attributed gifts to Santa Claus
and the other to Pre Nol and that they expressed different truths
when one said Santa Claus does not exist and the other said Pre
Nol nexiste pas.

VIII. CONCLUDING REMARKS

If this discussion has been on the right track, then at least the outline of a solution to some problems concerning nonexistence statements is available to the historical explanation theory. One point
emerged in the course of the last sections. We can perhaps point to
criteria for saying when two existence statements involving names
express the same proposition, but these criteria take a different form
from those for predicative statements involving names. In particular, it cannot be a matter of sameness of referent. For predicative
statements we were able to suggest a way of representing propositions, as ordered n-tuplets, but no obvious way of representing
propositions expressed by existence statements suggests itself. This
does not seem to me to count against the theory, since the notion of
a proposition is not, I think, a clear one that has established use outside of a theory. The fact that the representation suggested for predicative statements involving proper names has no counterpart for
existence statements, however, may account in part for the fact that
Russell took the alternatives for proper names to be either a Meinongian view or a concealed descriptions view. For the representation of propositions suggested is, I think, essentially Russellian and
either of these views of ordinary proper names would allow him to
apply it to existence statements.
112

SPEAKING OF NOTHING

NOTES
1. Earlier versions of this paper were read at a number of meetings and colloquia
and several important changes have resulted from those discussions. I am particularly grateful for detailed comments by Tyler Burge.
2. Reprinted in Logic and Knowledge, ed. by Marsh (1956), p. 47.
3. Why I am reluctant to use the word causal may become somewhat clearer further on, but the main reason is that I want to avoid a seeming commitment to all
the links in the referential chain being causal.
4. Translations from the Philosophical Writings of Gottlob Frege, ed. by Geach and
Black (1952).
5. Reprinted in Logic and Knowledge.
6. If we divide the theory into its negative aspects (see sec. III) and its positive (see
sec. IV), what the theory denies and the reasons for doing so have been, perhaps,
better delineated in the literature than the content of the positive theory. (This is
certainly true of my own contributions.) My papers dealing with various parts of
the theory as I see it are: Reference and Definite Descriptions, Philosophical
Review, LXXV (1966), 281304; Putting Humpty Dumpty Together Again,
Philosophical Review, LXXVII (1968), 20315; Proper Names and Identifying
Descriptions, Synthse (1970), reprinted in Davidson and Harman (eds.), Semantics of Natural Language (Dordrecht, 1972). By others, Saul Kripkes paper,
Naming and Necessity, in Semantics of Natural Language, is the most important
in that it gives not only arguments for the negative aspects of the theory, but also
a positive account (that, however, I do not altogether agree with).
7. The terminology, of course, is for convenience and not supposed to reflect a prejudgment that existence cannot be in some sense a genuine predicate.
8. See Middleton and Adair, The Mystery of the Horn Papers, William and Mary
Quarterly, 3rd series, IV (1947); reprinted in Winks (ed.), The Historian as Detective (New York, 1968).
9. The denial of Snow Whites existence, it should be noted, is in discourse about
actuality, while the statement that she enraptured a prince is in discourse
about fiction. (If the question of existence arose in discourse about fiction alone,
Snow White existed, whereas Hamlets fathers ghost, again presuming we are
talking about fiction, probably did not.) This does not disturb the point: no such
contrast can be made out for Jacob Horn; if Jacob Horn did not exist then there
are no true predicative statements to be made about him.
10. E.g., J. Searle, Proper Names, Mind (1958), and Speech Acts (Cambridge, 1968), Ch. IV.
11. See references in n. 10.
12. In Proper Names and Identifying Descriptions, op. cit.
13. That is to say, the first use of a name to refer to some particular individual might
be in an assertion about him, rather than any ceremony of giving the individual
that name. (In fact, my own name is an example: I discovered that colleagues

113

E S S AY S O N R E F E R E N C E

were pronouncing my last name differently than my parents doso, orally, they
referred to me by a different nameand I let it stand. But I was never dubbed by
that new name. I am sure that the first use of it was either an assertion, question,
or whatever about me and not a kind of baptism. And I think it is probable that
whatever audience there was knew to whom the speaker referred.)
14. See my Proper Names and Identifying Descriptions, op. cit.
15. Given that this is a statement about reality and that proper names have no descriptive content, then how are we to represent the proposition expressed?

114

[5]
SPEAKER REFERENCE,
DESCRIPTIONS, AND ANAPHORA

People refer and expressions refer. Let us call these phenomena

speaker reference and semantic reference respectively.1 What
connection exists between the two? The question has importance for
how we theorize about various referring expressions, demonstratives, for example. And it is fairly crucial for what significance should
be attached to a distinction I proposed some time back between what
I called two uses of definite descriptions, the referential and the attributive.2 At the time I thought of the distinction as having importance beyond a contribution to the understanding of how definite
descriptions work and the problems it seemed to pose for the theories of Russell and Strawson on that matter, precisely because it apparently showed the necessity of bringing in speaker reference for an
explanation of the semantic reference of certain expressions.
While the referential/attributive distinction proves, I believe, to
appeal to our intuitions, vagueness about the role of speaker reference
threatens its significance. Are there two uses of definite descriptions in
the sense of two semantic functions in one of which the description
First published in Syntax and Semantics, Vol. 9. Pragmatics. P. Cole, ed., New York:
Academic Press, 1978, pp. 4768.

115

E S S AY S O N R E F E R E N C E

conveys speaker reference and in the other not? Or is it rather that definite descriptions are used in two kinds of circumstances, in one of which
there is an accompanying phenomenon of speaker reference though it
has no effect on the semantic reference of the description? If the latter,
it is not clear what importance we should attach to the distinction in the
philosophy of language. It would not, for example, seem to have a bearing on the correctness or incorrectness of a semantic analysis of sentences containing definite descriptions such as Russell gives us.
I want to investigate this matter further in this paper and I offer
certain arguments derived from a consideration of the phenomenon
of anaphora to show that speaker reference cannot be divorced from
semantic reference.

I
It would no doubt be enlightening were I to begin by saying what
speaker reference consists in. We naturally gravitate to such expressions as what the speaker had in mind that echo locutions from
ordinary speechWhom do you have in mind? But what is it to
have someone or something in mind? Is it, for example, to possess a
body of descriptions that identify a particular person or thing? I will
not attempt, however, any general answer to this important problem. Instead, I will rely on what I think is the indisputable fact that
there is something corresponding to these locutions and go on to
ask where it fits into the theory of reference. I believe, in fact, that
even those who find little use for a notion of speaker reference acknowledge that the phenomenon occurs. Here, for example, is what
Peter Geach says about it in Reference and Generality:
Personal referencei.e. reference corresponding to the verb
refer as predicated of persons rather than of expressions
116

SPEAKER REFERENCE, DESCRIPTIONS, AND ANAPHORA

[I take it this is what I am calling speaker reference]is of

negligible importance for logic; and I mention it only to get it
out of the way. Let me take an example: Smith says indignantly
to his wife, The fat old humbug we saw yesterday has just been
made a full professor! His wife may know whom he refers to,
and will consider herself misinformed if and only if that person
has not been made a full professor. But the actual expression
the fat old humbug we saw yesterday will refer to somebody
only if Mr. and Mrs. Smith did meet someone rightly describable as a fat old humbug on the day before Smiths indignant
remark; if this is not so, then Smiths actual words will not have
conveyed true information, even if what Mrs. Smith gathered
from them was true.3

Whatever we finally decide about the nature of the distinction, I believe we have enough raw intuitions about it often to be able to tell,
given a sufficiently rich description of the circumstances, whether a
particular example falls on the referential or the attributive side.
Geachs description of this speech act pretty clearly makes it an example falling on the referential side.
Geach tells us that speaker reference is of negligible importance
for logic. I think we can fairly substitute for logic, semantics. In
any event the specific conclusion we are to draw from the illustration is clear: the speakers reference in no way determines the semantic reference of the expression he used. The two referents may,
of course, be identical, but they may also diverge. The person Mr.
Smith has in mind, the one he wants to inform his wife has been
made full professor, may not in fact fit the description Smith used,
may not be a fat old humbug whom they met the day before. In that
event, while Mr. Smith can be said to have referred to whoever it
was he had in mind, the uttered description did not. If it referred
at all, it would be someone who did (uniquely I suppose) fit the
117

E S S AY S O N R E F E R E N C E

descriptioneven someone whom Mr. Smith has forgotten entirely

and who never enters the heads of him or his wife during their conversation. Moreover, as Geach makes clear in the passage following
the one quoted, the truth or falsity of what was said by Smith, as opposed to what he may have wanted to convey to his wife, depends in
that event not on the properties of the person Mr. Smith has in mind
to inform his wife about but upon the properties of that person who
fits the uttered description. In this way Smiths words have, so to
speak, a semantic life of their own. I believe that Geach intends the
divorce of speaker and semantic reference to be a quite general
across-the-board matter. And I think he is not alone among philosophers in supposing this to be so.4
In Geachs example we may suppose that Mr. Smith believes that
a certain person he and his wife met the day before is a fat old
humbug; he believes, that is, that a particular person he has in mind
satisfies the description uttered. Perhaps we should add that this is
likely an active beliefone that he has currently before his mind.
Could this be all that speaker reference in such a case amounts to? If
it is, then the referential/attributive distinction would presumably
come to no more than this, that sometimes a speaker uses a definite
description with an accompanying (active) belief about somebody
or something that it fits the description and sometimes with no such
accompaniment. Construed thusly, the distinction would be real
enough, but, to be sure, of no interest for semantics and of little
interest, one would think, even for a total theory of speech acts.
A myriad of things accompany every speech actbeliefs, desires,
itches, and, of course, various external circumstances. A division of
speech acts or of a particular kind of speech act based merely on
cooccurrence or lack of it between the speech act and some other
event promises little for the theory of language. Even some connection between the accompanying item and, say, the content of the
118

SPEAKER REFERENCE, DESCRIPTIONS, AND ANAPHORA

utterance will not by itself warrant much attention. Some utterances

of the form All Xs are Ys, for instance, are undoubtedly accompanied by an active belief of the speakers to the effect that the world
contains too many Xs and others are not. The distinction is real
enough, but unlikely to find a place in any treatise on the philosophy
of language.

II
The account of speaker reference just given, however, is obviously
too thin. In the first place, it ignores the speakers intentions toward
his audience with respect to what he has in mind. In Geachs example, Mr. Smith intends his wife to recall some particular person met
by them the day before and through that, together with his statement, to become informed of a fact about that person. Moreover, it
is surely by his having produced the description the fat old humbug
we met yesterday together, perhaps, with the circumstances of the
utterance, that he intends she shall recollect that person. Such intentions can be crucial for the existence of speaker reference. Suppose
in some appropriate setting I say, The strongest man in the world
can lift at least 450 lbs. I might have as my grounds for stating this
some general considerations about the limits of human strength
with no belief about anyone in particular that he is the strongest
man in the world. In such an event there is, of course, nothing we
can identify as speaker reference. But my grounds might still be
those general considerations while I happen, in fact, to believe about
some particular person, Vladimir Jones, say, that he is the strongest
man. The addition of this belief should not make us talk of speaker
reference. Suppose, though, that my grounds for my statement are
that I believe of Vladimir that he is the strongest and I believe he can
119

E S S AY S O N R E F E R E N C E

lift 450 lbs. Still, those are my grounds and if I do not expect nor
intend that my audience shall recognize that I want to talk about
Vladimir and to become informed about his strength, we have no
reason to say that referred to Vladimir. What I have been describing,
of course, is a case of what I would call an attributive use of a definite
description. And what I am saying comes to this, that the referential/attributive distinction and the presence or absence of speaker
reference should be thought of as based on such speaker intentions
toward his audience or the lack of themnot on whether the
speaker believes or not about someone or something that he or it
fits the description.5 (Later I will give a more fundamental criterion
for speaker reference that will require modifications in what is said
here about speaker intentions for a certain range of cases, but for the
sort of example Geach gives us it will still be correct.)
The focus of interest of both speaker and audience in the sort of
situation Geach describes surely lies in the speakers reference. And
we have in language various resources for asking questions about
what the speaker refers to. The speakers audience will not infrequently fail to recognize what he has in mind to talk about just from
the uttered description and the context. It then becomes appropriate to ask him for more information. This may take the form of a
simple requestWhom do you mean? What are you talking
about?; or a request that the speaker provide further particulars in
order to distinguish among several things that seem to fit the description he usedWhich fat old humbug? We met a couple of
them yesterday; or a question asking whether the speakers referent
fits also some further descriptionDo you mean the boring old
geezer with the goatee? It also is not unnatural, we might note, to
speak as if the speakers reference were connected to the expression
uttered, for example, Whom do you mean by the fat old humbug
we met yesterday?
120

SPEAKER REFERENCE, DESCRIPTIONS, AND ANAPHORA

The very form of these questions makes it hard to deny that they
concern speaker reference. And in fact they are only appropriate
when the speaker did intend to convey a reference. In the example
given above in which when uttering The strongest man in the world
can lift 450 lbs., the speaker had no intention of having his audience
recognize him as referring to anyone in particular, it would have
been out of place for his audience to ask him, for example, Whom
do you mean? or Are you talking about the Russian who won the
gold medal? or By the strongest man in the world are you referring to Vladimir Jones? To do so would be to misread the situation
and the speaker in turn may correct the misimpression by saying,
for example, I dont mean anyone in particular, just whoever is the
strongest. Now if these questions were not about the speakers reference but about the denotation of the description (whatever satisfied it uniquely), then there would be no reason why they should be
appropriate only in the presence of speaker reference.

III
If the audience questions discussed in the last section are questions
about the speakers reference, as they certainly seem to be, I believe
we can also show that in the circumstances in which they are appropriate, audience responses can also contain expressions whose referents are determined by what the original speaker had in mind; that
here, at least, we have semantic reference determined by speaker reference. And that, I believe, was what Geach wished to deny took
place when he said that we could conveniently ignore speaker reference when pursuing logic or, as I would rather say, semantics. But
before showing this I want to introduce a notion I will have occasion to use in what follows.
121

E S S AY S O N R E F E R E N C E

In recent work Charles Chastain6 has pointed to the importance

for the theory of reference of what he calls anaphoric chains. By
these he means sequences of singular terms that are such that if one
of them refers to something, they all do. A pronoun with an antecedent in the same sentence provides the simplest example:
(1) Mary is coming to town and John will meet her.

The pronoun her is coreferential with the name Mary, and the
two form what Chastain calls an anaphoric chain. There can be more
than one chain in a single sentence, of course:
(2)

John went to meet Tom and he waited for him almost two
hours.

Here he is linked with John and him with Tom. If we ask how
we know which pronouns belong to which chains, I imagine that at
least in some instances it has to do with rules of syntax discoverable
by linguistsperhaps something to do with order of occurrence or
possibly with a linking of subject to subject and object to object.
The phenomenon I want to exploit is not confined to the limits
of single sentences. Chains can pass over sentence boundaries:
(3) John went to meet Tom. Does anyone know whether he has
brought him back yet?

Here the pronouns of the second sentence take as antecedents

names in the first. Or, to be more precise, they would do so on a
natural interpretation and within the same stretch of discourse. If
the sentences are uttered on widely separated occasions, for example, or if the utterance of the second sentence is accompanied by
122

SPEAKER REFERENCE, DESCRIPTIONS, AND ANAPHORA

demonstrations giving the pronouns independent reference, then

no such linkage will probably exist.
More important, chains can pass over speaker boundaries. In
(3) the second sentence could have been uttered by someone other
than the person who uttered the first. If that person is among the
first speakers audience, then the pronouns of the second sentence
may be parasitic on the names used by the first speaker forming
anaphoric chains with them. We will make use of this fact in what
follows.
In general, there will be one member of an anaphoric chain that
determines the referent of each of the other members and without
which the other members would be left dangling. In the examples so
far given proper names have served this function and pronouns have
been parasitic on them. I will assume, for simplicity, that this determining member of the chain occurs first and call it the antecedent.7

IV
Let us say that a definite description is uttered in a referential context when speaker reference exists relative to it. So far, all this will
mean is that the speaker intends to refer to something and intends
his audience to recognize his reference in part through his having
used that definite description. Thus in introducing this terminology
we are not begging the issue, for example, as against Geach. Similarly, a definite description will be uttered in an attributive context
when speaker reference relative to it is absent. Definite descriptions
in attributive contexts can serve as the antecedent in an anaphoric
chain. Imagine the following to be uttered in an attributive context
relative to the definite description that forms the subject of the first
sentence:
123

E S S AY S O N R E F E R E N C E

(4)

The strongest man in the world can lift at least 450 lbs. He can
also win a tug-of-war with a jackass.

In a suitable context the pronoun he in the second sentence forms

an anaphoric chain with the definite description in the first; they
both refer to the same thing if they refer at all. And what they both
would refer to is undoubtedly the denotation of the definite descriptionthat person, in this example, who uniquely fits the description. The pronoun, in fact, can be regarded as what Geach called a
pronoun of laziness8simply a means of avoiding repetition of the
description. Now let us return to Geachs example, which, we are
assuming, contains a definite description in a referential context,
and extend the dialogue a bit.
(5)

[Mr. Smith] The fat old humbug we met yesterday has just
been made a full professor. He must have bamboozled the
committee.

We seem once more to have an anaphoric chain. Certainly there is

some link between the pronoun in the second sentence and the definite description in the first. And one would suppose they are coreferential, if they refer at all. If so, what do they both refer to? On the
view that speaker reference and semantic reference are to be kept in
their separate realms, the answer is clear: they both refer to the denotation of the description, if it has one. What this means is that if
the person Mr. Smith has in mind, his reference, does not in fact fit
the description, the fat old humbug we met yesterday, what his
second sentence expresses will be true just in case there was a
(unique) fat old humbug he and his wife met the day before and
who bamboozled the committee. And this would be so even if that
person never enters into the Smiths heads. If semantic reference is
124

SPEAKER REFERENCE, DESCRIPTIONS, AND ANAPHORA

to be kept untainted by speaker reference, the pronouns in such anaphoric chains must be kept pure also.
But now recall that an audience may ask questions about the
speakers reference when it exists. So the conversation between the
Smiths might have proceeded in this way:
(6)

[Mr. Smith] The fat old humbug we met yesterday has just
been made a full professor.
[Mrs. Smith] Do you mean the funny little man with the
goatee?

