Professional Documents
Culture Documents
Harold Noonan
First published 2013
by Routledge
2 Park Square, Milton Park, Abingdon, Oxon OX14 4RN
The right of Harold Noonan to be identified as the author of this work has been asserted by him in
accordance with sections 77 and 78 of the Copyright, Designs and Patents Act 1988.
All rights reserved. No part of this book may be reprinted or reproduced or utilised in any form or by
any electronic, mechanical, or other means, now known or hereafter invented, including
photocopying and recording, or in any information storage or retrieval system, without permission in
writing from the publishers.
Trademark notice: Product or corporate names may be trademarks or registered trademarks, and are
used only for identification and explanation without intent to infringe.
Typeset in Garamond
by Taylor & Francis Books
Contents
ACKNOWLEDGEMENTS
1 Introductory overview
2 The background
Frege on sense
Russell on descriptions and names
The Frege–Russell synthesizers
Quine
3 Naming
The target
Giving the meaning, fixing the reference and rigid designation
The modal argument
The debate over the modal argument
The arguments against the cluster theory qua theory of reference-fixing
The historical chain picture
4 Necessity
5 Extensions
GLOSSARY
BIBLIOGRAPHY
INDEX
Acknowledgements
Russell made two contributions to the topic Kripke is concerned with: his
quantificational theory of descriptions and his description theory of
(ordinary) proper names. The two are logically independent: Kripke argues
against the second but raises no objection to the first; Frege, on the other
hand, although he is with Russell in assimilating ordinary proper names to
descriptions, does not group descriptions with quantificational phrases as
Russell does – in fact, Frege’s own account of descriptions is one of the
primary targets of Russell’s attack in ‘On Denoting’.
I begin with Russell’s theory of descriptions and then sketch his theory of
ordinary proper names. As we shall see, despite its centrality in Kripke’s
discussion, the latter is not much more than an appendix in Russell’s
treatment.
Russell’s theory of descriptions is most famously presented in ‘On
Denoting’. Here he regards himself as simultaneously arguing against Frege
as well as his own earlier self.
The easiest approach to Russell’s criticism of Frege and the development
of his theory of descriptions is via the problem of empty singular terms.
These pose a problem for Frege because he is led to conclude that sentences
containing such terms must be truth-valueless (except in cases where the
terms have their indirect referents); a conclusion he thought tenable
because, given the sense/reference distinction, he was able nonetheless to
ascribe such terms a sense and regard sentences containing them as
expressing thoughts.
However, in the case of empty definite descriptions like ‘the least rapidly
convergent series’ this position is deeply implausible. Russell points this
out in his first criticism of Frege (in which for ‘meaning’ read ‘sense’ and
for ‘denotation’ read ‘reference’):
… consider ‘the King of France is bald’. … this phrase, though it has a meaning … has certainly no
denotation, at least in any obvious sense. Hence one would suppose that ‘the King of France is bald’
to be nonsense, but it is not nonsense, since it is plainly false. Again, consider such a proposition as
the following: ‘If u is a class with only one member, that one member is a member of u’… This
proposition ought to be always true. But ‘the u’ or ‘that one member’ is a denoting phrase, and it is
the denotation, not the meaning, that is said to be a u. Now if u is not a unit class, ‘the u’ seems to
denote nothing, hence our proposition would seem to become mere nonsense as soon as u is not a
unit class.
(Russell 1956: 46)
This is not as clear as one might like (it is not, of course, a consequence of
Frege’s position that ‘the King of France is bald’ is nonsense), but the
general point is clear: Frege is committed to saying, both of subject-
predicate sentences like ‘The King of France is bald’ and of truth-
functionally complex sentences like ‘If u is a class with only one member,
then that one member is a member of u’, containing non-denoting definite
descriptions, that they are neither true nor false. However, that is
unacceptable, for such sentences can be false (the former case) and even
true (the latter).
The problem arises for Frege because he groups definite descriptions
with proper names, ordinarily so-called, and regards both types of
expression as having objects as their references. By contrast, he treats
quantifying expressions, that is, expressions which can be used in answers
to the questions of the form ‘How many Xs?’–‘at least one X’, ‘some X’,
‘at most one X’, ‘exactly three Xs’, ‘any X’, ‘all Xs’– as having as their
references properties of properties (in his terminology second-level
functions from first-level concepts to truth-values). Thus for Frege
quantificational statements, i.e. ‘How many?’ statements, are statements
about properties of objects, not statements about objects themselves. Hence
he analyses ‘Some King of France is bald’ as stating that a relation of
coinstantiation holds between the properties being a King of France and
being bald and not as making any statement about any object which is a
King of France.
Russell’s solution to the problem of empty definite descriptions is
essentially that definite descriptions should be treated, like ‘some King of
France’ on Frege’s theory, as quantifying expressions. We can see what ‘the
King of France is bald’ actually says, and what its true logical form is, if we
split it up into three components:
But the most important point to appreciate for our purposes is that, despite
his putting it this way himself, it is in fact misleading, without qualification,
to describe Russell as saying that names abbreviate descriptions. As
indicated above, he thinks that although every time a speaker uses a proper
name he must have some identifying description in mind, this may vary
from speaker to speaker and time to time. He does not think that for every
name there is a description which for every speaker on every occasion
‘gives the meaning’ or ‘fixes the reference’ of the name. He does not, in
short, think of names as having a shared public meaning at all. In fact, he
thinks that from speaker to speaker and time to time the only thing constant
so long as the name is rightly used is the object to which the name applies.
Public language is of no interest to Russell except as an overlapping of
idiolects. A striking passage from Russell’s Lectures on Logical Atomism
makes this clear:
When one person uses a word, he does not mean by it the same thing as another means by it. … It
would be absolutely fatal if people meant the same thing by their words. It would make all
intercourse impossible … Because the meaning you attach to your words must depend on the nature
of the objects you are acquainted with, and … different people are acquainted with different objects.
(Russell 1918, also in 1956: 195)
In the late 1950s and early 1960s various writers put forward accounts of
naming inspired by Frege and/or Russell. Two of the most important,
discussed by Kripke in Naming and Necessity, were Searle and Strawson.
Searle’s article ‘Proper Names’ (1958) is the locus classicus of the cluster
theory of descriptions which Kripke sets up as his most serious target. It is
intended to bring out what is shared by a community using a proper name
despite variation in descriptive associations. Searle’s contention is that
names are not descriptions, but ‘pegs’ on to which we can hang
descriptions. Referring uses of proper names do not specify any
characteristics of the object referred to, but Searle answers the crucial
question, how a name is connected to an object, by saying that the
descriptive force of ‘This is Aristotle’ is that a sufficient but so far
unspecified number of the statements regarded by users of the name as
stating essential and established facts about him are true. This is something,
he thinks, that reflects the essential function of names. For names would be
superfluous if they were logically equivalent to descriptions:
the uniqueness and universal pragmatic convenience of proper names in our language lies precisely
in the fact that they enable us to refer to objects without being forced to raise issues and come to
agreement on what descriptive characteristics exactly constitute the identity of the object. They
function not as descriptions, but as pegs on which we hang descriptions.
(Searle 1958: 167)
So, Searle says, it is not a necessary truth that Aristotle ever went into
pedagogy, as it is a necessary truth that the teacher of Alexander did, but ‘it
is a necessary truth that Aristotle has the logical sum, inclusive disjunction,
of the properties commonly attributed to him: any individual not having at
least some of these properties would not be Aristotle’ (1958: 168).
Searle thinks the answer to the question whether ordinary proper names
have sense is thus ‘yes and no’. ‘No’ because they are not used to specify or
describe characteristics of objects; ‘yes’ because they are associated ‘in a
loose sort of way’ with characteristics of the objects to which they refer.
The central contention of Searle’s article, which Kripke captures by
talking about names being associated with a weighted cluster of
descriptions, is that there is a vagueness or fluidity in our use of proper
names, which does not exist in the case of descriptions. Kripke’s arguments
are intended to show that this is not a far enough departure from the views
of Frege and Russell, if it is a departure at all, since we allow for the
possibility of far greater ignorance and error than Searle can accommodate.
The second significant discussion of naming after Frege and Russell to
which Kripke pays attention is Strawson’sin Individuals (1959). What
Strawson adds is the idea that speakers can use names deferentially to
others and so refer successfully even though individually they are woefully
ignorant or substantially erroneous in their views.
Strawson insists on the principle already noted from Donnellan that a
user of a name must be prepared to substitute a description for the name and
that ‘there must be some description [the speaker] could given which
applies uniquely to [the object] he has in mind’ (1959: 181–2). But he
explicitly notes, first, that this description need not be in purely general
terms (a point which is already implicit in Frege, Russell and Searle) and,
second, that the description may include a reference to another’s reference,
so ‘one reference may borrow its credentials from another and that from
another’ (1959: 182).
With these qualifications and additions to the Frege–Russell position we
have reached the description theory Kripke sets up as his target in Naming
and Necessity: names are associated by speakers with vague clusters of
descriptions, which may include demonstrative or indexical elements, may
differ from speaker to speaker and may incorporate deference by some
speakers to others.
Even with the qualifications introduced, Kripke argues, this theory can be
seen to be fundamentally flawed, and its fundamental error is precisely
captured in the principle of identifying descriptions expressed in Strawson’s
statement: ‘one cannot significantly use a name to refer to someone or
something unless one knows who or what it is that one is referring to by
that name’ (1959: 181–2).
Quine
and
Ralph believes that the man at the beach is a spy.
Quine says that these sentences are ambiguous. We can read them as
expressing the two-place relation of believing between a person and a
proposition. Then the first can be true and the second false consistently with
the identity of the man in a brown hat and the man at the beach. Or we can
read them as expressing the three-place believing of relation. Then they
have the same truth-value given the identity. But this is not a problem,
Quine thinks, because when we read ‘Ralph believes the man at the beach is
a spy’ in the second way, it does not entail what it says read in the first way.
So the principle of extensionality is adhered to: every singular term can be
replaced salva veritate with any other singular term for the same object
wherever it refers.
But Quine is not content with this treatment, since it requires an appeal to
intensions, which can only be defined in terms of the obscure notion of
synonymy. So he suggests that, just as we can replace the (lower-case)
statement operator ‘nec’, in the modal case, with the (upper-case)
semantical predicate ‘Nec’, so we can regard belief as a relation between
people and linguistic items: statements and predicates (or closed and open
sentences). So ‘Ralph believes the proposition that the man in a brown hat
is a spy’ is translated as ‘Ralph believes-true “the man in a brown hat is a
spy”’; ‘Ralph believes of the man in a brown hat the intension being a spy’
is translated as ‘Ralph believes-true “x is a spy” of the man in a brown hat’.
Being explicit about the language we get as a translation of the first
sentence: ‘Ralph believes-true “the man in a brown hat” as that is
understood in English’, and as a translation of the second ‘Ralph believes-
true “x is a spy” as that is understood in English of the man in a brown hat’.
Thus it seems that Quine has been able to establish the difference
between de re and de dicto ascriptions of propositional attitude without
offending against his policy of extensionality. De re ascriptions of
propositional attitude turn out to be no worse off than de dicto ascriptions.