But her query, which we have argued concerns Mr. Smiths

reference, could have been cast into a sentence employing a
pronoun:
(7)

[Mrs. Smith] Is he the funny little man with the goatee?

If this is just another way of posing the same question, as I think it is,
it too is a question about the speakers reference. But then the pronoun he must refer to the referent of the speakerto whatever
person Mr. Smith was referring to.
An even more striking way of getting this result comes from considering the fact that an audience will sometimes disagree about the
applicability of the description the speaker used. And such disagreement can hardly be a question of whether the description applies to
its own denotation. Thus in the Smiths conversation, Mrs. Smith
might have said at a certain point:
(8)

[Mrs. Smith] I dont think hes fat; hes just large boned. And as
for his being a humbug, he seemed quite genuine and above
board.
125

E S S AY S O N R E F E R E N C E

It would make nonsense of Mrs. Smiths comments to suppose that

the third-person pronouns they contain are pronouns of laziness
standing in for Mr. Smiths original description or that their referent
is to be the denotation, if it has one, of that description. Surely it is
Mr. Smiths reference that is in question and which determines the
referent of these pronouns.
The Smiths dialogue might have gone like this:
[mr. smith] The fat old humbug we met yesterday has just been
made a full professor. He must have bamboozled the
committee.
[mrs. smith] Is he the one with the funny goatee?
[mr. smith] Hes the one I mean.
[mrs. smith] I dont think we met him yesterday. Wasnt it
Friday?
[mr. smith] I think youre right. He was coming from a faculty
meeting, so it must have been a weekday.

We have an initial definite description followed by a string of pronouns in subsequent utterances. Some of those surely must take as
their referent the man Mr. Smith has in mind, the speaker reference
rather than whoever is denoted by the description, supposing it
does have a denotation.
Now the (third-person) pronouns in this discourse fragment
seem to form an anaphoric chain and the initial definite description
seems to be the antecedent. If this were so, it would follow that the
speakers reference determines the semantic reference throughout.
I suppose that it would be possible to maintain that despite these
data the truth or falsity of the utterance containing the definite
description depends upon the properties of the denotation of
the description, if it has one, and that its semantic referent is its
126

SPEAKER REFERENCE, DESCRIPTIONS, AND ANAPHORA

denotationeven if this is not so for subsequent pronouns that

seem on the surface to be anaphorically linked with it. There would
certainly be complications for such a view. In the conversation as
recorded Mr. Smith said He must have bamboozled the committee immediately following his utterance containing the definite description. Is this occurrence of the pronoun he anaphorically
linked with the definite description? If so, would it no longer be had
he produced this utterance after the Smiths had agreed that the original description was inapplicable?
In any event, some ground will have to be ceded. It will not be
possible to set aside speaker reference as of no importance in the
determination of semantic reference. For if speaker reference does
not determine semantic reference in certain instances of the use of
definite descriptions, it does for subsequent pronouns in some
stretches of discourse containing them. The referential/attributive
distinction, rested on the notion of speaker reference, will have semantic importance because it will mark the dichotomy between occurrences of definite descriptions that can initiate strings of
pronouns whose reference depends upon the speakers reference
and those that do not.

V
I now want to turn to some further facts about definite descriptions
and anaphoric chains that bear on the topic of speaker reference.
Philosophers have often contrasted definite and indefinite descriptions. Definite descriptions, it is said, are used to speak of some one
particular thing while indefinite descriptions are not. In the recent
essay already mentioned Charles Chastain9 has challenged this as a
universal rule. Without for the moment assessing this interesting
127

E S S AY S O N R E F E R E N C E

conclusion of his, I want to examine some of the facts that he uses in

arriving at it.
It is certainly true that sometimes a use of an indefinite description in no way involves a reference to a particular thing while a shift
to the definite article in the same context would indicate such a reference. We can correctly answer the question Have you ever seen
an elephant? by saying, Yes, I have seen an elephant even if one
has in fact seen many elephants. And in so answering, no one of the
elephants seen has been singled out. But a particular elephant is presumably referred to if the question is Have you seen the elephant?
and the answer is Yes, I have seen the elephant.
The definite article indicates in some sense that a particular reference is being made,10 but how does that get accomplished? We
know that according to Russells theory sentences containing definite descriptions (where they are what he called primary occurrences) assert in part that one thing uniquely has the properties
ascribed by the description. Thus, provided that this part of what is
asserted is true, some one particular thing gets singled out. This
contrasts with the Russellian treatment of indefinite descriptions
according to which sentences containing these expressions assert
the existence of members in a certain class but do not, relative to
the occurrences of indefinite descriptions, assert that there is a
unique member.
There are two well-known difficulties with this treatment of definite descriptions as far as ordinary speech goes. One is that the definite article can accept plural noun phrasesI have seen the
elephants. The second is that definite descriptions perhaps most
frequently found in ordinary discourse are too indefinite to allow
us to suppose that their users intend to assert that some particular
thing alone has the properties they ascribe. There are a large number
of elephants in the world even today and most people who might
128

SPEAKER REFERENCE, DESCRIPTIONS, AND ANAPHORA

have occasion to use the sentence I have seen the elephant are
surely aware of this.
It is not clear how an account along strictly Russellian lines
would get over the first of these difficulties, but a view suggested by
Zeno Vendler11 might provide a beginning. He says of the assertion
I know the men who fought in Korea that it would imply that . . .
in some sense or other, I know all those men. And It transpires,
then, that the definite article marks the speakers intention to exhaust the range determined by [the description].12 Singular definite
descriptions would be a limiting case where it is asserted or implied
that the range is limited to a single item. (Later I will try to show that
this account fails.)
The second difficulty remains even on Vendlers view. That the
descriptive content of many of the definite descriptions we actually
utter is too meager to suppose that we mean to imply that they fit
something uniquely has frequently not been taken seriously enough.
The usual attempt at saving the situation is well known. The idea is
that in ordinary speech we rely on attendant circumstances to supply
implicit qualifications on the description actually uttered. So, I have
seen the elephant uttered at the San Diego Zoo, with speaker and
audience aware of their location, succeeds in being an utterance about
a particular elephant because the speaker presumes that his audience
will take his description as qualified by the restriction at the San
Diego Zoo or some other qualification such as at this zoo that does
the same work. This is not without plausibility for a large number of
cases, but I believe it will not work as a general rule. Interestingly
enough, nowhere do we find more intransigent examples than among
the sorts of uses of definite descriptions investigated by Vendler that
led him to the view of the definite article just mentioned.
Definite descriptions not only serve as antecedents in anaphoric
chains, they also can act as later links in such a chain. Chastains
129

E S S AY S O N R E F E R E N C E

essay points to the importance of such occurrences of definite descriptions for the theory of reference. Now, interestingly, given the
great contrast that is supposed to exist between definite and indefinite descriptions, the antecedents of such anaphoric chains turn out
to be sentences containing indefinite descriptions. Consider the following discourse fragment:
(9) A man carne to the office today. He tried to sell me an
encyclopedia.

The pronoun beginning the second sentence seems clearly to

be connected in some fashion to the first sentence, though
whether anaphorically linked in particular to the indefinite description a man is another question.13 The pronoun can be
replaced without any shift in what we would understand as
being said by a definite description. In fact, as both Chastain
and Vendler point out, a definite rather than an indefinite description is required to preserve the anaphoric linkage. We
cannot read the following discourse fragment as saying that the
same man who carne to the office tried to sell the speaker an
encyclopedia; in fact, it is most naturally read as implying that
there were two distinct men:
(10) A man carne to the office today. A man tried to sell me an
encyclopedia.

The required definite description can be formed from the preceding

sentence in one of two ways, either from a generic noun obtained
from the indefinite description or from such a generic noun modified by adjectives or a restrictive relative clause obtained from what
was predicated in the preceding sentence:
130

SPEAKER REFERENCE, DESCRIPTIONS, AND ANAPHORA

(9a) A man carne to the office today. The man tried to sell me an
encyclopedia.
(9b) A man carne to the office today. The man who carne to the
office today tried to sell me an encyclopedia.
(11) A man carne to the office today carrying a huge suitcase. It
contained an encyclopedia.
(11a) A man carne to the office today carrying a huge suitcase. The
suitcase contained an encyclopedia.
(11b) A man carne to the office today carrying a huge suitcase. The
huge suitcase carried by the man who carne to the office
contained an encyclopedia.

In some of these cases repetition of information makes the discourse

sound like the awkward language of a childrens first reader; the
equivalences nevertheless hold.
There are several interesting features of this use of definite descriptions. One is that having used an indefinite description we immediately become entitled to use a definite description. This should
give one pause before accepting the usual account of indefinite descriptions. In fact the usual account simply will not fit the use of the
indefinite description here, as is suggested by trying to substitute for
the initiating sentence in our examples one that explicitly asserts
what that account says should be asserted:
(12)* At least one man carne to the office today. The man . . .
(He . . .)
(13)* One or more men carne to the office today. The man . . .
(He . . .)

The trouble seems to be that the pronoun or definite description in

the second sentence can no longer have an anaphoric link with the
131

E S S AY S O N R E F E R E N C E

first sentence.14 If the sentences containing the indefinite

description(s) in our examples do not assert merely that there is at
least one thing having certain properties, nor obviously do they
assert that there is exactly one thing having certain properties.
One would like to say that the initial sentences in these examples, the ones containing the indefinite description, serve to introduce a particular thing, a man, a suitcase, or whatever, and that this
is what justifies the subsequent use of a pronoun or a definite description. But in what sense is a particular thing introduced?
One initially attractive idea is that the introducing sentence as a
whole, not just the indefinite description, serves, if everything goes
right, to identify an individual by providing a unique description.
(It will be recalled that a subsequent definite description linked with
the introducing sentence can be formed from the sentence as a
whole by means of a restrictive relative clauseThe man who
carne to the office. . . . from A man carne to the office.) There are
many cases in which this has plausibility. Sometimes the introducing sentence yields a description that by its very nature could be
satisfied by one thing at most, as when the description contains a
superlative or a definite ranking in some ordering:
(14)

This set contains a least member. The least member of this

set. . . .
(15) A rank amateur carne in first in the Podunk Open today. The
rank amateur who carne in first. . . .

The speaker may assume certain background information available

to the audience that would limit the things fitting the description to
a single one:
(16) I believe she has a husband. He must be kind to her.
132

SPEAKER REFERENCE, DESCRIPTIONS, AND ANAPHORA

A background assumption that one can only have one husband (at a
time) may operate. Finally, we may be able to view a speaker as
sometimes implying (or asserting) that only one thing has the properties ascribed by the description:
(17)

A loose connection is causing the hum in the television. The

loose connection must be in the audio section.

But the problem is that in many cases we cannot suppose that the
speaker believes or intends the description plus any background assumptions to pick out something uniquely. Our initial examples
show this. (9), (9a), or (9b) might naturally begin an anecdote
about an event at ones office told to friends who know little or nothing about what goes on there. Even if one used the fully expanded
description, as in (9b), A man carne to the office today. The man
who carne to the office today tried to sell me an encyclopedia, the
speaker is surely not committed to, nor does he intend to suggest
that just one man carne to the office that day. Nor would he suppose
that there are background assumptions shared by the audience that
would allow them to recognize a particular man. There is a particular man presumably about which the speaker is talking, but that
person is not identified by the descriptions used plus the circumstances of utterance. An introducing sentence, in fact, may supply
no more than a generic noun:
(18) Once there was a king. The king rode a white horse.

This might begin a historical anecdote about a particular king. But

which king is not determined by the descriptive content and the
audience need not be presumed to have background assumptions
that would narrow down the range.
133

E S S AY S O N R E F E R E N C E

Similar examples employing plural definite descriptions can

easily be constructed:
(19) Some men carne to the office today. They wanted to sell me
a philosophical computer.
(19a) Some men carne to the office today. The men. . . .

Several groups of men may have come to the office today, although
only one is being talked about. In suitable circumstances the speaker
obviously may have no presumption that his audience can from any
shared assumptions, contextual clues, etc., supply further qualifications that distinguish the group talked about from the others.
These occurrences of definite descriptions in very prosaic discourse fragments show, it seems to me, that by using the definite
article a speaker need not be signaling any intention to exhaust the
range of the description, as Vendler suggests, nor even to exhaust
the range of an augmented description that his audience could be
presumed to be capable of deriving from the context of utterance.
Not only does this seem to show that Vendlers suggestion must be
mistaken, but it indicates that the usual way of attempting to save a
Russellian account of definite descriptions for ordinary discourse in
the face of the indefiniteness of the descriptive content found so frequently cannot succeed either.
Yet in these examples some particular person or persons are being
talked about and the definite descriptions and pronouns seem surely
to have particular semantic referents. If the descriptive content of the
uttered descriptions even augmented by background assumptions,
etc., are insufficient to determine the referents, how is this possible?
My answer will not be unexpected. The speaker having some person
or persons in mind to talk about can provide the needed definiteness.
Once more, then, we have a series of instances in which speaker
134

SPEAKER REFERENCE, DESCRIPTIONS, AND ANAPHORA

reference appears necessary to provide semantic reference. Here, not

just to provide the right reference, but to allow for reference at all.

VI
With the examples of the last section we have now considered two
ways in which definite descriptions can be introduced into a discourse: they may simply appear tout court, as did Mr. Smiths the fat
old humbug we met yesterday, in which case they may be followed
by pronouns apparently anaphorically linked to them; or they may
appear after a sentence containing an indefinite description that in
some fashion serves as an introduction, as in the examples of the last
section. In the latter case they may be replaced by pronouns. In both
cases we can distinguish between referential and attributive contexts depending upon whether or not there is speaker reference. But
it may have been already noticed that there has been an important
difference in the way in which I described the speakers attitude
toward his audience in referential contexts between the two. In the
first part of the paper, where we were considering definite descriptions appearing tout court in a stretch of discourse, when there was
speaker reference I said that the speaker intended his audience to
recognize, partly by means of the description used, what his reference was to. And this, I believe, is a correct account, for instance, of
the Mr. Smith example. But in the examples of the last section,
where the definite description appears after an introducing sentence
containing an indefinite description, the speaker need not be assumed to have any such intention toward his audience even where
I claimed there was speaker reference. In telling the anecdote about
what happened at the office beginning with A man carne to
the office today and continuing, for example, with The man . . . ,
135

E S S AY S O N R E F E R E N C E

the speaker need not expect nor intend his audience to recognize
anyone as the subject of the story. What then makes such an example a case of speaker reference at all?
Of course I have argued that in such examples some particular
person or thing is being talked about and the pronouns and definite
descriptions that occur subsequent to the introducing sentence and
linked with it seem surely to refer to that person or thing. But I believe we can put the point more precisely in terms of what the
speaker intends concerning the truth conditions of his utterances:
that he intends that truth or falsity shall be a function, in part, of the
properties of the person or thing he has in mind.
Suppose that Woodward and Bernstein in their account of their
investigation of the Nixon White House had said at a certain point:
(20)

We now had a telephone call from a man high in the inner

circle. He asked us to meet him at a certain suburban garage
where he would give us confirmation of some of our conjectures. We later decided to give the man the code name Deep
Throat.15

Woodward and Bernstein never, of course, reveal to their readers

who this high official is, and while some have tried to guess his identity and some may be certain of it, the authors did not intend that
there should be recognition. The truth value of what they would
have expressed in (20), however, depends upon the properties of
the person they are writing about. Suppose that the man they had in
mind did not ask them to meet him at a garage but at a certain bus
stop. Then the second sentence in (20) expresses a falsehood. Suppose they never decided to give him the famous code name, but put
this tidbit in the book because it seemed amusing. Then the third
sentence expresses a falsehood.
136

SPEAKER REFERENCE, DESCRIPTIONS, AND ANAPHORA

Let us complicate this a bit by introducing a second mystery

person. Suppose that the man that Woodward and Bernstein had in
mind when writing the passage was not ever referred to by them as
Deep Throat, but that in fact they had given this code name to a
second informant who entered into their investigations. So someone did possess the property attributed to the man in the third
sentence. Still, I believe it is clear that that would not save the sentence from expressing a falsehood. The second informant was not
being referred to here.
This determination of truth value by the properties of the speakers referent extends also to the initiating sentence. If the man Woodward and Bernstein are writing about did not call them, but, say, left
them a note in Woodwards copy of the New York Times, the first
sentence in (20) would express a falsehood. And it would not be
saved if some man high in the inner circles did in fact call them at the
time in question.
We have then a distinction among the cases in which definite
descriptions and pronouns anaphorically linked are preceded by an
introducing sentence containing an indefinite description comparable to the referential/attributive distinction. The distinction rests
on a notion of speaker reference, but one we were able to explicate
in terms of the truth conditions of the utterances. It would be nice
were we now able to do this across the board. The barrier in the way
of giving such an account for definite descriptions introduced tout
court, as in Geachs example of the conversation between the Smiths,
however, is that we did not show in any conclusive fashion that the
semantic reference of such a definite description is determined by
the speakers reference, when such exists, although we did, I believe,
show that pronouns apparently anaphorically linked to it were. And
to apply without further ado the present account of speaker reference would obviously beg the question. We have, however, been
137

E S S AY S O N R E F E R E N C E

able in the last two sections to link speaker reference with the
semantic reference of certain occurrences of definite descriptions,
so these expressions are not slipping through our net altogether.

VII
A diesis to be found in somewhat different forms in both Chastains
and Vendlers essays would, if correct, allow us to apply the results of
the last two sections directly to the case of definite descriptions introduced tout court. For both hold that the occurrence of definite
descriptions in anaphoric chains initiated by a sentence containing
an indefinite description is fundamental, and that when we find one
introduced into a discourse apparently on its own hook, either it is
not a genuine singular term or we should view it as derived from an
understood, presupposed, or deleted initiating sentence containing
an indefinite description. What we have been treating as two different sorts of cases are really not distinct.
If we accept their view, we could give the following sort of argument. Suppose, for example, that we are talking to someone we have
reason to think will be able to single out a particular person we have
in mind and we say:
(21)

The man who carne to the office today tried to sell me an

encyclopedia.