But, in fact, matters are more complex. Suppose a de dicto ascription is
true: Ralph believes the proposition that the man in a brown hat is a spy.
Does a de re ascription follow (given the existence of the man in the brown
hat)? Does Ralph believe of the man in a brown hat, i.e. the man at the
beach, the intension that he is a spy? In ‘Quantifiers and Propositional
Attitudes’ Quine says ‘Yes’. He puts this by saying that exportation from de
dicto propositional attitude claims to de re propositional attitude claims
must be regarded as implicative. This, for Quine, is crucial, because it gives
us an account of when an ascription of a de re propositional attitude claim is
true. It follows that if Ralph believes de dicto that the man in a brown hat is
a spy and also believes de dicto that the man at the beach is a law-abiding
citizen then he believes de re of the man in a brown hat, i.e. the man at the
beach, that he is a spy and also believes de re of the man at the beach, i.e.
the man in a brown hat, that he is a law-abiding citizen.
A comparable account in the modal case would involve that if it is
necessarily true that 9 > 5 and necessarily true that the number of planets
numbers the planets, then it is necessarily true of 9, i.e. of the number of
planets, that it is greater than 5 and also necessarily true of the number of
planets, i.e. of 9, that it numbers the planets, and, in fact, that any concept 9
numbers is one it necessarily numbers. No defender of a non-trivial
distinction between essential and accidental properties could say this, of
course, so Quine’s account of when a de re propositional attitude ascription
can be considered true does not provide his opponents in the debate over de
re modality with a rejoinder to his arguments.
However, a problem about exportation (due to Sleigh 1968; see also
Kaplan 1969) causes Quine to reject the position that de re propositional
attitude ascriptions can be objectively true or false. Suppose Ralph believes
correctly that there are spies. He also believes correctly that no two are
exactly the same height. So he believes that there is a shortest spy. Hence he
believes that the shortest spy is a spy. If we export we must say that there is
someone, the shortest spy, call him ‘Shorty’, of whom Ralph believes that
he is a spy. But this seems absurd. Surely, if there is someone of whom it is
true that Ralph believes that he is a spy, Ralph should be of interest to the
FBI. But he is of no interest to the FBI. Without exportation, however,
Quine has no general account of when de re ascriptions of propositional
attitude are true.
Others have suggested a way out. The idea is that we accept exportation,
but restrict the singular terms on which it is allowed. We distinguish, for
example, ‘the man in the brown hat’ from ‘the shortest spy’ and allow that
if Ralph believes that the man in a brown hat is a spy it follows that there is
someone of whom Ralph believes that he is a spy, but deny that it follows
from the fact that Ralph believes that the shortest spy is a spy that there is
someone he believes to be a spy. Kaplan (1969) suggests this idea and
introduces the terminology of a ‘vivid name’. So to have a de re belief
becomes to have a de dicto belief that can be expressed with a vivid name.
From Quine’s point of view, however, the notion of a vivid name is
wholly context-relative. The ability to use a vivid name for someone goes
with knowing who that person is. But, Quine argues, this ‘is utterly
dependent on context. Sometimes … we see the face and want the name,
sometimes the reverse. Sometimes we want to know his role in the
community. Of itself the notion is empty’ (1981: 121). So Quine’s final
position is that we should ‘omit propositional attitudes de re from our
overall scientific language couched in the extensional grammar of predicate
logic’ (1995: 97).
This might seem a great loss, but Quine says, ‘it grows on one. I now
think the distinction [between de dicto and de re ascriptions of propositional
attitudes] every bit as empty, apart from context, as … that of knowing who
someone is. In context it can be important’ (1981: 121).
Quine’s view is thus not that no distinction can be drawn between those
de dicto propositional attitude ascriptions which are relevant to a de re
propositional attitude ascription and those which are not, but that we cannot
make the demarcation once and for all, outside the conversational context.
De re propositional attitude ascriptions have no place in a context-
independent scientific language, but we can still sensibly make them in
everyday life. However, in doing so the truth of our ascriptions may depend
on our purposes and the context of our utterance.
Let us now return to the difference between Quine’s final position on de
re propositional attitude ascriptions and his position on de re modal claims.
Quine is prepared to allow that in a specific conversational context a de re
propositional attitude ascription makes sense. He similarly thinks that
ascriptions of essential properties can make sense in context. He continues
the passage quoted above by making the comparison explicitly: ‘It [i.e. the
notion of who someone is] and the notion of essence are on a par. …
Relative to a particular enquiry some predicates … may be treated as
essential’ (1981: 121). Elsewhere he writes:
the requirement that distinguishes de re from de dicto, namely knowing who or what, is a function of
the contextual situation and not a general distinction. This classifies it with the indexicals; also with
necessity and possibility, according to my view of them. The indexicals and necessity and possibility
are convenient in daily discourse … ; and we can say the same of propositional attitudes de re …
while admitting none of these idioms to absolute or nonindexical scientific discourse.
(Leonardi and Santambrogio 1995: 358)
The first target of Kripke’s attack in Naming and Necessity is the Frege–
Russell description theory of names, conceived as a Searlean cluster theory.
Kripke gives the theory a very careful, six-thesis, formulation. As he
formulates the theory it is indexical and idiolectic. This makes his claim to
have refuted it very ambitious indeed. If he is right, the principle that
reference requires an identifying description, however weakly formulated,
is refuted and externalism, as Putnam defines it, established. Nevertheless,
he is right to formulate it in this way since, as the previous discussion
makes plain, only an indexical, idiolectic theory would have been endorsed
by either Frege or Russell, or, of course, Searle or Strawson.
So Kripke’s opponent is no straw man. He breaks down the cluster theory
into the following six theses (the speaker is A) (1980: 64–5):
This thesis, Kripke says, is just a definition. The substantive theses are the
ones that follow:
This does not say that the properties pick out someone uniquely. It just says
that A thinks that they do. He thinks he knows an answer to the question
‘Which thing (who) is X?’.
(3) If most, or a weighted most, of the j’s are satisfied by one unique
object, y, then y is the referent of ‘X’.
This thesis captures the Frege–Russell intuition that the link between
language and the world must be made by the speaker’s beliefs and
intentions. The notion of a ‘weighting’ is to capture the cluster theorist’s
intuition that some properties are more important than others in the ‘vote’.
Who decides on the weighting? A. This does not mean that we have to
imagine that A is capable of saying explicitly which properties he regards
as constitutive of the meaning of the name. The weighting will be manifest
in his reaction to circumstances in which some but not all the properties in
the set are possessed by one individual. The reference to a ‘weighting’ is a
reference to A’s dispositions: his dispositions to respond to the presentation
in various circumstances of information relevant, as he sees it, to the
existence of X. We can elicit these dispositions by asking him to suppose
that these circumstances obtain, or by creating a situation in which he
actually believes that they do. We will in general be limited to doing the
former (it is by appeal to such suppositions that Kripke argues against the
Frege–Russell theory using his famous Einstein, Gödel–Schmidt and
Feynman examples).
Kripke adds three theses to complete the theory:
(4) If the vote yields no unique object, ‘X’ does not refer.
(5) The statement ‘If X exists, then X has most of the j’s’ is known a
priori by the speaker.
(6) The statement ‘If X exists, X has most of the j’s’ expresses a
necessary truth (in the idiolect of the speaker).
Kripke adds two subsidiary theses (1980: 66). Subsidiary to thesis (5) is the
thesis that it is a priori true for the speaker that if nothing has most of the j’s
‘X’ does not refer. Subsidiary to thesis (6) is the thesis that ‘If nothing has
most of the js, X does not exist’ expresses a necessary truth in the idiolect
of the speaker.
Theses (5) and (6) also have converses (Kripke 1980: 73). The converse
of thesis (5) is that it is known a priori by A that if any unique thing has the
j’s it is X. The converse of thesis (6) is that ‘if anything has most of the
j’sitisX’ is a necessary truth in the idiolect of the speaker.
This is the theory Kripke aims to refute in Naming and Necessity. But
before considering his criticisms we need to note a constraint he imposes
without which it would be circular. This is important because one of the
main responses to Kripke is that the constraint against circularity required
by description theories is much weaker than he suggests, and the
description theorist is not as restricted as Kripke thinks in the descriptions
to which he can appeal.
Kripke motivates the introduction of his non-circularity constraint by
noting what he calls an ‘incidental defect’ in the way the defenders of the
cluster theory often state it. He takes the ‘Moses’ discussion from
Wittgenstein to illustrate the point. Wittgenstein writes: ‘when one says
“Moses did not exist” one may mean: the Israelites did not have a single
leader when they withdrew from Egypt. Or: their leader was not called
“Moses”. Or: there cannot have been someone who accomplished all that
the Bible relates of Moses’. The gist of Wittgenstein’s discussion is that we
know a priori that if all the Bible relates of Moses is false, Moses did not
exist. But, Kripke points out, there is a distinction between a complete
fiction and a substantially false account of a real person. Such an account
may come to be associated with a name. Legends grow up about its bearer
and the less interesting truth gets distorted or forgotten. Eventually the tales
passed down through the generations contain a substantially false account.
But they are still tales about the real person, and not about a fictional
character of which they are true by definition (contrast here, perhaps, our
talk of King Arthur and our talk of Merlin; it is plausible that the second
character is wholly fictional). This distinction as such is no problem for the
description theory. But it indicates that an adequate account must appeal to
a Strawsonian deferential use of names. A natural response is to say that in
our talk of Moses we intend our reference to be to that person to whom the
biblical story relates, even if it gives a wholly false account of him.
It is at this point that Kripke imposes his circularity constraint.
Deferential reference is allowed, but circularity is not. We can use ‘Moses’
deferentially to the authors of the Bible only because we know that they are
not deferring to us. If they were, it is obvious that no reference would be
secured by the deferential intention. Hence Kripke imposes constraint (C)
on the properties in the cluster of properties (1980: 68):
(C) For any successful theory the account must not be circular. The properties which are used in the
vote must not themselves involve the notion of reference in a way that is ultimately impossible to
eliminate.
This is very reasonable. Suppose the only thing A can say when asked who
X is is: the person B calls ‘X’. Then if the cluster of descriptions theory is
correct, whether A refers at all when he uses the name depends on whether
B does. But, likewise, if all B can say in answer to the same question is: the
person A calls ‘X’,heis in the same boat. Neither is referring to anyone if
the cluster theory is correct.
Constraint (C) is reasonable because it is very weak. It does not exclude
the possibility that one of the properties in the reference-determining set is
one in whose specification the concept of reference must be used. Whether
such a property can play a reference-determining role depends on empirical
facts. I can con-fidently ‘pass the buck’ to the authors of the Bible because I
know they did not pass the buck to me. ‘Whatever Jones means by “X”’ can
perfectly well serve as a reference-fixing description for me if Jones does
not pass the buck back to me. The only way of specifying a reference which
constraint (C) rules out a priori is ‘the thing I here and now mean by “X”’ –
where ‘X’ is the very name whose reference I am specifying. It does not
even rule out someone referring to someone by an utterance of a name
when his only identifying description is one that makes reference to that
very utterance. The crucial question about Kripke’s attack on the
description theory in Lecture II is whether, in the cases he considers
counterexamples to it, there are any reference-fixing clusters satisfying
constraint (C), as he himself formulates it, available to the user of the name.
Giving the Meaning, Fixing the Reference and
Rigid Designation
Before getting further into Kripke’s arguments we first need to note the
crucial distinction between thesis (6) and the others. Thesis (6) is the thesis
in which modality is introduced (necessity is not mentioned in any of the
preceding theses). It must belong to the cluster theory if that is considered
as a theory of meaning. The other theses belong to it even considered, less
ambitiously, merely as a theory of reference-fixing. Interpreting the cluster
theory as a theory of meaning, it entails that any name is fully equivalent to
a description and hence interchangeable with it salva veritate in all contexts
(except where the expressions are not used but only mentioned). The cluster
theory, considered merely as a theory of reference-fixing, does not entail
this. In particular, the cluster theorist qua theorist of reference-fixation need
not require the full substitutability salva veritate of names and descriptions
in modal contexts. Kripke’s criticism of the cluster theory in Lecture I is
directed only at the cluster theory as a theory of meaning. So his target is
solely thesis (6).
Much of Lecture I is given over to explaining the difference between the
cluster theory as a theory of meaning, i.e. including thesis (6), and as a mere
theory of reference-fixing, i.e. minus thesis (6). Making the distinction also
leads Kripke into a discussion of the concept of a possible world, in terms
of which he first explains the notion of rigidity.
An expression is a rigid designator if it designates the same object with
respect to every possible world in which it exists. The fact that language
users in some other possible worlds use ‘Nixon’ to refer to someone else
does not show that ‘Nixon’ as we use it in the actual world is not a rigid
designator (it is a trivial fact that any word could have been used to mean
something different). Nor, of course, as Kripke is at pains to point out in the
Preface to Naming and Necessity (1980: 7ff.), does the fact that many actual
people are called ‘Nixon’. It is only of a particular use of the name, e.g. the
conventional use of the name as a name for a disgraced President, that it
makes sense to ask whether it is a rigid designator. Just so, there is no sense
in asking whether the description ‘the odd one’ is rigid unless one specifies
that the context is one in which one is talking of numbers and ‘odd’ means
‘not numerically even’ or one in which one is talking of philosophers and
‘odd’ means ‘eccentric’ (see Fitch 2004: 44).
But to see what the notion of rigid designation actually amounts to we
need to get rid of what Kripke calls the ‘misleading metaphor’ of identity
across possible worlds (1980: 47–8).
Consider Kripke’s example of Benjamin Franklin, who was the first
Postmaster General of the United States and the inventor of bifocals.
The sentences:
(1) The inventor of bifocals might not have been the inventor of
bifocals.
and:
(2) Someone other than the inventor of bifocals might have been the
inventor of bifocals.
(1*) Concerning the man who in fact invented bifocals: he might not
have been the inventor of bifocals.
or:
(1**) The following might have been the case: the inventor of bifocals
was not the inventor of bifocals.
(2**) The following might have been the case: someone other than the
inventor of bifocals was the inventor of bifocals.
can only be heard one way, as saying of the man Benjamin Franklin that he
might not have been the man Benjamin Franklin. And this is absurd.
Benjamin Franklin might not have been called ‘Benjamin Franklin’, he
might have had a wholly different career and done none of the things we in
fact know him by, but he would still have been Benjamin Franklin.
Similarly:
(4) Someone other than Benjamin Franklin might have been Benjamin
Franklin.
can only be heard as saying of the man, Benjamin Franklin, that someone
other than he might have been the man Benjamin Franklin, and this is
absurd. For although someone other than Franklin might have been called
‘Benjamin Franklin’ and might have done the things which made Franklin
famous, no one other than Franklin could have been Franklin.
That (3) and (4), unlike (1) and (2), are not ambiguous, Kripke says,
shows that the name ‘Benjamin Franklin’ is a rigid designator, whose
designation in any possible world is its actual designation, whereas ‘the
inventor of bifocals’ is a non-rigid designator, whose designation in any
possible world is whoever, in that world, satisfies the condition of being the
inventor of bifocals.
We thus see the rationale for the two intuitive tests for rigid designation
given in Chapter 1: that there are negative answers to the two questions (1)
Does ‘X might not have been X’ have a true reading? (2) Does ‘someone
other than X might have been X’ have a true reading? An account which
accords better with Kripke’s intentions than the admittedly misleading
explanation in terms of identity across worlds is the following: a singular
term ‘X’ is used as a rigid designator if it is so used that it refers, when it
figures inside counterfactual or other modal linguistic contexts, to what it
refers to, if at all, outside such contexts (this formulation is given by Hale
2004: 367). This makes no commitment to possible worlds and a fortiori
makes no use of any notion of identity across possible worlds. It explicitly
makes the rigidity of a proper name a matter of our intentions in using it,
but this is surely what Kripke thinks. Note that both tests are required
because ‘my father’ satisfies the second test, but it is not a rigid designator,
it does not designate my father with respect to every possible world in
which he exists. A stronger notion is that of an obstinately rigid designator
(Salmon 2005: 34): ‘X’ is an obstinately rigid designator if it designates the
same thing with respect to every possible world, regardless of whether that
thing exists in that world. So ‘Benjamin Franklin’ is an obstinately rigid
designator, since ‘It might have been that Benjamin Franklin failed to exist’
is unambiguously true, but ‘the man identical to Benjamin Franklin’ is not,
since ‘it might have been that the man identical to Benjamin Franklin failed
to exist’ is not unambiguously true.
An important distinction to note (not made in these terms by Kripke) is
between constant and inconstant rigid designators (Mackie 2002). An
example of a constant rigid designator is ‘the square of three’. This is what
Kripke calls a strongly rigid designator, since it designates nine, a necessary
existent. But in addition, given what it means, its conventional dictionary
significance, it could not have designated anything else (of course, it could
have meant something else and so been used by language users in some
other possible world to designate something else). Contrast the description
‘the actual number of planets’. This too is a strongly rigid designator by the
intuitive tests. But holding its conventional significance constant it could
have designated some other number (it would have done if there had been
only six planets). Hence although the necessitation of ‘the actual number of
planets is the number of planets’ (i.e. ‘it is a necessary truth that the actual
number of planets is the number of planets’) is false, ‘the actual number of
planets is the number of planets’ could not have been false, given its actual
conventional significance.
Kripke notes this distinction (1980: 59–60, n. 22) but says nothing
explicitly against the suggestion that names abbreviate such ‘rigidified’
descriptions, at one place (1980: 6, n. 8) remarking that his principal thesis
‘contrasts names with non-rigid descriptions as advocated by Russell’. But
it is highly relevant to his modal argument against the description theory, as
many have noted at least from Searle (1983) on, since the description
theorist can maintain his position, and, in particular, thesis (6), even whilst
acknowledging that proper names are rigid, if he says that such names are
inconstant rigid designators like ‘the actual number of planets’.
A distinction Kripke does emphasize, not in the original text of Naming
and Necessity, but in the Preface to the second edition, is one between de
facto and de jure rigidity.A de jure rigid designator is one whose reference
is stipulated to be a single object whether we are speaking of the actual or a
counterfactual situation, whilst a de facto rigid designator is a description
which happens to be constructed using a predicate true of one and the same
object in each possible world. A description like ‘the square of three’ or
‘the smallest prime’ is a de facto rigid designator. Kripke maintains that
names are de jure rigid designators. Given that they are rigid designators it
is trivial that this is so if it follows from the fact that they are simple
symbols and so have to have their meanings stipulated. But this is
consistent with a name’s being a mere abbreviation of an actualized
description like ‘the actual number of planets’ and hence an inconstant rigid
designator. How to classify such inconstant rigidly designating descriptions
with respect to the de jure/de facto distinction is not discussed in Naming
and Necessity. Kripke does say, however, that although he in fact believes
that names are de jure rigid, he is content with the weaker assertion of
rigidity.
We can now turn to Kripke’s account of the distinction between the
cluster theory qua theory of reference-fixing and qua theory of meaning. As
illustration he uses the example of the standard metre rod, taken from
Wittgenstein.
The puzzle he starts from is that it might seem that if ‘one metre’ is
defined to be the length of a stick S at a certain time t it must follow that it
is a necessary truth that S is one metre long at t. But, Kripke argues, this
consequence does not follow if the definition is understood merely as
giving the reference rather than giving the meaning. Someone who uses the
definition to give the reference can still say correctly ‘If heat had been
applied to stick S at t then at t stick S would not have been one metre long’.
The reason is that ‘the length of stick S at t’ is a non-rigid designator,
whereas ‘one metre’ is a rigid designator. A merely reference-fixing
definition of a name (‘one metre’ is a name of an abstract object, a length)
is then one in which the intention is to introduce a rigid designator via a
non-rigid designator. Whether it is a constant or inconstant rigid designator
is something Kripke does not discuss. Prima facie there are two possibilities
here. The introducer of the term could intend to introduce a rigid designator
which is synonymous with ‘the actual length of stick S at t’ and so
inconstant, or he could intend to introduce a constant rigid designator
whose meaning requires it to have a particular length as its designation. So
there is the question: is the meaning of ‘one metre’ in the idiolect of the
original introducer of the term such that the statement ‘one metre is the
length of stick S at t’, meaning what it actually means, has to be true? In
fact, although the notion of a constant rigid designator is unproblematic, as
shown by the example of ‘the square of three’,itis problematic to suppose
that by fixing its reference by a non-rigid designator one can introduce a
name by stipulation as a constant, rather than inconstant, rigid designator,
since to do so would be to stipulate that one’s very stipulation would have
determined a different meaning if things had been different. What would
show, so long as the name was used only by the original introducer, that it
was a constant rather than an inconstant rigid designator?
What Kripke does argue is that it is not a necessary truth that stick S is
one metre long at t since under other possible circumstances S would not
have been one metre long at t. However, he argues that for the introducer of
the term ‘one metre’, ‘S is one metre long’ is something he knows a priori
(just as, for someone who stipulates that the newly coined term
‘squatchelor’ is to apply to all and only unmarried men, it is a priori that
squatchelors are unmarried men). In this sense ‘S is one metre long’ is a
contingent a priori truth, for it is a contingent truth that S is one metre long,
but the introducer of the term knows a priori that S is one metre long.
Thus, given the distinction between rigid and non-rigid designation and
the possibility of a reference-fixing definition in which a rigid designator
has its reference fixed by a non-rigid description, the possibility in this
sense of contingent a priori truths appears to follow.
Kripke goes on to point out (1980: 57) that the distinction between
reference-fixing and synonym-giving definitions can be applied generally.
There is a difference between using the stipulation ‘Aristotle is the greatest
man who studied under Plato’ as a synonym-giving definition of the name
‘Aristotle’ and using it merely to fix the reference of the name. If we do the
latter once again there will apparently be a contingent a priori truth
forthcoming since for the introducer of the name it will be something he
knows a priori that, if he existed, Aristotle studied under Plato; but it is not
a necessary truth that Aristotle, if he existed, studied under Plato.