Now on the Vendler/Chastain view the definite description here

really stands as a member of an anaphoric chain, the antecedent of
which is either presupposed or in a covert discourse. The fully
explicit discourse fragment from which this should be viewed as
derived would be:
138

SPEAKER REFERENCE, DESCRIPTIONS, AND ANAPHORA

(22) A man carne to the office today. The man who carne to the
office today tried to sell me an encyclopedia.

But we have already argued concerning (22) that the truth conditions of these utterances would depend upon the properties of the
speakers referent. We should therefore have to say the same thing
about (21). And this, of course, would be directly contrary to the
view represented at the beginning by Geach that speaker reference
has nothing to do with the truth conditions of the utterance.
I am naturally not antagonistic toward the Vendler/Chastain
view, but I am also not sure, aside perhaps for some considerations
about simplicity and unification of the treatment of definite descriptions, that there is an argument that shows that the fundamental
grammatical construction is that of a definite description anaphorically linked to an antecedent sentence containing an indefinite description. Nevertheless, I believe we can give something like the
argument just mentioned.
First let me say something about how I view these two constructionsthe definite description introduced tout court and the definite
description anaphorically linked. In referential contexts, those where
speaker reference is present, the choice of which construction to use
is, I believe, a matter of the speakers expectations and intentions
toward his audience: does he expect and intend that they will recognize who or what he has in mind? If he does, then he will use a definite
description with no further introduction; if not, he will begin with an
introduction via an indefinite description. What the latter does, so to
speak, is to announce that the speaker intends to speak about a particular thing or particular things following under a certain descriptionfor example, he intends to speak of a particular king or a
particular man or particular men who carne to the office. There is no
implication in most cases that he will speak of everything falling under
139

E S S AY S O N R E F E R E N C E

the description. Having done this, he can then go on to use a definite

description or a pronoun to refer to what he wants to talk about.
Where the speaker intends and expects his audience to be able to recognize what he speaks about from the description used (plus attendant circumstances), such an introduction is otiose. In fact, it would
often be downright misleading because it would strongly suggest that
the speaker did not believe his audience to be in a position to recognize this reference. Suppose we have just been talking to our colleague
Joe who had with him an obviously close acquaintance. After they
leave I say to you, Joes friend seemed a bit daff y. It would be no
doubt obvious whom I was talking about. Had I said instead, however,
Joe has a friend. He seemed a bit daff y, I think you would be puzzled
because my use of the introductory sentence suggests that you will
not recognize my reference and that conflicts with the supposition
that I am referring to a person we both have just met.
This account does not imply that the one construction is grammatically derived from the other, although it is consistent with that
possibility. But if it does represent the difference between the two as
far as why one might be used in some circumstances and the other in
other circumstances, it also suggests that there should be no real difference in truth conditions or semantic references. Consider the indignant Mr. Smith again. Suppose he has two friends at the university,
one of whom is a close confidant and familiar with the politics of the
situation, the other of whom has just returned from a lengthy sabbatical in the nether regions of central Asia. To the second he says:
(23)

A guy in the English Department has been getting chummy

with the Dean. He, the guy in the English Department who
has been chummy with the Dean, just got promoted to full
professor. Shows what things have come to since you were
away.
140

SPEAKER REFERENCE, DESCRIPTIONS, AND ANAPHORA

About such a case we argued that a particular person is being referred to but that the particularity cannot be expected to result from
the descriptions actually employed applying to one person alone
(the speaker may know full well that several members of the English
Department have been getting chummy with the Dean), nor from
background assumptions (the speaker does not expect that his
newly returned colleague will possess facts that together with the
speakers descriptions will enable him to isolate a particular individual). Hence, we must go to the fact that the speaker has a particular person in mind, to speaker reference, to obtain the particularity.
And it is properties of that individual that determine the truth or
falsity of what the speaker uttered. Now to his more au courant colleague, the speaker might say,
(24)

The guy in the English Department who has been getting

chummy with the Dean has just gotten promoted to full professor. Shows what things are coming to.

In this sort of case we were able to argue that subsequent pronouns

anaphorically linked with the definite description have as referent
the speakers referent. In particular, Smiths colleague might reply:
(25) You are wrong about him. Ive investigated and he is far from
chummy with the Deanin fact they had a quarrel just last
week.

And the pronouns here could hardly have as a referent whatever

member of the English Department has been getting chummy with
the Dean, since there is an explicit denial of that description.
But now if the only factor that affects Mr. Smiths shift from using
an introductory indefinite description to introducing the definite
141

E S S AY S O N R E F E R E N C E

description tout court is his expectation or lack of it about whether

his audience will be able to recognize his reference, there can be no
reason why the semantic facts about reference in the two cases
should be different. And we can then, after all, apply the arguments
that show semantic reference to be determined by speaker reference
from the one sort of case to the other.

VIII
I have not touched upon the question of syntactic or semantic ambiguity. Are sentences containing definite descriptions that can be
used in either referential or attributive contexts ambiguous? Are
there two senses of the definite article? I have not done so in the first
place because it seemed to me I could argue for the semantic significance of speaker reference and the referential/attributive distinction without, directly at any rate, tackling the problem of ambiguity.
Secondly, the problem of ambiguity may require, for a solution, that
another question I have put to one side, what it is to have something
in mind, be answered first. And thirdly, we may need to work out
more clearly than we have just what we mean to attribute when we
speak of semantic or syntactic ambiguity, and that lies beyond the
scope of this paper.
It might be thought, however, that if the position of this paper
were correct that an ambiguity in the definite article would at least
be suggested and that it is intuitively very implausible to suppose
such an ambiguity. So that, until the question of ambiguity is resolved, a real doubt remains about whether that position can be correct. To this worry I will point out here that the situation of the
definite article, given the position of this paper, remains in no worse
shape in regards to ambiguity than several other operators and
142

SPEAKER REFERENCE, DESCRIPTIONS, AND ANAPHORA

connectives in natural languages already are after philosophical

scrutiny. We have already seen that sentences containing indefinite
descriptions, while they may often be used to assert existential generalizations, are also used to introduce particular individuals, as in
some of the examples in this paper where the sentence initiates an
anaphoric chain. The same sort of worry about ambiguity should
arise here. The if . . . then_____ construction provided another
example. We know that such a construction (in the simplest form of
declarative sentence) can be used to express what we might call a
true conditional, but also can be used to assert the proposition expressed by the consequent where the antecedent expresses
something like a condition the speaker presupposes for the asserted
proposition to be of interest or significance in the circumstances. So
I might tell someone to whom I am lending my car for a trip, If it
snows in the morning, there will be chains in the trunk. This might
be used to assert a true conditional; I might, for example, be informing my audience that I will see to it that there are chains in the trunk
on condition that I find it snowing in the morning. But I might also
be asserting that there will be chains in the trunk, this being of interest, of course, only if it is snowing. (Similarly, Austins There are
cookies in the cupboard, if you want some would normally be taken
as an assertion of the consequent, but might be used by ones fairy
godmother, who makes things true providing one wants them to be,
to assert a true conditional.) I am not certain that just because of
these data we should declare an ambiguity in the if . . . then_____
construction, although there certainly is a semantically interesting
distinction to be drawn.
If we abstain from passing a ruling about ambiguity, can we draw
any conclusions about the Russellian theory of definite descriptions? I am inclined to say that it cannot provide the proper analysis
for the referential context in the sense of telling us what proposition
143

E S S AY S O N R E F E R E N C E

is expressed. To fully sustain this would, I believe, require saying

something more about what it is to have an individual in mind, for
I can foresee the possibility of trying to obtain a Russellian proposition from descriptions in the speakers mind. Let me then just say
something about one way that may immediately occur to one attempting to obtain a Russellian proposition in the referential case.
The main problem brought up in the paper for a Russellian analysis in referential contexts was that of providing a unique denotation when neither speaker nor audience could be expected to believe
that the description actually uttered was true of just one individual.
It was argued that the usual way of trying to handle this problem
suggesting that the context of utterance would supply further distinguishing descriptionswould not in general work. It was at that
point that we turned to what the speaker had in mind. But it might
be thought that we could treat this as a special case of context supplying additional descriptions by supposing that what is implicit in
such cases is just the additional description which I, the speaker,
have in mind, or some close approximation. In the Geach example,
Mr. Smith could be represented as having said,
(26)

The fat old humbug we met yesterday [and whom I have

in mind] has just been made a full professor.

What occurs in brackets was not uttered by Mr. Smith, but intended
to be supplied by the audience, Mrs. Smith, by its recognition of
what Mr. Smith is up to. (26), in turn, would be analyzed, would be
taken as expressing a proposition, along Russellian lines.
The problem I envisage with this suggestion is that it gives us the
wrong proposition. The truth value of the Russellian proposition expressed by (26) would be partly a function of whom the speaker had in
mind. If the Smiths met two fat old humbugs yesterday, one of whom
144

SPEAKER REFERENCE, DESCRIPTIONS, AND ANAPHORA

got promoted and the other of whom did not, the Russellian proposition expressed by (26) would be true or false depending upon which
Mr. Smith had in mind. But, I should like to say, whom Mr. Smith had
in mind determines what proposition he expressed, not whether the
proposition he expressed in any case is true or not. To put it shortly,
while Mrs. Smith might say, You could have prevented that by expressing your views to the committee, it would be absurd for her to say,
You could have prevented that by having someone else in mind.
Much more needs to be said, of course, about the topics mentioned in this last section. My aim, however, has been to argue that
whatever the final view about them, we cannot divorce speaker reference from semantic reference.16

NOTES
1. The research for this paper was done in part under a grant from the National
Endowment for the Humanities and while a Fellow at the Center for Advanced
Study in the Behavioral Sciences.
2. Reference and Definite Descriptions, Philosophical Review 75 (1966): 281304
and Putting Humpty Dumpty Together Again, Philosophical Review 77 (1968):
203215.
3. Reference and Generality (Ithaca, 1962), p. 8.
4. Cf. Saul Kripke, Naming and Necessity, in Semantics of Natural Language, eds.
D. Davidson and G. Harmon (Dordrecht, 1972), fn. 3, p. 343.
5. As stated this is somewhat misleading. As will be seen in Section VI, I take such
a speakers intention toward his audience to be only a sufficient, not a necessary,
condition for speaker reference.
6. Reference and Context, Language, Mind, and Knowledge, ed. K. Gunderson,
Minnesota Studies in the Philosophy of Science, Vol. VII (Minneapolis, 1975),
pp. 194269. As will be obvious, I am especially indebted to this essay.
7. His smile is Johns best feature, for example, is an instance where the determining member occurs later.
8. Reference and Generality, p. 124f.
9. Reference and Context.
10. In this paper I use the term reference and variants on it in connection with
definite descriptions both in referential contexts and in attributive. This is not

145

E S S AY S O N R E F E R E N C E

intended to imply that there may not be a big difference between the two situations, as Reference and Definite Descriptions suggested.
11. Singular Terms, in Linguistics in Philosophy (Ithaca, 1967), Ch. 2.
12. Singular Terms, pp. 5051.
13. Chastains view is that it is. He thus holds that a man in the first sentence of
the discourse fragment would be a singular term with the same referent as the
pronoun in the second. I am not wholly convinced of this. In some sense the
first sentence introduces a particular person and, as I will argue, the truth
value of what it expresses depends upon the properties of a particular man.
I thus agree with Chastain that the sentence does not express an existential
generalization of the sort we are often given as the reading of sentence containing indefinite descriptions. But I am not sure that that is enough to show
that the indefinite description itself is a singular term. My reluctance stems
from wondering whether it is not possible for a sentence to introduce, so to
speak, an individual that subsequent pronouns refer to without itself containing an expression that refers to the individual. Perhaps the following would be
such a case: I had steak and lobster at Delmonicos before the play. It was a
wonderful meal.
14. The anomaly of the preceding discourse fragments is not the main argument for
the conclusion. That comes from a consideration of the truth conditions as will
be seen later.
15. All the Presidents Men (New York, 1975). The example, of course, is not an actual
quotation.
16. Too late for treatment in this paper, Saul Kripkes paper Speaker Reference and
Semantic Reference (this volume, pp. 627) was published in Midwest Studies in
Philosophy 2 (1977): 255276. My paper was then in the final stages of production for a different publication (Syntax and Semantics, Vol. 9, edited by Peter
Cole, Academic Press, 1978). The two papers now appear together in the present
volume. I had heard a version of Kripkes paper given as a talk, but having no
written copy to consult nor transcript of the talk, I did not want to rely on
memory. Kripkes paper deals with the same topic as mine and comes to an apparently contrary conclusion, although I believe he is more concerned with the
question of ambiguity than I have been here.

146

[6]
THE CONTINGENT
A PRIORI AND RIGID DESIGNATORS

In Naming and Necessity1 Saul Kripke describes the possibility of

introducing a term for something, say a name for a person or a star,
by citing a definite description, not to give a definitional equivalent,
but solely to fix the reference. A term so introduced will be a rigid
designator. Kripke calls something a rigid designator if in any possible world it designates the same object. . . .2 Kripke claims that as a
consequence of this there is the startling possibility of knowing a
priori contingent truths about the world. For suppose we propose to
introduce the term t as the name of the denotation, if there is one,
of the definite description the . And we do not intend that t shall
mean the or be a mere abbreviation for it. Rather, t is to designate whatever happens to be the ; the definite description serves
solely to fix the reference. If our procedure serves to introduce t as
a rigid designator, t will designate the same entity in all possible
worlds (in which it designates anything at all). It will designate the
same thing that it does in this, the actual world. But there will be possible worlds in which what is the in the actual world, that which t
First published in Midwest Studies in Philosophy, 1977, 2: 1227.

147

E S S AY S O N R E F E R E N C E

designates, is not the in that world. Thus the statement that if the
exists, t is the is merely contingently true, because there are possible
worlds in which it is false. Yet, if the reference of t has been fixed
solely by being the denotation of the description the it looks like
it can be known a priori that if the exists, t is the .
This apparent result goes counter to the way philosophers have
usually thought about what is knowable a priori, as Kripke would be
the first to acknowledge.3 If we offer as a somewhat vague explication
of knowable a priori (and I will not attempt to give a sharper account)
knowable without recourse to experience then perhaps some philosophers have considered certain statements to be both contingent
and knowable a priori; perhaps, for example, the Cartesian I exist or
some statements about language that can be known just through knowing the language. But the procedure Kripke is talking about is quite
general and the statements that supposedly come out as both contingent and knowable a priori do not enjoy any special status such as the
conclusion of the Cogito does. Here is one of his examples:
An even better case of determining the reference of a name by
description, as opposed to ostension, is the discovery of the
planet Neptune. Neptune was hypothesized as the planet which
caused such and such discrepancies in the orbits of certain other
planets. If Leverrier indeed gave the name Neptune to the
planet before it was ever seen, then he fixed the reference of
Neptune by means of the description mentioned. At the Urne
he was unable to see the planet even through a telescope. At this
state, an a priori material equivalence held between the statements Neptune exists and Some one planet perturbing the
orbit of such and such other planets exists in such and such a
position, and also such statements as if such and such perturbations are caused by a planet, they are caused by Neptune had the
148

T H E C O N T I N G E N T A P R I O R I A N D R I G I D D E S I G N AT O R S

status of a priori truths. Nevertheless, they were not necessary

truths, since Neptune was introduced as a name rigidly designating a certain planet. Leverrier could well have believed that if
Neptune had been knocked off its course one million years earlier, it would cause no such perturbations and even that some
other object might have caused the perturbations in its place.4

Before Kripkes discussion of the matter, most philosophers, I should

think, would have considered the sentences he mentions, as they
might be uttered by an astronomer today, as excellent paradigms of
sentences expressing contingent truths knowable only a posteriori.
As uttered by Leverrier, they would have said, the sentences might
express truths knowable a priori, but that would have been because
Neptune, at the time, was a mere abbreviation for a certain description and the sentences expressed mere tautologies. Kripke would
explain, at least in part, I believe, the failure of philosophers to consider the possibility of a priori contingent truths as, in the first place,
a failure to keep firmly separate the epistemic distinction between
the a priori and the a posteriori and the ontological distinction between the necessary and the contingent, the one defined in terms of
knowability with or without recourse to experience and the other in
terms of truth in possible worlds, and, in the second place, as a failure
to make a distinction between establishing a terms meaning and
fixing its reference. Whatever the extent of the influence of these factors there is another source of the uneasiness felt about the possibility of contingent a priori truths (I do not mean to suggest that Kripke
would disagree). It might be put roughly as follows: If a truth is a
contingent one then it is made true, so to speak, by some actual state
of affairs in the world that, at least in the sorts of examples we are
interested in, exists independently of our language and our linguistic
conventions. How can we become aware of such a truth, come to
149

E S S AY S O N R E F E R E N C E

know the existence of such a state of affairs, merely by performing an

act of linguistic stipulation? Or, to put it another way, contingent
truths are true in only a proper subclass of all possible worlds; how
by a mere stipulation of how the reference of a term is to be fixed, can
we come to know that our world is a member of that subclass?
I share these philosophical qualms and yet at the same time
I believe that Kripke has described a viable procedure for introducing a term as a rigid designator (viable in theory; as will become
clear I am dubious that it can be shown to be a part of our usual
linguistic practices). My way out of this dilemma in this paper will
be to suggest that the procedure does not have the consequence
claimed for it; it does not in fact yield the possibility of knowing
a priori contingent truths.

I
I should first say something, I believe, about the question of whether
or not a name can be introduced as a rigid designator, as opposed to
a mere abbreviation, by using a definite description. Michael Dummett, in his long discussion of Kripkes work in his recent book Frege:
Philosophy of Language,5 seems to want to challenge this. In discussing this it is important to be quite sure what we are asking. In particular, we must separate the factual issue of whether it is part of our
actual practice ever to introduce names in this way from the theoretical issue of whether names could be introduced in this way. Kripke
does sometimes talk about examples, such as the introduction of the
name Neptune by Leverrier, as if he wanted to assert that the example is one in which a name has been introduced as a rigid designator. And he may give the impression that he even has an argument
that shows that this is so. (I believe Dummett so reads him.) It is one
150

T H E C O N T I N G E N T A P R I O R I A N D R I G I D D E S I G N AT O R S

thing to fault him on these counts, if indeed those were his intentions, and quite another to show that there is something theoretically
wrong with the very notion of introducing a name in that way. It
seems to me that the philosophical worry about the possibility of
knowledge a priori of contingent truths should be just as strong if it is
no more than a theoretical possibility that a name should be introduced in such a way as to allow for such knowledge.
It is, of course, of importance whether or not names are ever introduced as rigid designators by the use of definite descriptions to
fix their reference. The philosophically interesting question is, what
would show that that was what had been done as opposed to introducing the name as an abbreviation for the description? For we
should not, of course, suppose that names cannot be introduced as
abbreviations; it is obvious that we can do that if we want to. Leverrier probably did not say anything that would disclose an intention
that the name should function one way rather than the other. Kripke
tells us that this is an example of the introduction of a name as a
rigid designator, but why is he so confident that it is not an example
of a name introduced as an abbreviation? At the end of the passage
quoted he may seem to be giving an argument. He says, of the things
he believes Leverrier could know a priori:
Nevertheless, they were not necessary truths, since Neptune
was introduced as a name rigidly designating a certain planet.
Leverrier could well have believed that if Neptune had been
knocked off its course one million years earlier, it would cause no
such perturbations and even that some other object might have
caused the perturbations in its place.6

It might be thought that there is the following sort of argument

being given: Take the modal sentence,
151

E S S AY S O N R E F E R E N C E

(A) Neptune might have existed and not been the cause of the perturbations in the orbit of Uranus.7

Following Kripke, it seems that having just introduced the name

via the description contained in (A), Leverrier might nevertheless
believe without any inconsistency what (A) expresses. But that
seems to show that the following sentence expresses a contingent
truth:
(B)

If Neptune exists, Neptune is the cause of the perturbations in

the orbit of Uranus.

But if Neptune were a mere abbreviation for the description

in question then (B) would be equivalent, by substitution of
the description for the name, to a mere tautology. Thus, or so it
might seem, we can show that Neptune was not introduced as an
abbreviation.
Anyone bent on maintaining that in Kripkes example the name
was introduced as an abbreviation can escape this argument. He
will, as Dummett does, point to scope differences. Let us suppose
that Neptune was introduced as an abbreviation for the description. Then we can explain why we are inclined to agree with Kripke
that Leverrier might without inconsistency believe what (A) expresses by claiming that (A) is ambiguous as between two readings,
one with the modal operator having wide scope and the other with
its having narrow scope. That is, if we substitute for Neptune in
(A) its claimed definitional equivalent we obtain:
(C)

The cause of the perturbations in the orbit of Uranus might

have existed and not been the cause of the perturbations in
the orbit of Uranus.
152

T H E C O N T I N G E N T A P R I O R I A N D R I G I D D E S I G N AT O R S

But this, it would be claimed, is ambiguous. We can represent the

two readings explicitly by:
(D) It might have been the case that [the cause of the perturbations in the orbit of Uranus did not cause the perturbations in
the orbit of Uranus].

and
(E) The cause of the perturbations in the orbit of Uranus might
have been such that it did not cause the perturbations in the
orbit of Uranus.

(D), of course, is plainly false and we cannot suppose that Leverrier

would have believed it. But (E) expresses something he should have
believed, for surely, the planet actually responsible for the perturbations might have met with an accident millions of years ago, etc.
That Leverrier might well have consistently believed what is
expressed by (A) does not then show that Neptune was not introduced by him as an abbreviation for the description nor that
(B) is contingently true. For on the hypothesis that it was introduced as an abbreviation (E) expresses one of the propositions
(A) could express and (E) is something Leverrier might well have
believed.
The attempt to construct an argument from modal beliefs to
show that a rigid designator has been introduced seems, then, always
open to this sort of evasion. As I will emphasize once more a bit
later, however, this does not show in the least that names cannot be
introduced as rigid designators by means of definite descriptions,
much less that there is, for example, anything wrong with the thesis
that names are, in general, rigid designators.
153

E S S AY S O N R E F E R E N C E

If we were concerned to know about an actual case whether a

name had been introduced, by means of a description, as a rigid designator rather than as an abbreviation, I am inclined to believe we
would be in some difficulty. Not only is there no conclusive argument of the sort described, I doubt that we could rely in general on
linguistic intuitions. In the absence of an explicit stipulation that the
name shall be taken as a rigid designator and the description as
merely fixing the reference (and historically we have not, I believe,
possessed a formula for making such a stipulation8) we would have
to rely on our intuitions about the syntactic, semantic, modal or
epistemic properties of sentences in which the introduced name
occurs. The two possibilities, that the name is a rigid designator and
that it is abbreviation, yield different predictions concerning the
properties of various sentences and classes of sentences. For example, if the name is a rigid designator then, where N is the introduced name and the the introducing description, the sentence
of the following form would express a contingent truth:
(F) If N exists, then N is the .

If, on the other hand, the name is an abbreviation the sentence of

that form would express a tautology. But this difference in the predicted status of a certain sentence does not seem to me to be a usable
difference. In the absence of an explicit stipulation, any profession
of an intuition about the status that we might make or even that the
stipulator himself might make would seem to have the same status
as Kripkes opinion that the sentence, in certain examples, expresses
a contingent truth or Dummetts, about the same examples, that it
does not. For my own part, when I think about such an example as
the Neptune case, I dont find myself with any strong intuition one
way or the other.
154

T H E C O N T I N G E N T A P R I O R I A N D R I G I D D E S I G N AT O R S

The same unusability of differences in predicted properties

seems to me to hold for other sentence types. If the name is an abbreviation, it follows that certain pairs of sentences will be paraphrases of each other, and that would not be the case if the name is
a rigid designator. Thus, in the Neptune example, Neptune is a
large planet would be paraphrasable, if the name is an abbreviation for the description, as The cause of the perturbations in the
orbit of Uranus is a large planet. When I think of Leverrier introducing the name by saying some such thing as Let us call the cause
of the perturbations in the orbit of Uranus Neptune, I do not find
myself with any strong intuitions that such pairs of sentences are or
are not paraphrases of each other. Similarly, as we saw, modal sentences such as (A) in the argument discussed above, will have two
non-equivalent readings if the name has been introduced as an abbreviation, but not if it is a rigid designator. Where N is a rigid
designator to say It might have been the case that [N is not ] is
equivalent to saying N might have been such that it is not (this
point is sometimes put by saying that the name always has wide
scopethat it can always be brought outside the scope of the
modal operator without change of meaning). But again I doubt
that anyone would have any strong intuitions about this concerning a situation such as that of Leverrier and the introduction of the
name Neptuneat least none not motivated by a theory about
the matter.
I believe the conclusion to be drawn from this is that in the absence of stipulation that the name shall be one or the other, it would
be indeterminate whether a name introduced by means of a description is a rigid designator or an abbreviation, so long as the name
continues to be pegged to the description. It may be interesting that
a name could have this indeterminacy, but I do not see that it should
cause concern. It certainly would not make any practical difference.
155

E S S AY S O N R E F E R E N C E

It will be indeterminate, for example, whether certain pairs of

sentences are paraphrases of each other, but there will not be any
question but that what they express will have the same truth value.
And should someone wonder how we know this, the right response
would be to remind him of the agreed upon stipulation, even though
it is indeterminate whether or not that stipulation introduced a rigid
designator or an abbreviation. This conclusion, of course, presupposes that the concept of introducing a name as a rigid designator
by means of a description is a coherent one.
It seems to me that the concept is a coherent one. If it is specifically intended and stipulated that this introduction of a name is to
be the introduction of a rigid designator, I see no theoretical reason
to suppose that the stipulation cannot accomplish what is intended.
Certainly none of the above facts about the status of names in cases
where there is no explicit intention one way or the other casts any
shadow on the theoretical possibility of such an introduction. (And
I must confess that while I am not confident I fully grasp Dummetts
attack on Kripke, a large part of it seems to me to accomplish no
more than what I have set out above.) I should like also to point out
even more emphatically that these facts certainly cast no doubt
upon the thesis that in general names are rigid designators. For we
are talking now about a special sort of situation, one in which a
name is introduced by and subsequently, for some period of time, is
pegged to a definite description. And there is no obvious reason
why what we would conclude about this sort of situation should be
extended to or have implications for what we should say about, say,
the present use of the names Aristotle or Ford. Those who hold
that such names are rigid designators need not hold that they are
now or were at any time pegged to a definite description that fixes
the referent. A name such as Neptune, which we will suppose was
first introduced by some description, is certainly no longer tied to
156

T H E C O N T I N G E N T A P R I O R I A N D R I G I D D E S I G N AT O R S

the description that was used to introduce it. This is shown by the
fact that if someone were to ask an astronomer how he knows that
Neptune causes the perturbations in the orbit of Uranus, he would
be treated, undoubtedly, to a discussion of such things as astronomical observations, details of orbits, gravitational forces and the like,
whereas, just after Leverriers stipulation the same question would
have received (whatever the names status as rigid designator or abbreviation) some such reply as That is just what we call Neptune
or It is just by stipulation that we call the cause of those perturbations, Neptune. Now those who hold some version of what I have
called the principle of identifying descriptions,9 a view that Kripke
and I have attacked, will describe the breaking away of the name
from the introducing description either as a case in which it has
come to be an abbreviation for (or associated with, to accommodate
certain versions) another description or as the introducing description having come to be merely one member of a set of descriptions
associated with the name (where the referent is determined by
being that which satisfies some proportion of the descriptions). But
there is no obvious reason why someone who held that the name
had originally been introduced as a rigid designator by means of one
description should have to say that now, when it has been disengaged from that description, there is some other description that
fixes its referent.10 And in fact those who hold that a name such as
Neptune is currently a rigid designator do not take, so far as I
know, any such route. Instead, they theorize that there is another
relationship that can hold between the user(s) of a name, the name
and the thing named in virtue of which the thing is named and that
does not involve the thing satisfying any descriptions associated
with the name. Thus Kripke speaks of a causal relationship, I have
talked about a relationship of being involved in an historical
explanation of the use of a name11 and David Kaplan introduced
157

E S S AY S O N R E F E R E N C E

a technical notion of a name being a name of something for someone, characterized in terms of how the user acquired the name.12 A
name that functions according to some version of such a theory will
be a rigid designator, but not in virtue of there being some definite
description associated with the name that fixes the reference. If any
such theory is tenable, then what we say about the situation, for example, when Neptune was tied to a description, will not have any
immediate application, at any rate, to the situation after it has
become divorced from the introducing description. Thus, where
there has been no explicit stipulation that a name being introduced
via a description shall be construed as a rigid designator, the fact
that there is no argument that would show that it is a rigid designator nor, as I have suggested, the possibility of firm intuitions that
might point that way, neither demonstrates (at least straight off )
anything wrong with the thesis that names in general are rigid designators nor even that there is anything amiss with the notion of introducing rigid designators by description. As for the latter, we will
from now on suppose that the stipulation that the name shall be a
rigid designator and the introducing description a device for fixing
the reference is explicitly made. Whether or not we in fact would
have any practical reason for doing this, rather than leaving the
matter indeterminate, the theoretical possibility that a consequence
would be knowledge a priori of contingent truths must be faced.

II
About Kripkes view that by the sort of stipulative introduction of a
name as a rigid designator one would thereby be in a position to
know a contingent truth a priori, Dummett remarks, Counterintuitive it undoubtedly is, but it appears to follow from Kripkes
158

T H E C O N T I N G E N T A P R I O R I A N D R I G I D D E S I G N AT O R S

arguments: something must, therefore, be amiss with those arguments.13 I am not sure what he meant by those arguments, but I
do want to question whether it follows from the possibility of introducing a rigid designator by means of a description that certain contingent truths are knowable a priori. And I will suggest that it does
not. (Or, more cautiously, if there are thereby any contingent truths
knowable a priori they will not be of the sort Kripke mentions and
will not cause the philosophical qualms mentioned at the beginning.) In doing this I am going to invoke a distinction between
knowing that a certain sentence expresses a truth and knowing the
truth of what is expressed by the sentence. I am going to suggest that
as the result of the introduction of a name as a rigid designator by
means of a description fixing the referent we can come to know, perhaps even a priori, that certain sentences express truths, but we do
not come to know, a priori, the truth of what they express.14 This
may turn out to sound as paradoxical as Kripkes position on the
matter, but at least it will not be for the same reasons.
We can illustrate the distinction I have in mind, in the first place,
by a familiar kind of example. A person could know that a certain
sentence expresses a truth if he has been told by an unimpeachable
source that it does. Thus I could learn from a German-speaking
friend that a certain German sentence expresses something true, but
if I do not speak German and do not know what the sentence means
I would say that I do not know the truth of what the sentence
expressesor rather, if by chance I do, it would not be just from the
information given me by my friend. Similarly, I could be assured by
a qualified mathematician that a certain mathematical sentence
expresses a theorem, without thereby knowing the truth of the theorem expressed, if I am ignorant of what the mathematical sentence
means. Now it is true that in such cases I can pass on to someone
else not only that the sentence in question expresses a truth, but also
159

E S S AY S O N R E F E R E N C E

the truth it expresses. For example, having read in an authoritative

article n Scientific American, The oblateness of Mars is .003, I may
be said to know that that sentence expresses a truth, but if I have not
the foggiest idea of what is meant by oblateness, I do not think I can
be said to know that the oblateness of Mars is .003. Yet if I subsequently happen to be in a group discussing the properties of the
planets and someone asks what the oblateness of Mars is, I might
answer It is .003. But I think I act here, to use an apt expression of
Gareth Evans,15 as a mere mouthpiece, passing on a statement about
matters of which I am ignorant.
In these sorts of examples the person is ignorant of the meaning
of at least part of the sentence involved. Where demonstratives or
pronouns are involved, however, one might know, at least in one ordinary sense, what a sentence means, know that it expresses something
true, but not know the truth of what it expresses. If you say to me,
That is mine, I know what the sentence means and if, for example,
I also know that you keep scrupulous track of your possessions, but
never claim something that is not yours, I may know that the sentence, on this occasion of its use, expresses something true. But if I fail
to grasp what it is that you are referring to by that I will not yet
know the truth of what the sentence, on this occasion, expresses.
None of this, of course, is new; I mention these examples only to
make sure the distinction I want is clear. A slightly more interesting example and one somewhat closer to the situation which, I will suggest,
holds in the case of introducing rigid designators via descriptions, is suggested by a remark of William Kneales (quoted by Kripke for another
purpose). Kneale says, While it may be informative to tell a man that the
most famous Greek philosopher was called Socrates, it is obviously trifling to tell him that Socrates was called Socrates.16 As Kripke points out
Kneale cannot be correct about the example as he gives it, because it is
surely not trifling to tell someone that Socrates was called Socratesit
160

T H E C O N T I N G E N T A P R I O R I A N D R I G I D D E S I G N AT O R S

is not even obviously true that Socrates had the same name we use for
him. But let us patch this up and suppose that someone says, Socrates is
called by me Socrates. This looks more like it may be trifling and let us
grant that it is. Then in general if anyone says something of the form N
is called by me N, he will have asserted something trifling, if he asserts
anything at all. Now suppose that you say to me Vladimir is called by me
Vladimir and I do not have the least idea who Vladimir is. From the
general principle I can see that if you have asserted anything, your sentence expresses a truth. But I do not believe that I know that Vladimir is
called Vladimir by you. I do not know the truth of what your sentence
expresses. This example shows that considerations other than the reliable testimony of those who understand what a sentence expresses can
put one in the position of knowing that a sentence expresses a truth, but
not knowing the truth of what it expresses.