Kripke’s picture of naming, in fact, is that in every case a name is
introduced by a procedure that can be thought of as reference-fixing – either
by description or, what can be thought of as a special case of the former,
ostension – and then the name is passed on from speaker to speaker without
the original reference-fixing description necessarily being passed on.
There is a final point to be made now, crucial to the assessment of
Kripke’s modal argument. If a name (‘Benjamin Franklin’) is synonymous
with a description (‘the inventor of bifocals’) thesis (6) must be true
(‘Benjamin Franklin, if he existed, was the inventor of bifocals’, must
express a necessary truth). But the converse does not follow. Thesis (6) can
be correct even if the name and the associated description differ in meaning
so long as this difference is consistent with the necessary truth in the
idiolect of the speaker of the identity statement ‘X is the j’. But
thoroughgoing synonymy is not needed for the necessary truth of this
identity statement. So defenders of Kripke’s target thesis (6) can still say
that names are not synonymous with the descriptions they are associated
with, and some, as we shall see, do.
The Modal Argument
Kripke states his modal argument against the description theory at a number
of places (1980: 53, 57, 61 and 74). Its full statement in the original text of
Naming and Necessity on page 61 is as follows:
The [description theorist] would say that the name is simply defined synonymously, as the cluster of
descriptions. It will then be necessary, not that Moses had any particular property in this cluster, but
that he had the disjunction of them. There couldn’t be any counterfactual situation in which he didn’t
do any of these things. … Such a suggestion … must clearly be false. … Most of the things
commonly attributed to Aristotle are things Aristotle might not have done at all.
In a nutshell the intended argument is this. All proper names are rigid
designators. Any descriptions (at least of contingently existent objects like
Aristotle) which can be at all plausibly thought of as fixing their references,
given what the users of the names believe and intend (in the case of
Aristotle, descriptions referring to achievements commonly associated with
him), are accidental designators. So such descriptions are not substitutable
salva veritate for proper names in modal contexts and hence a fortiori are
never synonymous with them.
To assess this argument we need to see how exactly it is supposed to
refute thesis (6).
Let us go back to the example of Benjamin Franklin and make the
supposition that the only description which is a candidate for synonymy
with ‘Benjamin Franklin’ is ‘the inventor of bifocals’. But as we have seen,
‘Benjamin Franklin’ is a rigid designator, and so it makes no sense to say
that Benjamin Franklin might not have been Benjamin Franklin or that
someone other than Benjamin Franklin might have been Benjamin Franklin.
On the other hand, ‘the inventor of bifocals might not have been the
inventor of bifocals’ has a true reading as well as a false one, as does
‘someone other than the inventor of bifocals might have been the inventor
of bifocals’. This shows that ‘the inventor of bifocals’ is a non-rigid or
accidental designator, and differs in meaning from ‘Benjamin Franklin’.
Moreover, ‘Benjamin Franklin might have existed and not invented
bifocals’ is unambiguously true. So the modal statement ‘it is necessarily
true that if Benjamin Franklin existed he invented bifocals’ is
unambiguously false. The next step in the argument is to the conclusion that
the non-modal statement ‘if Benjamin Franklin existed, then Benjamin
Franklin invented bifocals’ does not express a necessary truth (in the
idiolect of the speaker who fixes the reference of the name by the
description). Kripke concludes that thesis (6) is false since the same
reasoning can be gone through for any name and any description put
forward as the description which gives its meaning, including a description
equivalent to a cluster of descriptions. Of course, the description theorist
would not in fact offer the description ‘the inventor of bifocals’ as fixing
the reference of ‘Benjamin Franklin’. So the supposition Kripke is making
is unrealistic. But other examples make it easier to focus on the crucial
point. Kripke himself gives the example of ‘Jack the Ripper’. Dummett
(1973: 112) suggests ‘St Anne’ as a name which even educated users
associate only with the description ‘the mother of the Virgin Mary’ and
following Gareth Evans (1979) philosophers have come to use ‘Julius’ as
the name of the inventor of the zip, whoever he was.
The argument as stated is what appears to be the most obvious
interpretation of Kripke’s text, but in the Preface to the 1980 book he
refines it, in reaction to Michael Dummett’s statement (in his 1973) of the
first response we will look at below (that rigidity may be explained by
regarding proper names as synonymous with widescope descriptions).
Kripke’s Preface refinement will be explained after Dummett’s original
widescopist response has been set out, and we will then be in a position to
evaluate in turn the widescopist response to this which Dummett gives in
his subsequent book (1981). Kripke himself does not further respond, but
the debate has continued and we will look at one of the most prominent
recent attacks on widescopism.
The Debate Over the Modal Argument
What the defender of thesis (6) must do is to show how it can be true even
though ordinary proper names are typically used as rigid designators. He
cannot deny that ordinary proper names typically do function as rigid
designators so the only matter for argument can be the explanation of this
fact. He has at least four options. The first is the widescopist response,
originally put forward by Dummett (1973: 127ff.), though with
qualifications, together with the supplementary response, involving an
appeal to the distinction between ‘assertoric content’ and ‘ingredient sense’,
presented by Dummett (1981) in his reply to Kripke’s Preface refinement.
The second way of responding to the modal argument, within the
framework of Frege’s philosophy, is to appeal to the idea that proper names
differ from descriptions by maintaining their customary reference in modal
and counterfactual contexts, and do not refer therein to their ordinary
senses. This is a suggestion originally put forward by Burge (1979). A third
way of responding to Kripke’s modal argument is to say that the
descriptions with which names are synonymous are inconstant rigidified
descriptions, like ‘the actual inventor of bifocals’.As we noted, one of the
earliest defenders of this position was Searle (1983), but it has been
defended by many others including Jackson (1998). Finally, a fourth
response available to the descriptivist is to appeal to the distinction between
indicative descriptions and subjunctive descriptions, the distinction between
‘the person who is richer than anyone else’ and ‘the person who would have
been richer than anyone else’. All these responses are ways of defending
thesis (6) from Kripke’s objections by blocking the transition from the
modal claim, intuitively obviously correct in the cases he gives, that if
things had been different X would not have been j, to the conclusion that
the non-modal statement ‘If X exists, X is j’ does not express a necessary
truth in the idiolect of the speaker.
The fundamental idea of widescopism (which has also been endorsed by
Loar 1976, Schiffer 1977 and many others) is an appeal to a scope
convention. It cannot be denied that ordinary proper names are typically
used as rigid designators, but a possible explanation of this, Dummett
suggests (1973: 127ff.), is that names are in a sense abbreviations of
descriptions, but are used subject to the general convention that in modal
contexts they are to be read as having maximal scope relative to the modal
operators. Certainly, one can introduce an expression which is stipulated to
be synonymous with another except that, unlike the latter, it is to have a
wide scope relative to a certain class of operator. As McCulloch (1989: 110)
points out, one might introduce an expression ‘*’ by the stipulation that it is
to be read as synonymous with the multiplication symbol ‘×’ except that it
is to be read as having wide scope relative to other arithmetical operators.
Thus ‘2 3’ and ‘2* 3’ will identify the same arithmetical task, but ‘4+ 2
3=18’ will be ambiguous (in the absence of a convention specifying the
order in which the calculations are to be carried out), whilst ‘4 + 2 * 3 = 18’
will unambiguously have the same truth-value as ‘(4 + 2) 3= 18’. In a sense
then ‘*’ means the same as ‘×’ but is subject to a convention not governing
the latter, in another sense ‘*’ does not mean the same as ‘×’ since to
understand it requires knowledge of a convention knowledge of which is
not required for understanding of ‘×’. Does ‘2* 3’ means the same as ‘2 3’?
They identify the same arithmetical task. It is unimportant whether one says
it therefore has the same meaning (since here the difference between ‘×’
and ‘*’ is idle) or has a different meaning (since understanding one of its
constituents requires knowledge of a convention not required for
understanding any constituent of the other).
Similarly, it might be, for all that Kripke’s unrefined modal argument
establishes, that ‘Benjamin Franklin’ does mean the same as ‘the inventor
of bifocals’ except that it is subject to the convention that it must be read as
having maximal scope in modal contexts. If so ‘Benjamin Franklin might
not have invented bifocals’ is unambiguously true, but ‘the inventor of
bifocals might not have invented bifocals’ is not. Nevertheless, consistently
with thesis (6), the sentence ‘Benjamin Franklin, if he existed, was the
inventor of bifocals’ expresses a necessary truth. On this proposal the name
and the description do not mean the same inasmuch as understanding the
former requires knowledge that it is governed by a convention that does not
govern the latter. Do ‘Benjamin Franklin invented bifocals’ and ‘the
inventor of bifocals invented bifocals’ then mean the same? There is an
argument for saying so, since the only difference is that only one of them
contains an expression governed by the widescope convention, but that
convention is idle in the context. But there is an argument for saying that
they differ in meaning since there is a sense in which their constituents
‘Benjamin Franklin’ and ‘the inventor of bifocals’ differ in meaning.
Although Dummett is the originator of the widescopist response his own
position in his first discussion of Kripke is nuanced. He does not in fact
accept Kripke’s claim that all proper names are rigid designators, and
therefore, of course, only suggests that the widescope convention governs
those that are. His own example of ‘St Anne’ is a name that he argues is not
a rigid designator. It is no good, he asserts, arguing that ‘St Anne’ cannot
mean the same as ‘the mother of the Virgin Mary’ because St Anne might
never have become a parent. ‘The mother of the Virgin Mary might never
have become a parent’ is ambiguous, as Kripke insists, between a true
reading and a false one, as is ‘The mother of the Virgin Mary could not but
have been a parent’. But ‘St Anne could not but have been a parent’ is
equally ambiguous. ‘Even though there is an intuitive sense in which it is
correct to say “St Anne might never have become a parent”, there is an
equally clear sense in which we may rightly say, “St Anne cannot but have
been a parent” provided always that this is understood as meaning that if
there was such a woman as St Anne, then she can only have been a parent’
(Dummett 1973: 113).
Kripke acknowledges that there is an ambiguity in ‘St Anne cannot but
have been a parent’, Dummett points out, in the sense that he acknowledges
that if the reference of ‘St Anne’ is fixed by the description, ‘the mother of
the Virgin Mary’, the first users of the name know a priori that St Anne, if
she existed, was a parent. In this sense we can say that this statement is
epistemically necessary. But Kripke would insist that it is not
metaphysically necessary that St Anne was a parent; like every mother, she
could have remained childless. Hence the ambiguity is in the modal
operator. The situation is quite different, Kripke thinks, with modal
statements containing descriptions, like ‘the mother of Mary cannot but
have been a parent’. The ambiguity here is a scope ambiguity, which does
not have to be explained by appeal to two notions of necessity. Of course,
Kripke has to make this distinction between the cases if he is to maintain
that ‘St Anne’, unlike ‘the mother of Mary’, is a rigid designator. So
Dummett’s claim that the same ambiguity is present in both cases is a
challenge to a fundamental Kripkean claim.