III
If we imagine all the participants to have mastered the notion of
a rigid designator then we could make explicit the intention to
introduce a name as a rigid designator by using some such formula as, Let N be a rigid designator with its referent fixed by
the . But, because I think it somewhat illuminating to do it
this way, I am going to propose instead that we think of the introduction as consisting of stipulating that a certain sentence shall
express a contingent truth. If we want to introduce the name N
by means of the description of the then the formula we would
use would be:
(a)

Provided that the exists, let N is the express a contingent

truth.
161

E S S AY S O N R E F E R E N C E

It is a condition on the stipulation that the exists and should it turn

out that it does not, the stipulation, we might say, has been an unhappy one and not to be taken as being in effect.17 It should be clear
from the preceding remarks that I am not in any way suggesting that
this represents some sort of analysis of any practice that we have
ever engaged in.
No doubt the procedure looks strange. We are accustomed to
linguistic conventions that give meaning to words, phrases and
sometimes whole sentences, but not to conventions that specify
that a certain sentence shall express a truth. Perhaps the closest
approximation in ordinary practice is the laying down of axioms
in an axiom system. But, of course, what is most strange about the
procedure is that a sentence is by stipulation to express a contingent truth.
I can imagine a philosophical worry about the procedure not
unlike the worry about the possibility of the contingent a priori.
How, it might be asked, could one stipulate that something be contingently true? What makes something contingently true is how the
world is, the existence or non-existence of some state of affairs.
Surely only God, if even He, could perform the miracle of stipulating how the world shall be.
How shall we reply to this? In the first place, it is of course crucial
that the name in the sentence about which the stipulation is made is
a fresh new name or, if it has been used previously to name something, that it is being given a new use. I could not, for example, get
away with stipulating that, say, Gerald Ford will be starting quarterback for the Rams in 1976, where I intend all the words to have
their usual meanings and the names their usual referents, shall be
contingently true. Secondly, the philosophical worry assumes that
we are trying by mere stipulation to establish that a certain state of
affairs shall exist. That would be an exceedingly odd thing to suggest
162

T H E C O N T I N G E N T A P R I O R I A N D R I G I D D E S I G N AT O R S

we try to do. But that is not what the stipulation is intended to

accomplish nor does it do so. It stipulates that a certain sentence
containing a name for which there is no independently intended referent shall express a contingent truth. Now what state of affairs will
turn out, if any does, to correspond to the sentence is, so to speak,
up to the world. Imagine the detective, of philosophical bent, instead of using the hackneyed, Let us call the murderer X, saying
Provided the murderer exists, let Vladimir is the murderer express
a contingent truth. We can suppose that Jones is the actual murderer. So in the actual world the state of affairs that corresponds to
the sentence in question is Jones being the murderer and what the
sentence expresses is the existence of that state of affairs. But someone else might have been the murderer; in other possible worlds it is
Smith or Robinson. We can imagine everything being the same
except that someone else is the murderer (and whatever differences
are entailed by that). Then a different state of affairs would correspond to the sentence and it would express the existence of a different state of affairs. We could also put this by saying that in different
possible worlds Vladimir, as the name is introduced by the stipulation in those worlds, would turn out to rigidly designate someone
else. So the detective is not, by stipulation, attempting to create a
state of affairs (other than a linguistic one).
Not only does the detective not create, by stipulation, any state
of affairs (other than a linguistic one), but I should like to say nor
does he thereby come to know the existence of any state of affairs.
Kripke himself says, in connection with his example of the standard
meter bar, But, merely by fixing a system of measurement, has he
thereby learned some (contingent) information about the world,
some new fact that he did not know before? It seems plausible that in
some sense he did not, though it is undeniably a contingent fact that
S is one meter long.18 I should like to question, in these cases,
163

E S S AY S O N R E F E R E N C E

whether there is any sense in which by the sort of stipulation we are

talking about a person could come to know something of which he
was previously ignorant.
Let us look at an example from David Kaplans paper Quantifying In which he also uses in a subsequent paper Dthat,19 and about
which, interestingly, he draws diametrically opposite conclusions, a
fact I will touch upon later. I will adapt the example slightly to fit the
format we are using. Consider the description, the first child born
in the 21st century. Even though the denotation, if there is one,
does not yet exist and moreover, with certain assumptions about
determinism and freedom of the will, it may not even now be determined who that individual will be, we can by stipulation introduce a
rigid designator for that person, if there is to be one. So, following
Kaplan, we shall stipulate that providing the first child born in the
21st century will exist, the sentence Newman 1 will be the first
child born in the 21st century shall express a contingent truth. Let
us now imagine that just after midnight on New Centurys Eve a
child is born who is firmly established to be the first born of the
century. He is baptized John, but those of us who are still around,
remembering our stipulation, also call this child Newman 1. Now
it seems to me that it would be outrageous to say that some twentyfive years or so before his birth, we knew that John would be the first
child born in the 21st century. Suppose one of us, living to a ripe old
age, were to meet John after he has grown up a bit. Would it be true
to say to John, I call you Newman 1 and Newman 1, I knew some
twenty-five years or so before your birth that you would be the first
child born in the 21st century?
The Neptunians are watching on their interplanetary videoscope; they see and hear Leverrier perform his act of stipulation
(and since this is science fiction let us also imagine that, anachronistically, he intends the name he is introducing to be a rigid designator
164

T H E C O N T I N G E N T A P R I O R I A N D R I G I D D E S I G N AT O R S

and even uses our formula). They know that their planet is the cause
of the perturbations in the orbit of Uranus. Would they be justified
in concluding that the Earthling has learned or come to know that
their planet is the cause? It seems to me that the answer is obviously
that they would not. Suppose they call their planet Enutpen.
Would they be justified in saying, in Enutpenese, that the Earthling
now knows that Enutpen is the cause of those perturbations? Again
I think not. Suppose, finally, that they like the sound of the name
Neptune; one of them suggests they too adopt the Earthlings
convention and agree that Neptune _____ (where _____ is a
translation into Enutpense of is the cause of the perturbations in
the orbit of Uranus) shall express a contingent truth. Would they
be justified in saying that the Earthling knows the truth of what is
expressed by them as Neptune _____? I am still inclined to say
that they would not be.
The above are supposed to be considerations toward showing
that what some of us may know after the advent of the 21st century is
not anything we know now and that what the Neptunians may have
known is not anything that Leverrier knew just as a result of his stipulation. And I am assuming that if we know something now just as a
result of our stipulation it cannot be anything different from something we may know after the birth of the first child in the 21st century
and that if Leverrier did know something just as the result of his stipulation it could not be different from what the Neptunians knew.
I am assuming, to use the jargon, that if we now have any knowledge (other than about linguistic matters) just as a result of the stipulation concerning the sentence, Newman 1 will be the first child
born in the 21st century it would have to be knowledge de re. That
is, it would have to be knowledge about an individual in the sense
that there is (or will be) an individual about whom we now know
something and if that individual turns out to be John we now know
165

E S S AY S O N R E F E R E N C E

something about John. We could, of course, put this point by saying

that the sentence which expresses the proposition that we have such
knowledge, and which contains the name Newman 1 is open to
substitution on the name and allows of existential generalization on
it. This is as opposed to the knowledge we might come to have de
dicto, for example, that the firstborn child, if there will be one, in the
21st century will be bald, knowledge that we might come to have,
for example, because we get indisputable evidence that pollution
from aerosol sprays has caused genetic changes and every child born
after a certain time will be bald. That knowledge would not be
knowledge about a certain individual, not knowledge about John.
I make this assumption that the knowledge, if we have it, would
have to be de re not simply on the grounds that Newman 1 is a
rigid designator. It does not follow from the fact that a term is a rigid
designator that when it enters into a statement of propositional attitude, the attitude ascribed must be de re. It does not follow because
not all rigid designators lack descriptive content. Mathematical definite descriptions, when they do not lack a denotation, presumably
would be rigid designators, designating the same thing in all possible worlds. Although I know that the 98th prime number is not divisible by three, it does not follow that I know about the number
which is the 98th prime number that is not divisible by three.20 It is
rather, then, that as these stipulations introduce names they give the
names no descriptive content that leads me to say that the knowledge, if there is any, must be de re knowledge and then to try to show
that we do not have such de re knowledge.
Now the notion of de re propositional attitudes is a notoriously
difficult one. It is difficult to know what the proper analysis is of
statements ascribing such attitudes and it is difficult to know under
what conditions they are true. It also creates formidable problems
for the sort of theory of reference to which Kripke and I and others
166

T H E C O N T I N G E N T A P R I O R I A N D R I G I D D E S I G N AT O R S

subscribe, problems that have to do with identity conditions for

such propositional attitudes.21 I do not want to push the whole
matter under the rug, but I would like, if possible, to step around the
messy areas for the purposes of the point I want to establish in this
paper.
Putting aside any attempt to account for them and any theories
about how names function, let us look at a couple of loose principles
concerning names and propositional attitudes. I say loose because
I do not want to defend them as having no counter-examples.22 The
first one is this: If an object is called by one name, say N, by one
group of people and by another name by a second group, say M,
and if, in the language of the first group N is expresses a bit of
knowledge of theirs and if is is a translation of is into the
language of the second group then if the relevant facts are known to
the second group, they can say truly that the first group knew that
M is . Thus, for example, our historians would say that the early
Indian residents knew that Santa Catalina is a single island, unlike
the first Spanish explorers in the area, even though, of course, the
Indians have a different name for it. Now this principle fails in the
Newman and Neptune examples. If the first child born in the
21st century comes to be named John it would not be correct to
say then that although we had a different name for him we knew
twenty-five years beforehand that John would be the first child born
in the 21st century. Nor should the Neptunians admit that while
Leverrier, presuming he has introduced a name as a rigid designator,
had a different name for Enutpen, he knew that Enutpen was the
cause of the perturbations in the orbit of Uranus.
Another such principle is this: If one has a name for a person, say
N, and there is a bit of knowledge that one would express by saying
N is then if one subsequently meets the person it will be true to
say to him, using the second-person pronoun, I knew that you were
167

E S S AY S O N R E F E R E N C E

. I believe I knew that Bertrand Russell was a short man. Had I met
him I could quite correctly, though impolitely, say to him, I knew
that you were a short man. As we have seen this too fails in the
Newman 1 case. (A similar loose principle could be constructed
substituting a demonstrative such as this planet that would be
seen as failing in the Neptune example.)
In the absence of any other explanation of why these principles
should fail in these cases I suggest that the reason is that the stipulations have not given rise to any knowledge (other than of linguistic
matters). And so not to any knowledge a priori.

IV
If this is correct that indeed the introduction of a rigid designator by
a definite description does not give rise to the sort of knowledge
Kripke thought would be a priori and of contingent truths, what
then of the argument at the beginning that seemed to show that it
would? Where does it go wrong? The crucial step, I believe, was
where we said, Yet, if the referent of t has been fixed solely by
being the denotation of the definite description the it looks like it
can be known a priori that provided the exists, t is the . I believe
that this move looks very plausible because what can, I believe, be
known is the following:
(A)

Provided the exists, t is the expresses a contingent truth.

I am supposing here that (A) is not itself the stipulation, even though
it is close to it in phrasing, but a statement about the language into
which the name has been introduced. But it does not follow from
the fact that as the result of a stipulation one can know something of
168

T H E C O N T I N G E N T A P R I O R I A N D R I G I D D E S I G N AT O R S

the form of (A) that one thereby knows the truth of what the sentence of the following form expresses:
(B)

Provided the exists, t is the .

But, of course, it was the truth of what some sentences of form (B)
express that Kripke thought we could know a priori as the result of a
stipulation.
Now it might be asked whether we have not still got the contingent a priori on our hands, because does not (A) express a contingent
truth, albeit one about a language, and will it not be knowable a priori?
My answer to this is that I am not sure whether in the circumstances
what sentences of form (A) express are both contingent and a priori.
But if they are they are harmless varieties of the contingent a priori,
examples of which we could produce without recourse to stipulations
introducing rigid designators. Perhaps it would be argued that since
such sentences express a truth about a language that might have been
false (if, for example, the stipulation had not been made) they are
contingent and that because their truth can be assured merely by reference back to a linguistic stipulation, they are a priori. But if so, then,
for example, traditional definitions will yield the contingent a priori.
Suppose I stipulate that Widgit shall mean by definition green
cow, then I know the truth of what the following expresses:
(C) Provided there are any green cows, Widgits are green cows
expresses a truth.

But the same considerations would argue for this being a priori
knowledge of a contingent truth. If this proves that there can be a
priori knowledge of the contingent, then it seems to me that the
contingent a priori is not very scary and not very interesting.
169

E S S AY S O N R E F E R E N C E

V
At the end of his paper Dthat Kaplan says that It is now clear
that I can assert to the first child to be born in the 21st century that
he will be bald. In terms of the modifications in the example made
in this paper, he thinks he could do this, having introduced the
name Newman 1 in the way specified, by assertively uttering
Newman 1 will be bald. Kaplan is here recanting the position he
had in Quantifying In. Part of the task Kaplan set for himself in
that paper was to give the conditions under which a person can
correctly be said to have a de re propositional attitude. The
Newman 1 example was, in fact, used to show the necessity of
requiring for such an attitude that the person having it be in a special sort of relationship to the entity about which he has it. He
says, I am unwilling to adopt any theory of proper names which
permits me to perform a dubbing in absentia, as by solemnly
declaring I hereby dub the first child to be born in the twentysecond century Newman 1. 23 In Dthat he has become armed
with the apparatus for introducing rigid designators and definite
descriptions and this leads him, I believe, to think that he can perform the feat that he earlier thought obviously impossible. If the
position of this paper is correct, however, the fact that a name is
introduced as a rigid designator does not by itself put a person in a
position to have de re propositional attitudes toward the entity rigidly designated. For essentially the same considerations that were
adduced for denying that there was knowledge of an entity just in
virtue of the sort of stipulation that introduces a rigid designator
by means of a description can be applied to the other propositional attitudes. It would, for example, seem to me just as incorrect
to say to John who turns out to be the first child born in the 21st
century, I believed about you some twenty-five years before your
170

T H E C O N T I N G E N T A P R I O R I A N D R I G I D D E S I G N AT O R S

birth , I asserted about you some twenty-five years before

your birth , etc.
If this is so, we are in the somewhat odd position of possessing a
mechanism for introducing a name that rigidly designates something, but a mechanism that is not powerful enough to allow us to
use the name! But that it is odd does not show, of course, that it is
not true.
Still it may be thought that the result is so odd that something
must be wrong. It might be said, I can understand how a person
might know about a sentence in someone elses language or idiolect
that it expresses a truth without knowing the truth of what it
expresses; even where the sentence is in some sense, a sentence in
your language, as in the examples employing demonstratives or
proper names as used by someone else, at least the reference has
been fixed by the other person and that is why you might know that
the sentence, as used by him, expresses a truth without knowing the
truth of what it expresses. But you are trying to tell us that there is a
procedure for introducing a name, with its reference fixed, and
thereby an indefinite number of sentences employing the name,
into your very own language. And that is what makes it difficult to
see how, concerning some of these sentences introduced into your
language, you could be in the position of knowing that they express
truths, but not knowing the truth of what they express. In particular,
if P is a sentence in your language and we assume that the reference
of any names and demonstratives has also been fixed, how could
you fail to know that P expresses a truth if and only if P? But, of
course, if you know that in the cases you are concerned about, you
will know the truth of what is expressed.
I will, of course, have to agree that if I am right one result of introducing names via definite descriptions in the way suggested is
that there are sentences introduced into ones language such that
171

E S S AY S O N R E F E R E N C E

about any particular one P, one does not know that P expresses
a truth if and only if P. Or, rather, I will if this is not simply a terminological disputesimply a refusal on the one side to say that a sentence has been introduced into a persons language unless he knows
the biconditional. In that case, I do not mind what we say, as
Anscombe, I believe, once put it, so long as the facts of the matter
are straight. And the facts, as I have described them, are that the
procedure sets up the apparatus necessary for a set of sentences to
express truths and falsehoods without thereby putting the possessor
of the apparatus in a position to know what they express. And the
real issue, I believe, is whether there is any impossibility in this. If we
admit the possibility, then whether or not we say that the sentences
have really been introduced into his language is unimportant. Consider the following analogous situation: I close my eyes and say
(pointing), I will call the color of that Murple. I do not know
what I am pointing to, if anything. Let us suppose I am pointing to
something of a definite color. Have I not set up an indefinite number
of sentences, for example, Murple is the color of my true loves
hair, each of which expresses something true or false? But while my
eyes remain shut I do not believe I know what they express. The apparatus, however, has been constructed and I have only to open my
eyes to see, for example, how ludicrous it would be to think that
murple is the color of my true loves hair and that grass is murple.
And analogously to the introduction of names by the procedure we
have been talking about, I believe I know, while my eyes are shut,
that providing I am pointing to something of a definite color, where
were I to use assertively the sentence Murple is the color of that
with eyes open, I would express a truth, but with eyes closed I do
not believe I would know the truth of what I would have asserted.
There is, however, an alternative way of characterizing the situation that I do not know how, at the moment, to counter.24 And that
172

T H E C O N T I N G E N T A P R I O R I A N D R I G I D D E S I G N AT O R S

is that what these stipulations succeed in doing is reserving a name

for a future use. Thus Newman 1 is now reserved for use in referring to whoever turns out to be the first child born in the 21st century, much as prospective parents may choose a name for the next
female child they have. When we come to know who the first child
born in the 21st century is, when we come to be in a position to refer
to him and have beliefs about him, then the name, if we remember
and adhere to our former intentions, becomes a part of our repertoire and an indefinite number of sentences using the name enter
our language. A consequence of this account is that these sentences
did not express anything true or false prior to the time at which we
were in a position to use them ourselves. And thus it would be wrong
to characterize our position as that of knowing that certain sentences express truths, but not knowing the true of what they express.
Rather, what we can know is that certain sentences, if and when we
come to be in a position to use them, will express truths.
I do not, at the moment, see how to show that this way of viewing the matter, as opposed to how I have described it, is or is not the
correct way. I do not, however, find this particularly upsetting. The
alternative view still preserves the main points I wish to make,
namely, that there is a way of introducing a name, albeit as a reserved
name, via a definite description, but not as an abbreviation, and that
knowledge a priori of contingent truths (of a kind to raise philosophical qualms) is not a result.