That proper names are rigid designators seems obvious when we consider
examples like ‘Benjamin Franklin’ or ‘Aristotle’,or other proper names of
well-known historical figures. And so Dummett’s challenge might seem
heroic. However, as he emphasizes, the Kripkean modal argument is
supposed to show that even in the case of a proper name whose users all
associate with it just a single non-rigid description, and which is not used
deferentially by anyone, the name and the description are not synonymous,
since the name, unlike the description, is rigid. That is why Dummett
selects the name ‘St Anne’ as a name for which this assumption is plausibly
true. Another example, from Kripke himself, of which the same is true, is
plausibly ‘Jack the Ripper’, which we may suppose to have its reference
fixed by some such description as ‘the notorious nineteenth-century London
murderer’, which is known to everyone who uses it. Now Kripke’s position
has got to be that whilst ‘the notorious nineteenth-century London murderer
might not have been a murderer’ has two readings because of the scope
ambiguity, this is not so in the case of ‘Jack the Ripper might not have been
a murderer’. In this case the ambiguity, if one is allowed, must be due to a
difference between epistemic and metaphysical modality. But, Dummett
argues, to appeal to different explanations of the ambiguity in the two cases
is implausible. There appears to be the same phenomenon, which therefore,
other things being equal, should be explained in the same way. And this
point is reinforced, he argues, when we look more carefully at the
distinction between epistemic and metaphysical modality, to which Kripke
has to appeal to explain the ambiguity in ‘Jack the Ripper cannot but have
been a murderer’ or ‘St Anne cannot but have been a parent’. The sense in
which these are true, he says, is the sense in which it is true that the
standard metre rod could not have failed to be a metre long. This is an
epistemic sense. Our knowledge that Jack the Ripper cannot but have been
a murderer is knowledge derived from the way the words are used, i.e. from
the fact that ‘the notorious nineteenth-century London murderer’ is used to
fix the reference of ‘Jack the Ripper’. By contrast, the kind of necessity we
are concerned to deny when we say that ‘Jack the Ripper might never have
been a murderer’– for example, if he had died as infant – does not depend
on our use of words, it is a metaphysical necessity. But it appears that we
can draw exactly the same contrast between the two readings of ‘the
notorious nineteenth-century London murderer cannot but have been a
murderer’. When we say that the murderer cannot but have been a
murderer, in the sense in which it is true to say this, we are expressing a
priori knowledge based solely on our understanding of the words, precisely
similar to that expressed when we say that Jack the Ripper cannot but have
been a murderer, or that the standard metre cannot but have been one metre
long. And when we say that the murderer might not have been a murderer
we are concerned with the very same kind of metaphysical necessity
involved when we say that Jack the Ripper might not have been a murderer.
But if the ambiguity in this case is to be explained in terms of scope
ambiguity we cannot regard the modal operators as ambiguous between
epistemic and metaphysical readings.
Dummett’s suggestion that in cases like that of ‘St Anne’, where a single
description is associated with the name by every user and there is no
deferential use, the same ambiguity can be discerned in modal sentences
containing the name as in modal sentences containing the description, does
not require him to say that this is also the case for more typical names, like
‘Benjamin Franklin’ or ‘Aristotle’. His widescopist thesis is that there is a
convention whereby proper names are to be read as having wide-scope in
modal contexts. But if it is only by convention that something is not done
then it is possible to do it. So where the rationale for the convention fails to
apply, the convention will sometimes be flouted. In the case of unusual
names like ‘St Anne’ and ‘Jack the Ripper’, associated with the same single
description by all who use it, it is reasonable to suppose that whatever the
rationale for the convention that names are to be read as having wide scope
relative to modal operators, it does not apply. That is why the narrow scope
reading comes into view as a possibility. And, indeed, a reading of ‘Jack the
Ripper might not have been Jack the Ripper’ as saying something true (on
which reading the first occurrence of the name must have wide scope and
the second narrow) does seem to be available.
This account is consistent with the view Donnellan (1977) puts forward
in his response to Dummett in his paper ‘The Contingent A Priori and Rigid
Designators’. Donnellan explains that the uses of proper names Dummett is
discussing, like Kripke’s example of the use of ‘Neptune’ by Leverrier
subsequent to his introduction of it via the description ‘the cause of the
perturbations in the orbit of Uranus’, are quite unusual ones, in which a
name is pegged during the relevant time to the definite description via
which it was introduced. But in the case of such uses, he suggests, it is
indeterminate whether the name is a rigid designator or an abbreviation for
a description, for nothing in the actual use of the name or its original user’s
intentions can show which it is. The two proposals entail different
consequences. For example, the first entails that ‘If Neptune exists, Neptune
is the cause of the perturbations in the orbit of Uranus’ is a contingent truth,
the second that it is a necessary truth. But all the intuitions we actually have
about the case can be explained on either proposal, so the differences are
not ones that can be used to decide between them. Historically names have
not been introduced by an explicit stipulation that they are to be taken as
rigid designators and so long as a name not introduced by such an explicit
formula, like ‘Neptune’, ‘St Anne’ or ‘Jack the Ripper’, continues to be
associated with the original description by which its reference is fixed,
Donnellan argues, there is no fact of the matter about whether it is a rigid
designator or an abbreviation for a description – entirely unworryingly
since though, for example, it will be indeterminate whether ‘Neptune is a
large planet’ can be paraphrased as ‘the cause of the perturbations in the
orbit of Uranus is a large planet’, anyone who knows how the name was
introduced will know the sentences have the same truth-value.
The description of such situations as ones in which it is simply
indeterminate whether the name is a rigid designator is entirely in accord
with Dummett’s thought. For, as noted, these are plausibly cases in which
the rationale for the general widescopist convention fails to apply, so the
possibility of reading, e.g. ‘It might have been that Neptune existed but was
not the cause of the perturbations in the orbit of Uranus’, as saying
something plainly false (by giving the description the name abbreviates
narrow scope), though excluded by the convention, becomes evident. But,
given the general convention, to which users of the name are subject, it is
excluded. So a user of the name, aware of its peculiar features, yet party to
the general convention, might well feel undecided in the way Donnellan
describes.
Of course, this story depends on a rationale being available for the
widescope convention. But we shall see below that a plausible suggestion
(from Lewis 1984) fits well with the thought that names like ‘St Anne’ and
‘Jack the Ripper’ are ones to which it would be pointless to apply the
widescope convention.
In his Preface Kripke has in fact two objections to wide-scopism. The
first objection is that the proposed convention is ‘unaccountable’ (1980:
13), i.e. has no rationale. As just noted, Lewis’s proposed rationale will be
given shortly. The second objection (1980: 6ff.), which constitutes what I
have called Kripke’s Preface refinement of the modal argument, is that the
distinction between between rigid and non-rigid designators can be made
even when we consider simple sentences containing no modal operators. He
considers the two sentences ‘Aristotle was fond of dogs’ and ‘the last great
philosopher of antiquity was fond of dogs’. These are non-modal sentences.
Nevertheless he argues, they differ in their truth-conditions as they describe
counterfactual situations. There is a certain man of whom it is true that
‘Aristotle was fond of dogs’ truly describes a counterfactual situation if and
only if he would have been fond of dogs if that situation had obtained. The
truth-condition of ‘the last great philosopher of antiquity was fond of dogs’,
as it describes a counterfactual situation, cannot be so stated. The latter is
true as describing a counter-factual situation just in case exactly one person
would have been last among the great philosophers of antiquity and would
have been fond of dogs if that situation had obtained. This (Russellian)
condition for the truth of ‘the last great philosopher of antiquity was fond of
dogs’ can differ wildly from the specified condition for the truth of
‘Aristotle was fond of dogs’. With respect to a counterfactual situation
where someone other than Aristotle would have been the last great
philosopher of antiquity identifying the counterfactual truth-condition for
‘Aristotle was fond of dogs’ with the (Russellian) counterfactual truth-
condition for ‘the last great philosopher of antiquity was fond of dogs’
would (incorrectly!) make that other person’s fondness for dogs the relevant
issuefor thecorrectness of ‘Aristotle was fond of dogs’.
Kripke’s second objection to the widescopist can be restated as follows
(cf. Kripke 1980: 13–14). Consider the simple sentences ‘Benjamin
Franklin was the inventor of bifocals’ and ‘Benjamin Franklin was
Benjamin Franklin’. The assertion:
Benjamin Franklin was Benjamin Franklin. That might not have been the case.
is true. There can be no scope issues here since in each case the first
sentence is a non-modal sentence. Yet the two assertions differ in the way
specified. Hence the two initial non-modal sentences differ in their modal
status, in that one expresses a necessary truth and the other a contingency.
So ‘Benjamin Franklin was the inventor of bifocals (if he existed)’ cannot
express a necessary truth. But since any proper name is rigid the same
argument can be run for any proper name and any non-rigid description a
description theorist could plausibly put forward as fixing its reference. So
once again, thesis (6) of the description theory is refuted.
In his second discussion of Kripke, Dummett (1981) responds to this
refinement of the modal argument. David Sosa (2001) responds similarly
and quotes in a note a comment by David Lewis which nicely summarizes
the underlying thought:
If a description is eager enough for wide scope, its scope may cross sentence boundaries and also
boundaries of quotation and disquotation. … it’s fair to treat the following alike; and in each case to
say that ‘Aristotle’ may take wide enough scope to cover all the rest. (a) With respect to
[counterfactual situation] S, Aristotle was fond of dogs. (b) This is the case with respect to S:
Aristotle was fond of dogs. (c) Consider the simple sentence ‘Aristotle was fond of dogs’. That
sentence is true with respect to S.
… wide-scopers [have been accused] of solving the problem of (a) while ignoring (c). Our reply is
that (c) functions as a mere stylistic variant of (a)
(Sosa 2001, 34–7, n. 7)
and:
The inventor of bifocals invented bifocals. That might not have been true.
unlike:
Benjamin Franklin was rich. That might not have been true.
to get:
It is raining where Mrs Thatcher is. That is not always true.
The first of these (if we suppose the assertion made in Manchester) is
unambiguously false. The second is ambiguous (supposed still to be made
in Manchester). On one reading its truth requires that it not always be
raining in Manchester. On the second reading it requires that it not always
be raining where Mrs Thatcher is then located. (There is a parallel
ambiguity in ‘The man standing next to Mrs Thatcher is carrying a gun.
That is always true’ or ‘Mary’s boyfriend is very rich. That used to be
true’.) We can account for the difference by appealing to the different
ingredient senses of ‘it is raining Thatcherabouts’ and ‘it is raining where
Mrs Thatcher is’ in the context ‘it is always the case that …’ and
hypothesizing that the two occurrences of ‘That’ refer to these ingredient
senses.
Again consider:
Thatcherabouts is Thatcherabouts. That is always true.
and:
Thatcherabouts is where Mrs Thatcher is. That is not always true.
But now consider an instance of this in which ‘N’ is a name the reference of
which is fixed by the contingently satisfied description ‘the F’:
The proposition that if N is G something is both F and G is the proposition that if the F is G
something is both F and G.
The proposition that if the F is G something is both F and G is a necessary truth.