VI
In Quantifying In, Kaplan held that in order to have a de re propositional attitude toward an entity one must be, as he put it, en rapport with it. And he thought that being en rapport involved three
173

E S S AY S O N R E F E R E N C E

things: One must possess a name for the entity that (1) denoted
the entity, (2) is for the user a vivid name and (3) in a technical
sense, is a name for the speaker for the object. For the latter condition to be satisfied the entity must enter into the genetic account
of how the speaker came to acquire the name, the beliefs that he
would express using the name, etc. I do not believe that he succeeds in spelling out exactly how the entity must enter into the account any more than have others who have suggested some similar
condition, but he has a nice analogy with the notion of when a picture is a picture of something (as opposed to merely resembling
something) where it seems clear that a similar genetic requirement
operates. Now for my own part I am inclined to drop Kaplans first
two requirements and try to go with some variant of this third condition as being what is required for a name to be a name that a
speaker can use to assert de re something about an entity. I am, of
course, not going to attempt any defense of that sort of claim here,
but if it were correct it would, of course, account for why the sort of
stipulations we have been discussing do not put us in a position to
assert and, thus, to know anything about the entity for which we
have introduced a rigid designator.
One final remark: Having indicated the direction in which I am
inclined to go, I find myself wanting to ask the question, why, if
indeed it is true, is one in a position to assert and know de re things
about an entity when the entity becomes (in the right way) a part of
the history of ones use of the name? What does that accomplish
that allows for this possibility? But perhaps that is a misconceived
question. Perhaps the only answer is that that is just when we do
ascribe de re propositional attitudes. Perhaps the only task we can
perform is the one Kaplan was attempting, to make sure that we
have spelled out as exactly as possible the conditions under which
such attitudes are correctly ascribed.
174

T H E C O N T I N G E N T A P R I O R I A N D R I G I D D E S I G N AT O R S

NOTES
1. In Semantics of Natural Language, ed. D. Davidson and G. Harman (Dordrecht,
1972), pp. 253355.
2. Ibid., p. 269. Strictly we should add, as Kripke points out, if the term designates
anything in that possible world.
3. Ibid., p. 263.
4. Ibid., pp. 3478, n. 33.
5. (New York, 1973), pp. 110151.
6. Naming and Necessity, p. 348, n. 33.
7. I modify Kripkes description of the case to provide a definite definite description to work with.
8. On the other hand it must be admitted that we have possessed various natural ways
of expressing the intention to introduce a term as an abbreviation for an expression.
9. Proper Names and Identifying Descriptions, Synthese 21 (1970): 335358,
reprinted in Semantics of Natural Language, ed. D. Davidson and G. Harman,
pp. 356379.
10. Suppose, however, that someone were to hold the following sort of view about the
status of a name such as Aristotle: At some point there was an original dubbing
of a person giving him that name. The name is passed along to others not present
at the dubbing. At least in many cases this is accomplished by one person deciding
to use the name to refer to what another who already possesses the use of the
name refers to. And such a person is to be considered as introducing the name
into his own repertoire as a rigid designator using some such description as the
individual to whom Jones was referring when he used the name Aristotle to fix
the reference. I believe it would be a consequence of the position of this paper that
the person would not thereby be put in a position to use the name. And that would
seem to show that sort of view of how names are passed along wrong.
11. Speaking of Nothing, Philosophical Review, 83 (1974): 331.
12. Quantifying In, Words and Objections, Essays on the Work of W. V. Quine, ed.
D. Davidson and J. Hintikka (Dordrecht, 1969), pp. 206242. In this paper,
however, Kaplan adds two other relations so that his view is not a pure example
of the kind of view I have in mind.
13. Frege, p. 121.
14. The conclusion that this is the correct characterization of the matter is not original to this paper. It is suggested by at least two authors that I know of, Alvin
Plantinga [The Nature of Necessity, (Oxford, 1974), pp. 89, n. 1] and Michael
Levin [Kripkes Argument Against the Identity Thesis, The Journal of Philosophy 72 (1975), 152, n. 21. My excuse for a longer discussion is the need for a
fuller exposition and an investigation of the consequences.
15. The Causal Theory of Names, Proceedings of the Arstotelian Society, Suppl.
Vol. 47 (1973): 192.

175

E S S AY S O N R E F E R E N C E

16. Modality, De Dicto and De Re, in Logic, Methodology and the Philosophy of Science: Proceedings of the 1960 International Congress (Stanford, 1962), pp. 62930.
Kripkes discussion occurs on pages 2834 of Naming and Necessity. Kripkes
quark example (p. 284) could be used to make the same point using a common
noun instead of a name.
17. I utilize a stipulation about the contingent truth of a certain sentence for heuristic purposesI do not believe it plays an essential role in my argument. There
may be some difficulties involved with this sort of stipulation. One possible objection to the device in the simple form I give it in the text (suggested to me in
correspondence by Plantinga) is that anyone who thought names to be disguised descriptions might hold that the device would succeed only in introducing N as an abbreviation for a certain description. The only plausible
candidates, however, for what that definite description would be seem to be
either the or the thing denoted by the . The thing denoted by the is
the and the is the are, it could be argued, only contingently true when
true because they express something false of worlds in which the does not
exist. We could avoid this objection, I believe, in one of two ways. We could
define a notion of strictly contingent according to which a sentence expresses
a strictly contingent truth only if the sentence would express a falsehood in
some worlds in which there is a denotation for any definite descriptions contained in the sentence. Or we could change the form or the stipulation to read:
Provided the exists, let If the exists, N is the express a contingent truth.
(I suppose there would also have to be some other conditions understoodfor
example, that the words in the quoted sentence, other than the introduced
name, have their usual meanings and references.)
18. Naming and Necessity, pp. 3467, n. 26. emphasis his.
19. This volume, pp. 383400.
20. If some of Kripkes other views are correct, I believe there will be definite
descriptions that denote things other than abstract formal objects that will also
turn out to be rigid designators. He holds, for example, that certain identity statements concerning natural kinds are, if true, necessarily true, though discoverable
as true only a posteriori. One such statement he mentions is the statement that
gold is the element with atomic number 79. This would entail that the definite
description, the element with atomic number 79 is a rigid designator, designating the same thing, i.e. same substance, in all possible worlds, namely, providing
the statement in question is true, gold. Now I know that the element with atomic
number 79 has more than seventy-eight protons in its atomic nucleus, but since
I have not checked to make sure Kripke made no mistake in his example, I would
not claim to know this about gold.
21. The problems have been posed forcefully by Diana Ackerman in comments in a
symposium at the 1974 Eastern Division A.P.A. meetings and in an unpublished
paper, Proper Names and Propositional Attitudes.

176

T H E C O N T I N G E N T A P R I O R I A N D R I G I D D E S I G N AT O R S

22. I have in mind the Hesperus and Phosphorus kind of cases. According to the
usual story the ancient Babylonians used these as names for what they took to be
two distinct heavenly bodies but which was in fact one, Venus. They believed
in fact, I think we can say, knewthat Hesperus was the first heavenly body to
appear in the evening, but they would have denied that they believed this of
Phosphorus. Did they nevertheless know this about Phosphorus, which, after all,
is Hesperus? And should we say about them that they knew about Venus, though
they did not call it by that name, that it was the first to appear in the evening? The
structure of this sort of case that gives rise to the problems does not seem to me
to be present in the Neptune and Newman 1 examples and so there seems no
reason why the loose principles that I give should not apply to them if they do
in fact present examples of knowledge. Also, I do have some temptation, at any
rate, to say of the Babylonians that they knew about Venus, etc., but I have no
corresponding temptation in the Neptune and Newman 1 examples.
23. Quantifying In, pp. 22829.
24. This was suggested in discussion by Rogers Albritton.

177

This page intentionally left blank

[7]
KRIPKE AND PUTNAM
ON NATURAL KIND TERMS *

The theory of natural kinds terms developed by Saul Kripke and Hilary
Putnam is seen by both authors, I believe, as being intimately connected to Kripkes views about reference, perhaps even a consequence
of them.1 The views about reference, however, were first applied to singular terms and one wonders whether the transition to a theory about
general terms, terms for kinds of things, can be made without more
complications than either author, in my opinion, clearly indicates.
I want to examine, then, the application of some of the central
ideas Kripke developed concerning singular terms to general terms
for kinds. In particular, I will be concerned with his notion of rigid
designation and the generation of, what I call, exotic necessary
truths. I will try to show that the theory of natural kind terms must
in fact use additional ideas and different argumentation. In doing
this I will not, generally, be uncovering things only implicit in
the writings of Kripke and Putnam; for the most part they are
there to be seen. But I think it is easy to be misled about what is
going on and, I think, the authors themselves sometimes go wrong
First published in Knowledge and Mind: Philosophical Essays, C. Ginet and
S. Shoemaker, eds., New York: Oxford University Press, 1983, pp. 84104.

179

E S S AY S O N R E F E R E N C E

in describing their own procedures. My own feeling is that when

one sees more clearly what is involved some interesting puzzles
emerge. At the end I try to develop an outline of one of these.

I
Before getting down to details it would be useful to look at some general features of the Kripke-Putnam treatment of natural kind terms. In
the first place most of the examples they use are words and expressions
in everyday use, such as water, tiger, gold, and heat. While the
theory calls for a certain relationship between the semantics of these
terms and science, the terms obviously are not borrowed from the vocabulary of science and were part of English long before the advent of
modern science. I think it is no accident that terms with these characteristics were chosen. In the first place, although one might suppose
that if terms for natural kinds are to be found anywhere the language of
science would be replete with them, it is not obvious that the KripkePutnam theory is applicable to kind terms in science. Nor is it obvious that it will apply to terms which the vernacular has borrowed
from the language of science, such as plutonium or electron.
Putting aside, however, the doubts just expressed, there is a
second reason for choosing as examples terms from the vernacular
that antedate the rise of science: The Kripke-Putnam theory offers
an answer to an important puzzle about the relationship of vernacular kind terms and scientific discovery. We seem willing to tailor the
application of many of our vernacular terms for kinds to the results
of science and if necessary to allow our usual means of determining
the extension of these terms to be overridden. There is, for example,
a product on the market composed half of sodium chloride and half
of potassium chloride. It looks like and tastes like ordinary salt. In
180

K R I P K E A N D P U T N A M O N N AT U R A L K I N D T E R M S

most ordinary circumstancesin talking about how much to put in

the stew, for examplewe would be happy to call this product salt
even if we knew its chemical composition. But if pressed to say
whether this product is really salt, I think we would, if we know
some elementary chemistry and the chemical composition of the
product, concede that it is only half salt. To take a couple of more
examples, I would give up calling a stone purchased as a diamond a
diamond if assured by experts that it did not possess a certain crystalline structure of carbon and I am prepared to be corrected when
what I take to be a wolf in a cage at the zoo turns out to be identified
by zoologists as being of a quite distinct species.
This apparent reliance on results and classification raises a question about the extension and meaning of kind terms in the vernacular prior to investigation. It is at least plausible that prior to modern
chemistry, crystallography, and zoology a mixture of sodium and
potassium chlorides, a sparkling clear crystal of some other substance than carbon, and an animal looking exactly like a wolf, but of
that other species, might be called salt, a diamond, and a wolf
without a way for anyone to be the wiser.
Several questions come up at this point. Is the extension of
the terms in question as used prior to the relevant results the
same as or different from the extension of the terms as used today?
Is the meaning of the terms the same or different? And what is the
status of the results? To take the term salt as an instance, it seems
very plausible to say that at a certain point it was experimentally
discovered that salt is the compound sodium chloride. Subsequently, knowledgeable speakers appear to use the property of
being sodium chloride as the ultimate criterion for whether or
not a bit of stuff is really salt. One common response to these
observations is that prior to the scientific discovery being sodium
chloride was not a part of what it meant to be salt, but that after
181

E S S AY S O N R E F E R E N C E

the scientific discovery (for whatever reason), it became a part of

(perhaps the whole of ) the meaning of the term salt. This, however, has some unintuitive consequences. First, it implies that in
this instance as in many others a word has changed its meaning
due to a scientific discovery. But we do not generally suppose that
when we see a word such as salt in, say, John Lockes Essay Concerning the Human Understanding, it has a different meaning.
There is also the not unlikely possibility that if such a view of the
situation is correct then the extension of such a term has also
changed. And, finally, it is difficult to see how we can now speak of
an experimental result if we have made it part of the very meaning
of salt that salt is sodium chloride.
The Kripke-Putnam view seems to have the virtue that none of
these unintuitive results need be accepted. The facts as presented do
not show, on their view, that, e.g., the word salt has changed its
meaning from, say, Lockes day to ours or that it has changed its extension. People in Lockes day might have called some stuff salt that
we would not (on the grounds that the stuff is not sodium chloride)
but that does not show that the extension is different, only that
(providing that our chemists have got things correct) people in
Lockes day made mistakes which they had no means to correct.
And finally that salt is sodium chloride was and remains a statement
of an experimental discovery.

II
The way in which the results mentioned in the last section may seem
to come about is through the application to kind terms of a now
well-known, but still startling, result that Kripke obtained for proper
names. Proper names, according to Kripke (and I agree with this)
182

K R I P K E A N D P U T N A M O N N AT U R A L K I N D T E R M S

are what he calls rigid designators. It will suffice for our purposes
to characterize a rigid designator as a term which designates the
same individual in all possible worlds. It will be recalled that in both
Naming and Necessity and in Identity and Necessity Kripke
introduces the notion first in connection with what are loosely
called singular terms. In particular, proper names are, in general,
rigid while definite descriptions, for the most part, are not.
It will be recalled that one of the startling results Kripke obtains
is that in regard to terms rigidly designating contingently existing
objects (as opposed, say, to mathematical objects), certain sentences
involving such terms turn out to express necessary truths, although
the fact that they express truths is to be learned by empirical means.
I have called such truths exotic necessary truths.2
I suppose one reason that this result is startling, aside from the
fact that it represents a possibility not countenanced in philosophy
before, is that since necessary truths are true in all possible worlds
they do not distinguish the actual world from other possible worlds
and so it looks as if an examination of how the actual world is could
have no bearing on establishing the truth.
In any event, providing we accept the notion, rigid designators
for individuate produce exotic necessary truths in a relatively simple
manner. I stress this because I think matters to be different when we
turn to terms for kinds.
Restricting ourselves to contingently existing individuate, exotic
necessary truths arise when there are two rigid designators for the
same individual. Kripkes main examples are sentences expressing
the relation of identity in which, to put it loosely, the identity sign is
flanked by two different rigid designators, designating the same
individual. So if D and D are two such terms the sentence
D is identical with D
183

E S S AY S O N R E F E R E N C E

expresses an exotic necessary truth. And the argument is simply

this: since by hypothesis D and also D designate the same individual in all possible worlds they must designate that individual in
each case. But since, by hypothesis, they designate the same individual in this world they must do so in all possible worlds. Hence,
the sentence expresses a necessary truth. The sentence expresses an
exotic necessary truth for the following reason: While it may be
possible for someone not to have to do any research about the
actual world in order to know that such a sentence expresses a
truthas when, for example, one is introduced to a person and in
the same breath given two names for himit is obviously possible,
and does occur, that a person or a culture has two names for the
same individual without realizing it. So we have the stock examples
of someone having both the names Cicero and Tully without realizing that, as they are used by him, they name one and the same
person, and (what historically seems to me questionable) the story
of the Babylonians having given the names Hesperus and Phosphorus to the planet Venus without realizing that a single heavenly
body had been named. Under these circumstances, empirical investigation is required to ascertain that the identity sentence in
question expresses a truth.
If this short exposition of Kripkes view about proper names is
correct, it generates exotic necessary truths using essentially just
the notions of rigid designation and identity.
What I want to argue is that this is not enough to give us the
theory of natural kind terms that Kripke and Putnam want. Rigid
designation (and identity), which gives the spectacular results for
proper names, is not enough to do the same thing in the case of general terms for kinds. I do not mean to say that Kripke and Putnam
would disagree with this, but Putnam says, for example, in his treatment of the term water, Our discussion leans heavily on the work
184

K R I P K E A N D P U T N A M O N N AT U R A L K I N D T E R M S

of Saul Kripke although conclusions were obtained independently

and then . . . we may express Kripkes theory and mine by saying, for
example, that the term water is rigid.3
My view is that here Putnam is wrong about the mechanism of
his own view. However, I also find in some of Kripkes descriptions
of his argument the same mistake. Kripke does use an argument in
the earlier version of his view, Identity and Necessity, which does
seem to extend in a simple manner the results and argumentation
from proper names to (natural) kind terms. The problem is that it
will not work.
But Putnam and Kripke (in the later work, Naming and Necessity) use, on my reading, a more complicated argument. I want to
question the latter argument as well, but to do so requires considerations which are more difficult to assess.
I want to argue that one cannot transfer, in a straightforward
way, the Kripke results about proper names to terms for kinds, even
though in the next section I assume that Kripke and Putnam must
treat kind terms as being singular terms.

III
In trying to extend the notion of a rigid designator to non-singular
terms we immediately come up against a problem which, I believe,
neither Kripke nor Putnam touch upon. Rigid designation was initially defined by Kripke for singular termsthose terms which in
some way carry along with them the requirement that there is a
single individual referred to. A term such as tiger, unless it occurs in
some such context as The tiger who bit my friend, does not at first
glance seem to refer to any individual. Philosophers talk about the
extension of a term and usually apply this notion both to singular
185

E S S AY S O N R E F E R E N C E

terms, such as names, and general terms such as tiger. The extension of a term is, roughly, what it applies to correctly.
Now if by talking about what is designated by a term in all possible worlds we were to think of the extension of the term tiger we
would not find it to be a rigid designator. In different possible worlds
there are more or fewer tigers than in this and I suppose that the particular tigers of the actual world do not exist in all possible worlds.
For the purposes of this paper I am going to assume that construing terms for kinds, such as water, tiger, etc., as rigid designators and giving the Kripke-Putnam view the best run for its money
is to think of them as what Mill calls abstract nouns. Tiger is not
to be thought of as designating its extension. Rather, it designates (is
the name of) a certain species. Water designates, not its
extensionpuddles, pools, ponds of stuffbut the substance,
water. Thought of in this way, kind terms are in one way like proper
names: they designate a single entity, albeit an abstract entitya
species or a substance in these cases. I am aware that there are other
suggestions about how to construe kind terms for the Kripke-Putnam theory. I dont believe that any are better suited, but the issue
needs to be argued. All I will say here is that I think I can show that
no other way has fewer problems for their view.4
If I am correct, a problem now arises. One would suppose that
there is going to come out of the Kripke-Putnam view a distinction
between general terms which designate natural kinds and general
terms which do not. And there is at least the suggestion that rigid
designation is the key. But, taken as abstract nouns, terms that
seem intuitively to be obvious examples of non-natural kind terms
seem just as rigid as those Kripke and Putnam use as examples of
natural kind terms. Taken as abstract nouns there seems to be no
difference at all between terms we would naturally suppose to fall in
one class and terms we would naturally suppose to fall into the other
186

K R I P K E A N D P U T N A M O N N AT U R A L K I N D T E R M S

in respect to rigidity. This was noted by David Kaplan in a footnote

to his paper Bob and Carol and Ted and Alice.5
One would naturally suppose that some such term as bachelorhood would fall on the side of non-natural kind terms. But as an
abstract noun it seems to be just as much a rigid designator as
tiger or water. Rigid designation seems to provide no difference
between natural and non-natural kind terms.