Hence:
The proposition that if N is G something is both F and G is a necessary truth.
and
The proposition that it is raining Thatcherabouts is the proposition that it is raining where Mrs
Thatcher is.
The proposition that it is raining where Mrs Thatcher is is always true.
Hence:
The proposition that it is raining Thatcherabouts is always true.
(1) The inventor of bifocals might not have been the inventor of
bifocals
are similarly present (as noted by Quine) in ascriptions of propositional
attitudes such as:
and:
The relevant ambiguity in (1) can now be brought out by the two
readings:
(1*) It might have been the case that [the inventor of bifocals was not
the inventor of bifocals].
and:
In the case of both (1) and (2) some restriction on type of aspect is
required if the second reading is to capture the intuitive English meaning.
We have already seen what is at issue in our discussion of Quine on
exportation and the ‘shortest spy’. In the case of propositional attitude
ascriptions we do not invariably allow exportation of a singular term from
an opaque position to a transparent position. We do not allow the inference
from ‘Ralph believes that the shortest spy is a spy’ to ‘Ralph believes of the
shortest spy that he is a spy’ and thence to ‘There is someone Ralph
believes to be a spy’. Similarly, from the fact that George IV wondered
whether the author of Waverley was Scottish it is not immediately evident
that we should infer that there was someone of whom he wondered whether
he was Scottish.
The Fregean must say that the existential quantifier form (2**) captures
the intuitive English meaning with the description read as having wide
scope only if the quantification is restricted to what we might call ‘identity-
revealing’ aspects – where an aspect is identity-revealing only if knowing
that [α is X] suffices for knowing who X is.
An identity-revealing aspect may be what Kaplan (1969) calls a vivid
name. Alternatively, if Quine is right that ‘knowing who’ is a highly
context-dependent notion, the notion of an identity-revealing aspect will
also be highly context-dependent. But in most contexts of discussion of
George IV’s interest in the author of Waverley it will be correct to say that
he did not know who the author of Waverley was, although he knew that the
author of Waverley was the author of Waverley, and hence in most contexts
of discussion it would be incorrect to say that there was an identity-
revealing aspect under which George IV thought of the author of Waverley
as the author of Waverley.
The Fregean must proceed similarly to account for the intuitive English
meaning of (1), according to the reading on which the first occurrence of
the description is taken as lying outside the modal operator. So read the
statement says of the man, Benjamin Franklin, that he might never have
become the inventor of bifocals. To capture this meaning we have to regard
the quantification in (1**) as restricted to what we might call ‘essence-
revealing’ aspects. Suppose that Benjamin Franklin was essentially a
human being, so that ‘the inventor of bifocals might not have been human’,
on the intended wide scope reading, is false. Still, bifocals might have been
invented by extraterrestrials. So:
For some α, α is an aspect of the inventor of bifocals and it might have been that [α was not human].
is true if the quantification ranges over all aspects, since ‘the inventor of
bifocals’ expresses an aspect of Benjamin Franklin and it might have been
that the inventor of bifocals was not human.
So the Fregean needs the notion of an essence-revealing aspect if he is to
capture the intuitive thought that the inventor of bifocals had to be human,
but might not have been the inventor of bifocals, just as he needs the notion
of an identity-revealing aspect in order to capture the intuitive thought that,
in some contexts at least, knowing that the shortest spy is the shortest spy
does not count as knowing who the shortest spy is. An aspect α is an
essence-revealing aspect of an object x if and only if it is necessary that [if
anything is α it is F] for any ‘F’ which denotes a necessary property of x,
and it is not necessary that [if anything is α it is F] for any ‘F’ which
denotes an accidental property of x. An essence-revealing aspect of an
object is one that presents it as the possessor of all and only its essential
properties.
There are complications. One possibility is that even if objects have
essential properties, they do not have individual essence, but only general
essences. It may be that I am essentially human but have no other essential
property (so that I could have been anything any human being could be). To
deal with this possibility we need to extend the notion of an ‘aspect’ of an
object x. Say that aspects of x are the senses of predicates satisfied by x.
Aspects as originally defined, senses of Fregean proper names, can be
thought of as senses of predicates of the form ‘is identical with X’, where
‘X’ is the proper name; the aspect expressed by ‘Socrates’ can be thought of
as the sense of the predicate ‘is Socrates’. In the case of quantification into
propositional attitude contexts the aspects quantified over must be
understood as aspects as originally defined.
Of course, this notion of an essence-revealing aspect, like that of an
identity-revealing aspect, may have different applications in different
contexts of use. Maybe, as Quine thinks, we have no context-independent
notion of an essential property. Maybe in some contexts it is correct to say
that Benjamin Franklin might not have been a human being. And, of course,
I have not explained the – possibly highly context-dependent – difference
between essential and accidental properties by reference to the notion of an
essence-revealing aspect; I have just defined the latter in terms of the
former. Just so, the difference between knowing and not knowing who
someone is has not been explained in terms of the notion of an identity-
revealing aspect; rather the latter has been defined in terms of the former.
We know, in most contexts anyway, that knowing that the shortest spy is the
shortest spy does not count as knowing who the shortest spy is, and we
know that the fact that the inventor of bifocals might not have been a
human being does not establish that the actual inventor of bifocals might
not have been a human being. The Fregean must be able to accommodate
the differences between the two readings of ‘the inventor of bifocals might
not have been the inventor of bifocals’, just as he must be able to
accommodate the differences between the two readings of ‘Ralph believes
that the shortest spy is a spy’. The Russellian does so by appeal to scope
distinctions. The Fregean must do so by appeal to the possibility of
quantifying over senses.
Now a Fregean response to Kripke’s modal argument can be given, i.e. a
Fregean account of what the rigidity of proper names amounts to, entirely
parallel to the Dummettian ‘widescopist’ Russellian response. Namely,
although every name has on any occasion of use a sense which might well
be that of a description, there is a convention in force whereby a proper
name, as opposed to a description, must not be used in a modal context to
refer to its indirect reference, i.e. its customary sense, and, in general, must
not be used by a speaker to refer to the sense he associates with it in
contexts where replacement by a proper name with the same direct
reference but a different customary sense is not guaranteed to preserve
truth-value. This proposal, originally put forward by Burge (1979), is
further defended by Hale (2004). Burge expresses the point by saying that
proper names always maintain their customary reference in counterfactual
contexts. Hale similarly writes: ‘We could say: d is used as a rigid
designator if and only if it is used in such a way that it refers when it figures
inside counterfactual and other contexts to what it refers to, if to anything at
all, outside such contexts’ (2004: 367). (Of course this is not, in itself, a
distinctively Fregean account of rigidity; it becomes so only if the
possibility being excluded for rigid designators is viewed as reference,
inside counterfactual or modal contexts, to sense, which is Hale’s intention.)
With this convention in force, there is no reading of ‘Benjamin Franklin
might not have been Benjamin Franklin’ corresponding to the reading of
‘The inventor of bifocals might not have been the inventor of bifocals’ as ‘It
might have been the case that [the inventor of bifocals was not the inventor
of bifocals]’.
Obviously, as noted, this Fregean response to Kripke’s modal argument is
strictly parallel to Dummett’s widescopist response. In fact, it is Dummett’s
response in Fregean guise. (Dummett would agree, since on his view
quantification into modal contexts, like quantification into any other non-
extensional context, is to be understood as quantification over senses
(Dummett 1973: ch. 9).)
Hence its rationale is the Lewisean one already provided for the
widescopist. Since the senses of proper names are idiolectic we will lack a
common subject matter if we use them to refer to the senses we individually
associate with them. Burge puts this point by saying that the (near) rigidity
of names and other indexical devices is ‘the offspring of a marriage of
convenience between cognitively promiscuous linguistic devices and
contexts where Sinn [sense] does not matter’ (2005: 225).
If the convention has this rationale other linguistic behaviours, not
involving the use of proper names to refer to senses, will be ruled out for
similar reasons. If I say ‘Cicero is Tully. That is necessarily true’, what I say
will be true if the thought I express with the sentence ‘Cicero is Tully’ is a
necessarily true thought because I associate the same sense with the two
names, and I refer to that thought when I utter ‘That’ in my comment. But if
you associate different senses with the names, what you say, if you use
‘That’ to refer to the thought you express with ‘Cicero is Tully’, may be
false. So if you repeat what I say but use ‘That’ in your comment to refer to
the thought you expressed with your previous sentence and I used ‘That’ in
my comment to refer to the thought I expressed with my previous sentence,
there is again no guarantee that our utterances will have the same truth-
value, since they may have different subject matters. This indicates how a
Fregean can respond to the refined Kripkean modal argument of the
Preface.
We can now return to the puzzles relating to the failure of substitutivity
salva veritate of codesignating proper names within propositional attitude
contexts, which we noted as puzzles to be addressed by the widescope
descriptivist given the Lewisean rationale for his proposed convention.
Within the present Fregean framework the problems are: to account for the
non-equivalence of ‘Ralph believes that Cicero was an orator’ and ‘Ralph
believes that Tully was an orator’ (and hence to resolve Frege’s puzzle of
identity) if ‘Cicero’ and ‘Tully’ are not here understood as standing for their
ordinary senses; to explain on the same basis how ‘Ralph believes of Tully
that he denounced Catiline’ might be true whilst ‘Ralph believes that Tully
denounced Catiline’ is false; and to account for the possible truth of
ascriptions of propositional attitudes using empty names in the specification
of the content clauses.
There is intuitively a difference between George IV’s believing of Scott
that he is Scottish and his believing that Scott is Scottish, just as there is
between his believing of the author of Waverley that he is Scottish and his
believing that the author of Waverley is Scottish. ‘George IV believed that
Scott was Scottish’ (like ‘George IV believed that the author of Waverley
was Scottish’) is ambiguous. It can mean (a) ‘for some α, α is an (identity-
revealing) aspect of Scott and George IV believed that [α was Scottish]’ (in
which sense it is equivalent to ‘George IV believed of Scott that he was
Scottish’) or (b) ‘for some α, α corresponds to the sense of “Scott” in my
idiolect and George IV believed that [α was Scottish]’. Here
correspondence is that relation between senses which is required if an
ascriber is to be correct in his ascription, via the use of a proper name, of a
propositional attitude to an ascribee. Correspondence is not identity. If two
people use a name to refer to the same object, one deferentially to the other,
the senses they associate with the name are different, but correspondent.
Kripke’s Pierre (1979), before his kidnap, uses ‘Londres’ in a sense which
is different from, but correspondent with, the sense I associate with
‘London’ (as does any Frenchman), which is why we must say, on the basis
of his repeated assent to the French sentence, ‘Londres est jolie’, that he
believes that London is pretty. Correspondence does, however, require
identity of reference (though both senses may have no reference as in the
case of ‘Santa Claus’ and ‘Père Nöel’). But identity of reference is not
sufficient. It may be that nothing more general can be said and that what
correspondence requires is contextually determined (see for further
discussion Noonan 1979 and 1981; there are other similar discussions in the
literature, see in particular Chalmers 2011).
The solution to the puzzle of failure of substitutivity salva veritate within
propositional attitude ascriptions of codesignating proper names
(equivalently, of Frege’s puzzle of informative identity statements) is now
evident.
… believes that Phosphorus …
and
… believes that Hesperus …
Because the latter is unambiguously false ‘the one person who is richer than
anyone else in the world’ designates rigidly. The distinction between the
two types of description can be confirmed by noting the difference between
the questions ‘If Bill had given half his money to Ivana, who would have
been the person who would have had more money than anyone else in the
world?’ and ‘If Bill had given half his money to Ivana, who would have
been the person who has more money than anyone else in the world?’ The
answer to the first question may be ‘J. K. Rowling’ (suppose Bill has 100
billion dollars, J. K. 80, Ivana 20, and no one else is close). The answer to
the second question is, trivially, ‘Bill’.
So it appears that present-tense indicative descriptions in English, like the
ones given, are rigid designators by Kripke’s tests. The same is true of all
future-tense indicative descriptions, e.g. ‘The person who will give the John
Locke Lectures in 2150’, ‘the book Mary will be reading next month’. Past-
tense descriptions, e.g. ‘the book Mary was reading last month’, are
ambiguous: ‘If Mary had joined a different book club, the book Mary was
reading last month would not have been the book Mary was reading last
month’, has a true interpretation because the past tense is ambiguous
between an indicative and subjunctive reading. This also accounts for the
ambiguity of ‘If Mary had been sicker on Monday than Jane was on
Tuesday …’ (contrast the unambiguous ‘If Mary had been sicker on
Monday than Jane is today …’ which unambiguously makes a cross-
comparison between how things would have been in counterfactual
circumstances and how things are). Past-tense descriptions are rigid on the
indicative reading. In the case of the rigid present-tense description ‘the
book Mary is reading now’ change to the passive voice and omission of the
tensed verb introduce flexibility: ‘If she had changed book clubs the book
now being read by Mary would not have been the book now being read by
Mary’ has a true interpretation (note, then, that the temporal indexical does
not impose modal rigidity and can also occur in the subjunctive description,
‘the book Mary would have been reading now’). Insertion of the rigidifying
adverb, ‘actually’, to yield ‘the book actually being read by Mary now’
brings back rigidity. The only non-rigid descriptions, then, are ones, like
‘Mary’s current husband’, ‘the richest person in the world’ and ‘the
inventor of bifocals’, not containing tensed indicative verbs.
If this is accepted what is to be the response to the refined Preface
version of the modal argument in Naming and Necessity already
encountered? Here the modal argument goes as follows, adapted to
Burgess’s example. Suppose I fix the reference of a name ‘Richie’ by the
description ‘The person who has more money than anyone else in the
world’. Then the simple, non-modal, sentence ‘Richie is a female’
unambiguously expresses something that could not have been true. On the
other hand the simple, non-modal sentence ‘the person who has more
money than anyone else in the world is a female’ does not. Hence the name
and description differ in a way that the description theorist cannot
accommodate.
But why think that ‘Richie is female’ unambiguously expresses
something that could not have been true? Presumably because Richie is Bill
Gates and so it couldn’t have been that Richie was female (assuming one’s
gender is an essential property). But then it couldn’t have been that the
person who has more money than anyone else in the world was female
(though it could have been that the richest person in the world/the person
who had more money than anyone else was female, as it would have been if
Bill had given all his money to Ivana). And both ‘It could have been that
Richie is a female’ and ‘It could have been that the person who has more
money than anyone else in the world is a female’ are ungrammatical
(contrast the grammatical ‘It could be that Richie/the person who has more
money than anyone else in the world is a female’). But, a defender of
Kripke’s argument may rejoin, is it not the case that with respect to certain
counterfactual circumstances, e.g. if Bill Gates had given all his money to
Ivana, what ‘Richie is a pauper’ says would have been true, whereas with
respect to those circumstances, what ‘the person who has more money than
anyone else in the world is a pauper’ says would not have been true? But
why say that with respect to those circumstances what ‘Richie is a pauper’
says would have been true? Because it is unambiguously true that if those
circumstances had obtained Richie would have been a pauper (recall here
Dummett’s modal involvement principle). But equally, it is unambiguously
true that if those circumstances had obtained the person who has more
money than anyone else in the world would have been a pauper, though, of
course, it is not the case that if those circumstances had obtained the person
who would have had more money than anyone else in the world would have
been a pauper. When asked whether what ‘Richie is a pauper’ says might
have been true I answer the question by asking whether it might have been
true that Richie was a pauper, i.e. whether there is a way things could have
been such that if things had been that way it would have been that Richie
was a pauper. When asked whether what ‘the person who has more money
than anyone else in the world is pauper’ says might have been true I answer
the question by asking whether it might have been true that the person who
has more money than anyone else in the world was a pauper, i.e. whether
there is a way things could have been such that if things had been that way
it would have been that the person who has more money than anyone else in
the world was a pauper. In both cases I must answer affirmatively, though,
of course, I acknowledge that there are no ways things could have been
such that if things had been those ways it would have been the case that the
person who had more money than anyone else in the world was a pauper.
Of course, this final response only speaks to Kripke’s challenge to the
descriptivist to come up with descriptions in English which may plausibly
be thought of as equivalent to proper names, given that such names function
as rigid designators. Languages can be imagined, and, in fact, actually exist,
in which there is no distinction between indicative and subjunctive forms
and no descriptions function as required by this response. But it is no part of
the descriptivist position to deny the possibility of such languages. This is
obviously so since the descriptivist can acknowledge the possibility of a
language in which proper names occur but there are no definite descriptions
at all (what Kripke (1977) calls a ‘strong Russell language’); indeed, a
descriptivist who follows Russell, for whom definite descriptions are a mere
abbreviatory convenience, must insist on the possibility of such languages,
expressively equivalent to languages in which definite descriptions can be
formed. Thus the possibility of such languages is irrelevant to the issue
between Kripke and the descriptivists. If definite descriptions formed in
English using indicative verb forms are rigid designators it suffices to refute
the modal argument since the second premiss of that argument is that no
descriptions (at least of contingently existent objects) that can be at all
plausibly thought of as synonyms of proper names are rigid designators.
But indicative descriptions, like ‘the book Mary is reading’, form an
extremely large class of commonplace English expressions and are
plausibly available to users of proper names to define them.
I now conclude this section. The simple and straightforward modal
argument that dominates the early part of Naming and Necessity is hotly
contested. At least four ways of opposing it are in the literature. There is
much more to be said about each of these. However, Kripke has, of course,
a second, more ambitious line of argument against the descriptivist’s cluster
theory, beginning in the second lecture, which is designed to show that even
as a theory of reference-fixing it fails (and the debate over this argument
survives translation into a strong Russell language). If this succeeds the
modal argument is redundant (or at best necessary to sweep up ‘stragglers’,
atypical proper names like ‘Jack the Ripper’ and ‘St Anne’).
The Arguments Against the Cluster Theory Qua
Theory of Reference-Fixing
When he presents his alternative to the cluster theory, Kripke stresses the
word ‘picture’. He says that he is not attempting a theory (a set of necessary
and sufficient conditions, which he thinks is impossible), but only an
alternative way of looking at reference. Recall that the descriptivist story is
motivated in large part by the need to provide an account of how we
manage to connect symbols with the things to which they refer. The
descriptivist’s answer is: by association with properties (specified by
description). Since Kripke rejects this story he recognizes that he needs to
say something else. He does so by sketching his alternative picture:
Someone … is born; his parents call him by certain name. They talk about him to their friends. Other
people meet him. Through various sorts of talk the name is spread from link to link as if by a chain.
A speaker who is on the far end of the chain, who has heard about Richard Feynman, may be
referring to Feynman even though he can’t remember from whom he first heard of Feynman or from
whom he ever heard of Feynman. … He is then referring to Feynman even though he can’t identify
him uniquely. … a chain of communication going back to Feynman himself has been established, by
virtue of his membership in a community which passes the name from link to link, not by a ceremony
that he makes in private in his study: ‘by “Feynman” I shall mean the man who did such and such’.
(1980: 91–2)
This divides the account of reference into two parts: the initial fixation of
reference and the maintenance and transmission of reference. The former
concerns the first use of the name by the first user of the name, the latter,
later uses of the name by the first user and the first and later uses of the
name by later users who acquire it from an earlier one. The first part of the
theory is in accord with the description theory, except that the name is
introduced as rigid.
This division between the initial fixation of reference and the
maintenance of reference is important in relation to Kripke’s claims about
the contingent a priori and the necessary a posteriori. His examples of the
contingent a priori are cases in which something is a priori for the initial
baptizer who has used a description to fix the reference. These examples
depend on the contrast between the rigidity of the name and the non-rigidity
of the reference-fixing description. His explanation of the ‘illusion of
contingency’ associated with necessary a posteriori truths, i.e. the fact that
they are knowable only a posteriori, also depends on the association of the
name with a non-rigid description in the initial baptism. In both cases the
description linked initially with the name may cease to be associated with it,
so that what is initially contingent a priori for the first user may, for
subsequent users or the same user at a later time, be merely a posteriori, and
the explanation of the a posteriori status of the necessary a posteriori truth
known initially by the baptizer and retained by him or transmitted to others
will not be available for him later or for subsequent users.
At first sight it appears that there is a difference in kind between Kripke’s
explanation of reference and that of the descriptivist. One is given in terms
of facts about linguistic communities and the other in terms of facts about
the psychology of individual language users. But the contrast is overdrawn.
Individual language users’ intentions are still crucial to the Kripkean story:
(a) at the time of the initial introduction of the name and (b) when the name
is first passed on from one user to another. On the other hand, as we have
seen, the Frege– Russell descriptivist account allows for deference and the
thought that it is essential to my reference that I am a member of a
community.
This brings us back to the thought that no counterexample of the kind
Kripke envisages to the description theory can possibly succeed. As just
noted, speaker intentions play two crucial roles in Kripke’s picture. There is
an initial baptism in which the baptizer intends the reference of a name to
be fixed by description. The name is then passed on to a second user who
must intend when he picks it up to refer to the object to which the person
from whom he is picking it up referred – he cannot decide to initiate a new
use of the name as a name for his pet aardvark, say. And the same must be
true of all subsequent users. Now when the second user picks up the name
he will not count as intending to use the name with the same reference as
those from whom he acquired it just because he does in fact intend to use it
as a name for his pet aardvark and, in fact, quite unbeknownst to him, the
people he heard using the name were already using it as a name for his pet
aardvark. Rather, it must be that the content of his intention is: to use the
name to refer to whatever was the referent in the mouths of those from
whom he acquired it. Similarly, when the baptizer introduces the name he
must have an intention whose content is: to use the name to speak of the so-
and-so (or that object which … if he fixes the reference of the name
ostensively). But then neither the second user’s first use, governed by his
deferential intention, nor the baptizer’s first use, governed by his baptismal
intention, can possibly be a counterexample to the Frege–Russell
description theory as Kripke defines it. Counterexamples can only emerge
once the second user (to focus on him for the moment) ceases to have the
intention to refer to the object to which the people from whom he acquired
the name referred with it and acquires no other reference-determining
intention in its place (he cannot acquire the intention to refer to the object to
which he most recently referred with the name, for example, or to what
others in his community now refer to by the name, or to his own or Kripke’s
pet aardvark). Similarly, uses of the name by the baptizer can only emerge
as counterexamples to the description theory once the baptizer ceases to
have the intention to refer to the thing satisfying the description which
originally fixed the reference of the name for him and acquires no other
reference-determining intention in its place, not even the intention to refer
to the object to which he previously referred using the name. Such uses can
be counterexamples only if their reference is the object standing in a certain
causal-historical relation R to them, to be described by Kripke – which
requires that an object X would be acknowledged by the name user to be
the reference if evidence were provided to him that X did indeed stand in
relation R to his utterances – but the name user has no intention to refer to
the object standing in relation R to his utterances, nor any other intention
capable of determining that object as his reference. But Kripke never
describes a case satisfying these requirements. In fact, the intelligibility of
the stories he tells depends crucially on the fact that the user of ‘Gödel’,or
whatever, is a member of a community and intends to use the name in the
same way as others do – it is only this that makes it plausible that, despite
his ignorance and mistaken beliefs, he does not refer to Schmidt but does
succeed in referring to someone when he uses the name.