IV
Now, I want to identify the argument about natural kind terms
which seems to me to be the straightforward extension of what Kripke
shows about proper names and which seems to me not to work for
general terms. I quote from a passage in Kripkes Identity and Necessity in which he seems to argue merely from the rigidity of the abstract nouns (or, to be accurate, nouns and noun phrases) that a certain
statement is an exotic necessary truth. Kripke has just been discussing
the statement that heat is the motion of molecules. He says,
To state the view succinctly: we use both the terms heat and
the motion of molecules as rigid designators for a certain external phenomenon. Since heat is in fact the motion of molecules,
and the designators are rigid, by the argument I have given here,
it is going to be necessary that heat is the motion omolecules.6

Some comments on this passage: First, Kripke takes the terms heat
and the motion of molecules as functioning in the sentence Heat
is the motion of molecules as what I have called an abstract noun
(or an abstract noun phrase). Second, he is treating the statement
in question as an identity statement. Third, when he says by the
187

E S S AY S O N R E F E R E N C E

argument I have given here I take him to mean the general argument of the paper as applied to this example. And the general
argument started out with a consideration of proper names and
includes the following passage:
If names are rigid designators, then there can be no question
about identities being necessary, because a and V will be rigid
designators of a certain man or thing x. Then even in every possible world, a and b [sic] will both refer to this same object x, and
to no other, and so there will be no situation in which a might
not have been b.7

So the argument seems to be the same one given in the case of

proper names and to depend simply on the rigidity of the terms
involved. And this may seem to go against my saying that it wont do
to sum up the view about natural kind terms by saying that suchand-such a natural kind term is a rigid designator. For if the argument about proper names is sound, this one seems sound also. But
there are problems. For example: I would suppose that both Putnam
and Kripke would consider it likely that the following statements
are, if true, necessarily true:
(a) Tigers are mammals.
(b) Water is a compound, not an element.

Neither of these, obviously, is an identity statement. So even given

that the terms mammal and a compound not an element are rigid
designators of certain kinds, the simple argument which can be applied to names cannot be used. Of course, if we already knew the
truth of some identity statement from which (a) or (b) followed,
and if the identity statement were an exotic necessary truth, then a
188

K R I P K E A N D P U T N A M O N N AT U R A L K I N D T E R M S

simple extension of the argument would give the desired result. For
instance, if we know that heat is the motion of molecules is an exotic
necessary truth then we can simply deduce that heat is some state or
other of moleculesthe latter not being an identity statement.
But we need not know any such identity statement in order to
discover (and historically I would guess this is the actual situation)
that (a) or (b) is true. I am not even sure that in the case of (a), at
least, we even now have any identity statement from which it can be
deduced that tigers are mammals.
It might be said that all I have shown is that we can know that (a)
and (b) are true without being in a position to know that they are
exotic necessary truths: that to know the latter we would have to
await the discovery of some exotic necessarily true identity statement. My point, however, will be that on the more complex theory,
as I understand it, we could know that (a) and (b) are exotic necessary truths without awaiting the appearance of some identity statement which we can know to have that property. And this is another
reason why we should not sum up Kripkes or Putnams view by
saying that this or that term is a rigid designator.

V
Now I want to imagine time-transporting John Locke to our era
and convincing him of our scientific results and also teaching him
what is meant by a rigid designator. I choose to transport Locke
because I think his view about terms for natural kinds (he uses the
expression names of substances, but the topic is the same) represents an enemy for Kripke and Putnam. Locke, on my reading, is
willing to allow that there might be something like natural kinds
(divisions in nature as a consequence of the primary qualities of
189

E S S AY S O N R E F E R E N C E

what we might now think of as such things as atoms, molecules,

genes, etc.), but that we dont have terms in the vernacular the extensions of which are determined by such divisions in nature.
Instead, he believed that while men are inclined to believe that
many of their kind terms are dependent upon nature in some such
way, in fact the extensions of our vernacular kind terms are determined by the properties each of us uses to decide on the application of them. (To use Putnams apt phrase, Locke holds that the
meaning of kind terms are in the head.) One might say that he
held that while there may be natural kinds, there are no natural kind
terms, at least in the vernacular.
But Locke transported here and now, given current philosophic
and scientific theories, could, I believe, agree with both and still
maintain his position. He could, that is, if the philosophic training
only utilizes the argument based on rigid designation and identity
that I find in the passages quoted from Kripkes earlier paperthe
argument which seems to be the simple application of the ideas
developed for proper names.
Using Kripkes example, he could admit that heat and the
motion of molecules used as abstract noun and noun phrase are
rigid designators and, given that he believes our science (or what we
are pretending here is our science), that the extensions of these
terms are the same. But he could not hold that any discovery shows
that the identity statement expressed by the sentence using these
terms as abstract nouns is true.
Given that he would suppose, as I think he would, that we have
different necessary and sufficient criteria for the correct application
of the terms heat and the motion of molecules in our minds, there
is no way for science to discover that these kinds of phenomena are
identical; in fact, there is no way in which the kinds could be identical. The situation for Locke would be the same as it is in fact for the
190

K R I P K E A N D P U T N A M O N N AT U R A L K I N D T E R M S

terms hearted thing and livered thingas abstract noun phrases

they designate different kinds, although their extensions are, in the
actual world, identical.
What Locke easily concedes is that each kind term (or expression) names the very same kind in all possible worlds and so is a
rigid designator. What he would deny is that we have discovered an
identity between kinds. Perhaps we have discovered a co-extension in
the actual world, but that wont produce the desired result.
The reason why the argument we have looked at, using rigid designation and identity alone, works with proper names and not with
kind terms is, of course, that proper names are singular and kind
terms are general terms. If two singular terms are both rigid designators and their extension is the same, then there is just one individual
that both terms designate in this, the actual, world and the same
individual must be so designated in all possible worlds.
But, if we are dealing with two kind terms, their extensions might
be the same in this, the actual, world without it being true that the
two terms are names of identical kinds. What Locke might say, then,
at this point is that co-extension of instances of two kind terms, even
though each is a rigid designator as the name of a kind, doesnt show
at all that the two kinds are identical.

VI
To obtain the same results for kind terms as Kripke obtained for
proper namesespecially the generation of exotic necessary
truthswe need something more. And both Kripke and Putnam, as
I read them, try to do this. I am going to use Putnam, but Kripke has
the same augmented apparatus. Here are two connected quotes
from Putnam about water:
191

E S S AY S O N R E F E R E N C E

To be water, for example, is to bear the relation sameL [same

liquid] to certain things.8

And
x bears the relation sameL to y just in case (i) x and y are both
liquids, and (2) x and y agree in important physical properties.9

These are two passages from the same paper in which Putnam says
that his view and Kripkes can be summed up by saying that, e.g.,
water is a rigid designator. But what new material is introduced in
these representative passages! Nothing about important physical
properties is needed or used in Kripkes arguments about proper
names to generate exotic necessary truths. Here then is the something more.
As a first shot I believe we can capture the idea in the following
principle:
(P) Something is water just in case it is a liquid and agrees in its important physical properties with the general run of stuff which English speakers call water.

(That the term water is mentioned does not, I believe, lead to

difficulties as it would were we giving a classical definition.) The
use of the general run of stuff which English speakers call water
can be explicated as follows: A word such as water comes into a
language as a word for a certain kind of stuff and ordinary users of
the word will usually be able to apply the word by having learned
to recognize certain characteristic surface propertiesroughly
those one can discern by relatively unsophisticated uses of the
senses.
192

K R I P K E A N D P U T N A M O N N AT U R A L K I N D T E R M S

One complication I will ignore is that even before scientists,

who are supposed to uncover the important physical properties,
come into the picture there may be room for what Putnam calls a
division of linguistic labor which is not reflected in (P). There may
be users of the language who, although they have a term in their
vocabulary and even may be said to know the meaning, do not
have a sufficient grasp of what surface properties to look for to be
able to recognize instances of what the term is supposed to apply to.
I think that this is probably what Putnams situation is concerning
the terms elm and beech, a situation which I share with him.
I cannot tell the difference between these two trees, but what
I needat least for such purposes as ordering from a tree farmis
not a scientist, but more likely a knowledgeable gardener. Since it
would be absurd for anyone who wants to investigate the important
physical properties of elms or beeches to consult me, even though
I am an English-speaking user of the terms in question, a modification
of (P) may be in order. But, as I say, I think we can ignore this here.
It does, however, seem in order to modify (P) so as to reflect the
important fact that Putnam wants his mechanism to function across
possible worlds. To capture this, let us modify (P) to get:
(P) For all worlds W, something is water in W just in case it is a
liquid and agrees in its important physical properties with the
general run of stuff which English speakers call water in the
actual world.

It follows from (P) that the extension of the term water in any
possible world is not determined by the surface properties which
ordinary people may use to decide whether or not some bit of stuff
is water, but rather by the important physical properties. And these,
presumably, fall within the province of scientists. Now if being H2O
193

E S S AY S O N R E F E R E N C E

is one of the important physical properties of the general run of stuff

English speakers call water, then that water is H2O is what I have
called an exotic necessary truth. Its truth is a discovery of science
and by P it is true in all possible worlds. But notice that this result
is not a consequence simply of the fact that the term water, taken as
an abstract noun, is a rigid designator, but rather from the more
complicated features of (P), in particular its use of the notion of
important physical properties of samples of stuff in the actual world.
There may seem to be another interesting and important consequence of Putnams mechanism. Since it is the important physical properties, properties to be discovered by what probably has to
be a sophisticated science, and not the surface properties, which
determine the extension of a term such as water, it may seem to
follow that prior to the rise of such a science users of the term, relying solely on surface properties, can be wrongat least in some
caseswhen even after a most careful examination they call some
bit of stuff water. For it may not be H2O. This is a result Putnam
in fact insists upon, and one I wish to raise a question about.

VII
In his article, The Meaning of Meaning, Putnam makes use of an
ingenious set of Twin-Earth examples. I want to do something
similar. I want to imagine two cultures which up to a certain point
in their history are as alike as one can make them. In particular, the
languages they speak are identical, or, at any rate, there is nothing
up to the date mentioned which affords a basis for positing a difference. I want also to imagine that at a certain period in their histories each culture develops a sophisticated science and view of the
world and that these also are identical down to the smallest detail.
194

K R I P K E A N D P U T N A M O N N AT U R A L K I N D T E R M S

Now let us take one of the terms for kinds which existed in the vernacular languages prior to the rise of science. I want to argue that
although that term, by hypothesis, would exhibit no linguistic differences in the two languages prior to the rise of science, and although the sciences developed are identical, the term could come
to have a different extension in each culture. Or, to be more accurate, I want to argue that Putnams account of natural kind terms
allows for this result.
It will be somewhat simpler for my purposes if I temporarily
shift exampleschanging water into gold. One of Kripkes examples of what I have been calling exotic necessary truths is the statement that gold has atomic number 79. I feel sure that Putnam would
regard atomic numberthe number of protons in the nucleus of
the atom of an elementas an important physical property.
If I understand Putnams mechanism correctly, he would also
agree that John Locke might have been mistaken, say, about the
composition of a ring he valued for its gold content if, in fact, the
material it was made of did not consist mainly of atoms having
atomic number 79, even if the foremost goldsmith of his day would
have, on the basis of careful scrutiny of its surface properties, said it
was undoubtedly gold.
For convenience we may as well imagine that my two cultures
exist on Earth and Twin-Earth, much as in Putnams examples. I will
imagine then that my Twin-Earth has the same history as Earth up to
some point in the earlier part of their twentieth century. In particular
there has been a group of people, including the doppelganger of John
Locke, speaking a language called by them English indistinguishable in every respect to an outside observer from the language of the
same name spoken on Earth. At some point early in the twentieth
century they developed an atomic theory, once more indistinguishable from that developed on Earth. In particular they see the atom as
195

E S S AY S O N R E F E R E N C E

having a nucleus composed of positively charged particles, which

they call protons, and neutrally charged particles, which they call
neutrons. In their laboratories their scientists speak of elements
and use this term in the same way that our scientists have, for those
non-compound substances whose atoms have a particular number of
protons in the nucleuswhat they and we call the atomic number.
They also recognize, and use the word, isotopes, isotopes being individuated by the combined number of protons and neutrons in the
nucleus of the atomthat number being the isotope number.10 They
also have, as our scientists have, come to realize that the same element may have, and usually will have, several different isotopes and
they give them names for use in the laboratory as our scientists have
given the names protium, deuterium, and tritium to the three
known isotopes of what we call hydrogen. These similarities are all
that I really need for the point I wish to make, so for the rest of the
two developments in science, imagine that Twin-Earth and Earth are
identical. I have thus, I believe, so built my story that the scientific
picture of the world is the same in the first parts of the twentieth
century for both cultures.
Although I believe that what I will now imagine is not strictly
speaking necessary for what I will propose, it may help to make the
situation psychologically more plausible. I imagine then that on
Twin-Earth not only do elements usually have several isotopes, but
also it is a general rule that one of the isotopes of a particular element makes up the bulk of the element as it occurs in naturethe
other isotopes being fairly rare. I do not off-hand know whether this
situation in fact obtains on Earth. If it does then I have made no
change as yet between Earth and Twin-Earth. But even if it does
not, the change I will have made is simply one concerning the
relative distribution of things and will have nothing to do with a

196

K R I P K E A N D P U T N A M O N N AT U R A L K I N D T E R M S

difference in scientific outlook, any more than the fact that we have
less and less oil changes our scientific theories.
Now what isotope one is dealing with makes a big difference in
how one will expect some bit of matter to behave. Just to mention a
couple of the few things that I know about it, it is often the case that
some isotopes of an element are radioactive and others not, and this,
of course, has had profound consequences. Also some isotopes of
the same element are unstable and break down into an isotope of
another element. I believe this is the basis of carbon 14 dating. Of
course, what element is in question is also of very great importance,
especially in chemical reactions. But it might be a close question as to
whether isotope number or atomic number has more importance.
Now with these suppositions and facts, it seems to me not psychologically implausible for my Twin-Earthlings to be more taken
with, so to speak, the isotope number of a bit of substance rather
than with its atomic number and also not implausible for them to
diverge from our practice and to identify the substance designated
by some of their vernacular natural kind terms not with a certain
element, but with the isotope which makes up the bulk of what had
been previously called by that term. Hence, for example, they identify gold, not with the element having atomic number 79, but with a
certain isotope having a certain isotope number. The rare isotopes
of the element having number 79 would then be dismissed as not
really being gold, although, to be sure, in various ways very much
like gold.
If I have described a possible Twin-Earth situation, then were
my doppelganger on Twin-Earth to ask someone in the TwinEarth UCLA chemistry department what gold is, he would be
told that it is a substance having such-and-such isotope number,
while I, in the analogous situation, would be told that it is a

197

E S S AY S O N R E F E R E N C E

substance having atomic number so-and-so. It should be obvious

that if I and my doppelganger defer to scientists as Putnam supposes we would, the extension of his term gold and of mine will
at this point diverge.
My story can, obviously, be extended to the vernacular term
water. On Twin-Earth, in my story, because isotopes are taken more
seriously for one or another practical or historical reason, we can
suppose that Twin-Earthlings will identify water with protium oxide
and exclude what we call heavy waterdeuterium or tritium
oxide.
After the scientific discoveries and the mapping of non-scientific
kind terms onto them, there will be a difference between TwinEarthlings and Earthlings in regard to the truth-value of what certain sentences express. For example, there will be a sentence of the
form Some gold has isotope number x and some gold has isotope
number y which Earthlings will take to express a truth and TwinEarthlings a falsehood. Correspondingly, while the sentence Gold
has atomic number 79 will be regarded as expressing a truth by
Twin-Earthlings provided has does not carry with it the notion of
identity, Gold is identical with the element having atomic number
79 will be regarded by them as expressing something false, while
Earthlings will regard it as expressing something true.
We cannot immediately conclude that in my story it turns out
that Twin-Earthlings regard certain propositions as true which Earthlings regard as false and vice versa, for it is open to one to say that the
corresponding sentences do not express the same thing. In fact, in
my story it would be paradoxical to conclude that the two parties
have different beliefs about what is true or false because that would
imply that there is a right and a wrong in the matter. And there does
not seem to be any way of showing that either party is right while
the other is wrong.
198

K R I P K E A N D P U T N A M O N N AT U R A L K I N D T E R M S

My talk about mapping ordinary language terms onto science

may suggest a historical two-stage process: first there is science and
its divisions of things into kinds, and then there is the hookup of
natural language termswith some leeway in theory, if not in psychological reality, about how this second stage is carried out. I do
not believe I need suppose this to be possible. Perhaps science has
to begin with questions about kinds of things couched in ordinary
language: What is the nature of gold?, What is the nature of
water?, etc. Even if that is so, Twin-Earthlings in my story come up
with verbally different things from Earthlings even though they
start from what is to all appearances the same language and even
though their science turns out to be identical. (Even if science must
begin from questions couched in ordinary language, it soon develops its own terms for natural kinds: The names of elements low on
the periodic chart, in our science, are taken from ordinary language;
those discovered later have invented names, and there are vastly
more invented names for compounds than ones derived from ordinary language.)
What do I conclude from my story? I do not draw the conclusion that Putnam has failed to describe how natural kind terms in
the vernacular function. The story does not show that. But I do think
that two important points emerge. First, there is a certain slackness
in the machinery which Putnam does not, I feel, prepare us for. In
my story science develops apart on Earth and Twin-Earth. The
slackness comes from how ordinary language terms for kinds are
mapped onto the same classifications. In my story I have envisaged
only a small wobble; how much latitude there might be in theory I
do not know. The point is that we can agree with Putnams theory
about how natural kinds terms function in ordinary language and
still see that we might have done things differently even with the
very same results.
199

E S S AY S O N R E F E R E N C E

The second consequence flows out of the first. Putnams view,

it will be recalled, is that the extension of a term such as gold or
water would be the same before discoveries about important
physical properties as it was afterward. John Locke would have
been wrong, however justified he might have been, had he thought
something to be gold or water which did not have the important
physical properties discovered later. My strong inclination is to
agree with Putnam that when I see the words gold and water in
Lockes Essay on the Human Understanding there is no change of
meaning. And I also am strongly inclined to agree that I would
not count something as, strictly speaking, water unless it were
H2O, or as gold unless it had the atomic number our scientists say
that gold has. But I am an Earthling. In my story Twin-Earthlings
do things slightly differently, although for no reasons having to do
with a different linguistic basis or a different science. What should
we then say about the extension of such ordinary language terms
for kinds?
If you accept that my story of Earth and Twin-Earth contains no
inconsistencies or other mistakes, then I do not see how we can
accept Putnams view that it is clear that natural kind terms in ordinary language have the same extension before and after scientific
discoveries and the mapping of those terms on to those discoveries.
The slackness I have talked about seems to allow that from the
very same linguistic base we may, after the very same scientific discoveries, move in different directions. Which way we move will
change the extension. How can we then say that, given that we have
moved in one direction, the users of the ordinary language terms
might just have been dead wrong in applying the term to some particular instance? Lockes ring, which he took to be gold, might have
been made of stuff having atomic number 79. In my story, the Earthling Putnam would say that Locke was correct. But suppose his ring
200

K R I P K E A N D P U T N A M O N N AT U R A L K I N D T E R M S

were made of one of the rare isotopes, as we would say, of gold.

Twin-Earthlings would say the ring wasnt made of gold at all. Putnams doppelganger would conclude that Locke was dead wrong.
But nothing about language or science explains this difference.
You might object at this point that I have not kept my TwinEarthlings as much the same as us as I have made it seem. After all,
they do take a different turn; and must this not mean that they had
at least one psychological difference from us?they had the disposition to be struck more by isotopes than by elements. And perhaps this disposition in some way ought to count as showing a
difference in the language, however much it looked to be the same
as ours prior to scientific discoveries. Even if we admit that there
was at some point a disposition on their part which we do not
have, the disposition may not always have been there. We need not
suppose that because the twentieth-century Twin-Earthlings had
this disposition, the eighteenth-century ones did also. Nor do we
need to suppose that there was such a disposition even in twentieth-century Twin-Earthlings prior to their looking at their scientific theory. Thus we seem to have no reason to think that their
John Locke and ours differed psychologically or linguistically. Or
rather we have none unless we accept what seems to be an outrageously bizarre view of languagethat the extension of ones
terms may be determined by the psychological quirks of some
people several centuries hence.
There is one point about the scientific situation in regard to
atomic theory that I have left out of my story. Historically, the discovery of the possibility of isotopes, and of the big difference it makes
which one is involved, carne after the development of the simpler
atomic theory which included the notion of elements and atomic
numbers. Possibly this historical accident had an influence on the
way in which we mapped our vernacular natural kind terms onto sci201

E S S AY S O N R E F E R E N C E

ence. But this would be of little comfort to Putnam, I should think. It

seems just as bizarre to suppose that the extension of ones terms
should depend upon such future historical accidents as it does to
suppose that they depend upon future psychological quirks.
To sum up then, if we go along with Putnam a certain distance
we seem either to have to embrace unacceptable views about language or to admit that nature, after all, does not fully determine the
extension of vernacular natural kind terms, and science is not wholly
responsible for discovering their true extensions.

NOTES
* Norman Malcolm has been a great influence on my philosophical thinking and
remains so. I wish that my contribution to this volume in his honor were on a
topic for which he is well known. A seminar on dreaming which I gave last year
did not result in anything which I considered an original contribution on my
part. Hence, this contribution of mine simply represents my own current philosophical interest. It benefits from discussion with, among others, Jay Atlas,
Tyler Burge, David Kaplan, and Jon Wilwerding. The latter has independently
developed an example somewhat like the one given by me in the last section,
but from which he has been able to derive somewhat stronger conclusions than
I do here.
1. For Kripkes views I use Identity and Necessity in M. K. Munitz, ed., Identity
and Individuation (New York: New York University Press, 1971), pp. 13564; and
Naming and Necessity in Davidson and Harman, eds., Semantics of Natural
Language (Dordrecht: D. Reidel, 1972), pp. 253354. For Putnams views I use
The Meaning of Meaning in his Mina, Language and Reality (Cambridge:
Cambridge University Press, 1975), vol. 2, pp. 21517.
2. I introduce the expression exotic necessary truths not just to dramatize the interest of Kripkes discovery. The more obvious term a posteriori truths obscures
an important point. If we distinguish a sentence from the proposition it expresses
then the terms truth and necessity apply to the proposition expressed by a
sentence, while the terms a priori and a posteriori are sentence relative. Given
that it is true that Cicero is Tully (and whatever we need about what the relevant
sentences express) Cicero is Cicero and Cicero is Tully express the same
proposition. And the proposition is necessarily true. But looking at the proposition through the lens of the sentence Cicero is Cicero the proposition can be

202

K R I P K E A N D P U T N A M O N N AT U R A L K I N D T E R M S

seen a priori to be true, but through Cicero is Tully one may need an a posteriori investigation.
3. The Meaning o Meaning, pp. 230 and 231.
4. For one alternative to my assumption see Monte Cook, If Cat Is a Rigid Designator, What Does It Designate? Philosophical Studies, 37, no. 1 ( January 1980).
If Cook is correct much of what I say is wrong. I believe I can counter him, but to
do so requires more than I can put into a footnote.
5. Hintikka, Moravcsik, and Suppes, eds., Approaches to Natural Language,
pp. 51831.
6. Identity and Necessity, p. 160.
7. Ibid., p. 154.
8. The Meaning of Meaning, pp. 23839.
9. Ibid., p. 239.
10. For the sake of simplicity I am pretending that isotopes are distinguished by their
isotope numbers alone. To accord with atomic theory as we know it I need a
combination of atomic number and isotope number. This would complicate the
telling of the Twin-Earth story, but the same point would emerge.

203

This page intentionally left blank

SUBJECT INDEX

abbreviation 147, 14957, 173, 1756.

abstract entity (and abstract noun) 1867,
1901, 194.
ambiguity 202, 1423, 146.
pragmatic 21.
semantic 21, 142.
Syntactic 21, 142.
amended statement 2930.
anaphoric chain (also anaphora,
anaphorically) 115, 1224, 1267,
12931, 135, 1379, 141, 143.
a posteriori 149, 176, 2023.
a priori (also knowledge -, known -,
knowbale -) 14749, 151, 1589,
1689, 2023.
contingent - 14751, 1589, 162,
1689, 173.
atomic number (see also isotope
number) 176, 1958, 2001, 203.
attributive (use) (also attributively, attribute)
xv, 716, 18, 206, 28, 328, 47, 52, 115,
117, 118, 120, 123, 127, 135, 137, 142,
145.
audience 79, 126, 189, 28, 324, 416,
69, 77, 114, 11921, 123, 125, 129,
1326, 13940, 1425.
backing of descriptions,
see description, backing of.

baptism
see dubbing.
block 1039, 111.
calling attention (to a person or thing)
(see also having in mind) 7, 83.
causal connection 76, 81, 99, 113, 157.
contingent (see also a priori, contingent -,
and necessary) 147, 149, 154, 161,
1635, 168, 176.
(linguistic) convention (see also practice) 31,
43, 149, 162, 165.
definite article 1289, 142.
definite description see description
denotation (also denote, denoting) 34, 167, 534,
59, 61, 656, 68, 78, 124, 126, 174, 176.
de re
- attitude 166, 170, 1734.
knowledge - 1656, 174.
description (also description, describe,
describing) (see also misdescription)
(backing of) - 4951, 53, 603, 769,
94, 100.
composite - 568.
concealed - 51, 53, 82, 89, 923, 1012,
10910, 112.
definite (also -, identifying -.
introducing-) viixv, xix , 330, 3241,
437, 4979, 82, 879, 92102, 105,
205

SUBJECT INDEX

description (continued)
10810, 112, 11521, 12335,
13742, 1445, 1478, 15061, 164,
166, 168, 1701, 173, 1756, 183.
indefinite - (also -) 128, 133, 135, 13741,
146.
indefinite definite - x , xiixiii, 46.
question-begging (and non-question
begging -, also secondary -) 54, 612,
64, 66, 73, 123.
What would the user(s) of the name describe
in this way? 70, 76.
What would the speaker describe in this way
on this occasion? 70.
descriptive content xiv, 52, 88, 114, 129,
1334, 166.
de re propositional attitude 1656, 170,
1734, 176.
determine the referent (also the extension) 32,
40, 42, 52, 68, 70, 77, 94, 99, 117,
121, 123, 1267, 129, 133, 134,
137, 142, 148, 157, 180, 190, 1934,
2012.
determine the truth value 6, 1367, 141, 144.
Donnellans most famous examples
her husband 226.
Smiths murderer 89, 112, 15, 18,
38, 47.
the book is on the table 1011.
Jacob Horne 837, 92, 113.
J. L. Aston-Martin 6870, 79.
the king (the usuperer) 14.
the man drinking a martini ix , xiv,
xviixix, 911, 18, 21.
the man carrying a walking stick 1920, 47.
the rock 39, 41, 435.
the table 101, 32, 39, 46.
dubbing (also baptism) 99, 114, 170.
expectation (also expect) 8, 414, 47, 97,
120, 136, 13942, 144.
extension 1812, 1856, 1901, 195, 198,
2002.
fiction (discourse about -) 29, 835, 95, 104,
107, 113.
fix the reference 147, 151.

general terms 179, 1867, 191.

having in mind 8, 910, 29, 38, 612, 73, 81,
11621, 124, 126, 134, 1369, 1412,
1445.
historical connection (see also omniscient
observer of history, and causal connection)
76, 813, 878, 92, 96100, 1048,
1102, 157, 174, 189.
reference in historical context 713, 75, 84,
92, 969, 1037, 109, 112, 133, 167,
189, 1945, 198, 2012.
Humpty Dumpty 312, 405.
identify (also identification) (see also principle
of identifying descriptions, refer) 9, 15,
20, 24, 26, 28, 324, 46, 51, 5760,
623, 657, 74, 767, 79, 96, 104, 116,
119, 1323, 181, 1978.
intention (also intend, intending) (see also
speakers intention) xiv, xviixx , 3947,
835, 95, 97, 104, 106, 123, 1336,
14664, 16676, 184.
introduction (also introduce) of a name, a term,
334, 56, 59, 79, 88, 104, 13940, 147,
1501, 1556, 1589, 168, 172.
introductory indefinite description 1323,
13541, 143, 146.
irrelevant (ostensible referring
expression) 33,4, 3940, 47.
isotope number (see also atomic
number) 1968, 203.
misdescription (see also description) 145,
39, 47.
meaning 31, 403, 501, 106, 149, 155, 162,
176, 1812, 190, 1934, 200.
-change 182, 200.
name (also (ordinary) proper name) 7, 15,
257, 45, 4979, 82, 84, 879, 91114,
1223, 1367, 14759, 1618, 1707,
1956, 199.
name (in the narrow, or strict, logical sense)
(also genuine (proper) -, ) xiiixvii,
xixxx, 4, 27, 534, 82, 8791, 93,
1023, 18292.

206

SUBJECT INDEX

name of substance (see also natural kind

term) 189.
natural kind 176, 190.
natural kind term 17980, 184, 18690,
195, 197, 199202.
near miss (and missed by a mile) 35, 379, 47.
necessary (see also contingent) 149, 151,
17980, 1834, 1879, 1912, 1945,
202.
Neptune 14858, 165, 1678, 177.
noun phrase 128, 187, 1901.
existential and negative existential statement
See statement
omniscient observer of history 75, 967.
particular (- individual, - entity, - person, - someone
or something) 913, 234, 289, 32,
389, 49, 61, 89, 92, 113, 116, 1189,
1279, 1324, 136, 1389, 141, 143,
146, 196.
possible world 147, 14950, 163, 166,
1756, 1834, 186, 191, 1934.
practice (see also convention) 38, 413, 150,
162, 197.
predicate (also predicated) 38, 71, 76, 78,
956, 103.
presuppose or imply (presupposition or
implication) 58, 146, 18, 223,
2830, 35.
presuppose (presupposition) 9, 112, 36,
835, 100, 138, 143.
principle of identifying descriptions (also set of
identifying descriptions) (see also
backing of descriptions, identify,
identifying statement, refer) 4977, 79,
93, 95, 97, 99102, 105, 108, 157.
counter-example to - 50, 55, 61, 63, 658,
70, 76, 78, 100.
properties
important physical - 1924, 200.
surface - 1935.
proposition 84, 889, 912, 1012, 106,
10812, 114, 1435, 166, 202.
general - 28, 901.
propositional identity 91.
singular - vii.

recognize (also recognizing, recognition) 15,

38, 42, 44, 47, 60, 1201, 123, 133,
1356, 140, 142, 144, 1923, 196.
refer (also reference, referent, referential,
referentially, referential use, referring,
referring use) (see also fix the reference)
viiixi, xiiixxi, 341, 437, 4966,
6879, 81101, 1036, 10828, 130,
13451, 154, 156162, 16566, 168,
171, 173, 1756, 179, 185, 188.
- in a direct way 534.
expression - 90, 115.
people - 115.
referential/attributive 115, 118, 120, 127,
137, 142.
referential context 1234, 135, 139, 1435.
rigid designator (also rigidly designate, rigidly
designating) 147, 14951, 15361,
1634, 16671, 1746, 18392, 194.
Santa Claus 86, 95, 101, 1045, 107, 109,
1112.
satisfies the description (also satisfying the
description, satisfies the condition)
(see also talk about) 34, 38, 46, 55, 57,
63, 82, 105, 118, 157.
say of (also saying of ) 19, 22, 257, 177.
semantic reference (also semantic referent)
(see also speakers reference) xxi,
1156, 118, 121, 124, 1267, 1345,
1378, 140, 142, 145.
sense (Freges Sinn) 501.
simple sentence (containing a singular term)
90, 93.
singular term xiv, 82, 878, 902, 122, 138,
146, 179, 183, 185, 191.
single out 10, 98, 128, 138.
speak about (see also talk about) 9, 28, 84,
13940.
speakers intention (also speaker
intends) 201, 40, 11921, 129,
1336, 13940, 145.
speakers reference (also speaker reference,
speakers referent, speaker refers, speaker
referring) (see also semantic reference)
18, 245, 37, 3940, 44, 46, 60, 69, 97,
11421, 1237, 13539, 1412, 145.

207

SUBJECT INDEX

speech act 15, 35, 46, 76, 85, 95, 1178.

standard meter bar 163.
state of affairs 150, 1623.
statement
(the -) 24, 26, 2930, 46.
existential - 83, 94, 1003, 1068,
1102, 114.
negative existential - 55, 812, 856, 889,
92, 111, 113.
predicative - 836, 96, 101, 10610,
1123.
- about 89, 91.
identity - 81, 108, 110, 18790.
stipulation (also stipulate, stipulating) 42,
150, 1548, 1616, 16870, 1734,
176.
substance 186, 189, 1968.

talk about (also talking about, think about)

(see also speak about) viii, x, xiii, 3, 9,
337, 3940, 446, 60, 64, 66,
835, 90, 113, 1201, 134, 140,
150, 164, 172, 181, 1856.
true
saying something true (also expressing
or stating something true) 15, 18, 22,
24, 26, 2930, 35, 378, 84, 94,
15960, 172, 198.
truth 104,
contingent - 154, 161, 1635, 1689, 173,
176.
truth-condition ix , xix , 94, 103, 106, 10811,
1367, 13940, 143, 146, 166, 170,
174.
Twin-Earth 194201, 203.

208

AUTHOR INDEX

Acherman, Diana 176.

Adair, Douglass 113.
Albritton, Rogers 79, 177.
Almog, Joseph xx.
Anscombe, Gertrude E.M. 172.
Aristotle 513, 569, 64, 714, 1067,
156, 175.
Atlas, Jay 202.
Bacon, Francis 756.
Budlong, Theodore 78.
Burge, Tyler vii, xi, 113, 202.
Canfield, John 28.
Carriero, John xi.
Caton, Charles 29.
Chastain, Charles 122, 127, 12930, 1389,
146.
Cook, Monte 203.
Coolidge, Sarah xi.
Cumming, Sam xi.
De Galle, Charles 67.
Donnellan, Keith viixi.
Dummett, Michael 150, 152, 154, 156,
158.
Evans, Gareth 160.

Foot, Philippa xviii.

Frege, Gottlob 51- 556, 58, 81, 108, 110,
113, 150, 175.
Geach, Peter 11621, 1234, 137, 139, 144.
Goldwater, Barry 17.
Grice, Paul 31, 42.
Herman, Barbara xi.
Herodotus 714.
Kamp, Hans xi.
Kaplan, David vii, xi, 78, 93, 157, 164, 170,
1735, 177, 202.
Kneale, William 160.
Kripke, Saul A. v, x , xx, 789, 113,
14552, 154, 156- 59, 163, 166,
1689, 1756, 17980, 18292,
195, 202.
Leonardi, Paolo xi, xx
Le Verrier, Urbain 14953, 155, 157,
1645, 167.
Linsky, Leonard 223, 26, 2930.
Locke, John 182, 18991, 195, 2001.
MacKay, Alfred F. 312, 347, 3941, 436.
Malcolm, Norman 202.

209

AUTHOR INDEX

Martin, Tony xi.

Meinong, Alexius 92.
Middleton, Arthur P. 113.

Searle, John R. 513, 557, 64, 93, 113.

Shakespeare, William 756.
Shoemaker, Sidney 28, 179.
Strawson, Peter F. xv, 37, 11, 168, 203,
2830, 46, 49, 56, 59, 612, 64, 78,
115.

Perry, John xi, 78.

Putnam, Hilary 17980, 182, 1846,
18895, 198200, 202.

Thales xiv, 715.

Russell, Bertrand xi, xv, xx , 37, 167, 201,
23, 278, 46, 516, 58, 64, 78, 812,
8794, 1012, 112, 1156, 1289, 134,
1435, 168.
Sachs, David 45.
Schriffin, Seana xi.

Vendler, Zeno 12930, 13,

1389.
Wettstein, Howard xi, xvi, xix.
Wilwerding, Jon 202.
Wittgenstein, Ludwig 32.

210