One thing that Kripke explicitly allows is the possibility of reference
change. In the ‘Napoleon’ example the man who decides to use ‘Napoleon’
as a name for his pet aardvark consciously opts out of a previous referential
practice and institutes a new use. But there is also the possibility of
referential shift without any such conscious decision. Kripke notes this
himself, acknowledging the contribution of Gareth Evans:
Gareth Evans has pointed out that … cases of reference shift arise … from one real entity to another
of the same kind. …‘Madagascar’ was a native name of part of Africa; Marco Polo, erroneously
thinking that he was following native usage, applied the name to an island. Today the usage of the
name as a name for an island has become so widespread that it surely overrides any historical
connection with the native name.
(Kripke 1980: 163)
Neither Marco Polo nor any subsequent European users of the name
consciously initiated a new practice. Each intended to use the name when
he received it with the same reference as those from whom he got it.
Nevertheless, reference shift occurred. Kripke offers the following
explanation:
… a present intention to refer to a given entity … overrides the original intention to preserve
reference in the historical chain of transmission. The matter deserves extended discussion. But the
phenomenon is perhaps roughly explicable in terms of the predominantly social character of the use
of proper names … We use names to communicate with other speakers in a common language. This
character dictates that ordinarily a speaker intends to use the name the same way as it was transmitted
to him, but in the ‘Madagascar’ case this social character dictates that the present intention to refer to
an island overrides this distant link to native usage.
(1980: 163)
Kripke here indicates his awareness of the crucial role of the present
intention to refer to the island in determining the present reference of the
name. Surely this is right. Initially Marco Polo was referring to part of the
African mainland and merely had mistaken beliefs about it – at this stage he
would have withdrawn his statement ‘Madagascar is an island’ if informed
of his error by a native speaker. But Kripke also indicates that he does not
think that such examples of reference change, and the role in explaining
them of speakers’ intentions, are problems for him or evidence in favour of
the description theory. He suggests later that the case may be understood as
one in which what is a speaker’s reference by an earlier user is
unintentionally transmitted as a semantic reference to later users (Kripke et
al. 1974: 512).
But this is problematic. Dummett argues as follows:
Kripke expressly wishes to allow that the association with a name of a description that does not apply
to the person … for which the name was originally introduced does not deprive the name of reference
to that person … It merely reveals a false belief about the reference. There is therefore no room in
Kripke’s account for a shift of reference in a chain of communication … Intuitively … there is no
such guarantee [of preservation of reference]. … Once this is conceded, the account crumbles away
altogether. We are left with this: that a name refers to an object if there exists a chain of
communication … at each stage of which there was a successful intention to preserve its reference.
This proposition is indisputably true; but hardly illuminating.
(1973: 150–1)
It goes without saying, of course, that it is consistent with this that the facts
reported in context by assertions of de re modality are entirely objective
non-modal facts of similarity. Which facts of similarity are relevant is
determined by context, but these facts, once selected, are in no sense
subjective.
It may be objected to this Lewisean story that it allows too much
inconstancy. De re modal claims about one and the same thing do not
exhibit wide context-dependent variation in truth-value in the way Lewis
suggests. It is not the case that whatever de re modal statement is in
question there is some context in which a counterpart relation is selected
which makes it literally true and another in which a counterpart relation is
selected which makes it literally false. However, the examples given
already suggest that that it will be hard for an opponent of Lewis to defend
the claim that any non-trivial modal predicates have context-independent
referents. But it is important to see also that where there is constancy it can
be accommodated within the Lewisean framework. It may be, for example,
that our intentions and conventions are so entrenched that certain proper
names are associated with connotations that determine certain counterpart
relations irrespective of any other features of the context. It may be that the
use of any name of a person brings to such salience the personal counterpart
relation that it is impossible for other features of the context to override it,
and it may be that the personal counterpart relation is such that nothing can
count as a personal counterpart of Saul Kripke unless it is a human being
with a father who is a personal counterpart of Rabbi Kripke. If so it will be
true in any context to say, ‘Kripke could not have had a different father’,
and the statement resulting from replacement of ‘Kripke’ by any other
proper name of Kripke, thought of as a name of a person, will also be true
in any context. But this will simply be a matter of what is rightly called ‘a
personal counterpart of Saul Kripke’. Nothing in this account is inconsistent
with the thesis of the context-relativity of essence. So, whilst the Lewisean
story is tailor-made to account for the inconstancy of de re modal
predication, it can also account for constancy if it occurs.
Turning back to Quine: he does not, of course, use the sort of examples in
his critique of quantified modal logic that have since become familiar and
have been given here since his focus is merely on establishing that when ‘is
necessarily so-and-so’ is understood in terms of ‘is necessarily true’,
interpreted as ‘is an analytic truth’, quantified modal logic is either trivial or
nonsense; once this has been established, he takes the context-relativity of
essence, in any other sense of essence, as evident and needing no
illustration. And, in fact, from the Lewisean viewpoint sketched Quine’s
example of 9/the number of planets – which, like his other examples, is
intended merely to illustrate that anything true of a thing follows
analytically from some way of specifying it – would not be best-chosen to
illustrate the context-relativity of essence. Inconstancy of modal
predication, as underpinned by a multiplicity of counterpart relations must
be irrelevant in this case since 9 is its own only counterpart under the
counterpart relation evoked by either ‘9’ or ‘the number of planets’ (though
that this is so, for a Lewisean, reduces to the mere fact that we do not treat
anything else, however similar, as a counterpart of 9, i.e. as relevant to the
truth-value of de re modal assertions about 9 (Lewis 1984: 235)).
Nonetheless, appeal to the sort of examples sketched here is entirely in
conformity with Quine’s position. For Quine material objects are simply the
material contents of spatio-temporal regions. So Goliath is Lumpl. The
piece of plastic is the dishpan. So the difference between our intuitive
ascriptions of truth-value to the claims ‘Goliath might have been rolled into
a ball and not destroyed/undergone complete material change’ and ‘Lumpl
might have been rolled into a ball and not destroyed/ undergone complete
material change’ is grist to his mill. Quine does not accept Lewis’s plurality
of (concrete) possible worlds, but as emphasized, the thesis that modal
predication is inconstant does not need development within the Lewisean
modal realist framework. What the thesis does require is that modal
predication is reducible to non-modal predication in the sense that what is
ascribed in context by the use of a modal predicate can be ascribed
independently of context by the use of a non-modal predicate. But there is
nothing here in tension with Quine’s thinking.
To sum up this section: Kripke insists against Quine that ascriptions of
essential and accidental properties to individuals make objective sense and
argues that this is the intuitive, common-sense, position. But there are
familiar intuitions Kripke cannot accommodate, and so it is doubtful
whether de re modal ascriptions have the sort of intelligibility he thinks
they have. The alternative view of essence as context-relative can explain
the recalcitrant intuitions, and even if there is more constancy than Lewis
thinks, it can be accommodated within his framework.
The Rejection of the Problem of Transworld
Identity and the Critique of Counterpart Theory
After defending essentialism Kripke goes on to criticize the idea that there
is a genuine problem of ‘transworld identity’ or ‘transworld identification’.
He suggests that this idea is one motivation people have for thinking that
the distinction between essential and accidental properties is unintuitive.
Given the context of these remarks it is tempting to read Kripke as thinking
of Quine as one of the people so motivated, though he does not say so
explicitly. Quine does indeed raise some worries about transworld identity
in ‘Propositional Objects’:
How is Catiline to be identified in … various possible worlds? Must he have been named ‘Catiline’
in each, in order to qualify? How much can his life differ from the real life of Catiline without his
ceasing to be our Catiline and having to be seen as another man of that name? Or again, how much
can the [Great] Pyramid differ from the real one? … Is it sufficient, for its identification in other
worlds that it have been built by Cheops? How much then can his life differ from the life of the real
Cheops without his ceasing to be our Cheops?
(Quine 1969: 153)
But it is clear that the primary focus of Quine’s worry is that, unlike Kripke,
he does not have any strong convictions about the truth-values of de re
modal statements. He does not have any answer to the question ‘How
different from the way he is might Catiline have been?’ Quine’s worries
about transworld identity are not the basis of his doubts about de re modal
predication. He is merely using the language of possible worlds to express
them. The basis of his doubts is his long-standing rejection of context-
independent essence.
Other philosophers who were waxing sceptically at the time about
transworld identity were similarly primarily sceptical about the
intelligibility and truth-values of de re modal ascriptions rather than about
statements of transworld identity expressed in the language of possible
worlds. Thus Chisholm discusses the possibility of a world in which Adam
and Noah ‘swap roles’ and the consequent apparent possibility that there
could be two worlds (the actual one and this one) discernible only inasmuch
as in one Adam plays the Adam role and Noah the Noah role and in the
other things are reversed. The existence of this pair of worlds can be easily
ruled out if it is accepted that Adam and Noah differ in their essential
properties. However, this is precisely what worries Chisholm. He does not
see how we can tell which of Adam’s properties are essential and which
accidental:
What … properties … are essential to Adam? Being the first man? Having a name which, in English,
begins with the first letter of the alphabet? But why these properties? … It seems to me that even if
Adam does have such essential properties, there is no procedure at all for finding out what they are.
And it also seems to me that there is no way of finding out whether he does have any essential
properties.
(Chisholm 1967: 6)
But:
A possible world isn’t a distant country … a possible world is given by the descriptive conditions we
associate with it. …‘Possible worlds’ are stipulated, not discovered by powerful telescopes. There is
no reason why we cannot stipulate that, in thinking about what happened to Nixon in a certain
counterfactual situation, we are talking about what would have happened to him.
(Kripke 1980: 44)
Of course, if someone makes the demand that every possible world has to be described in a certain
qualitative way, we can’t say ‘Suppose Nixon had lost the election’. Why … make this demand? That
is not the ordinary way we think about counterfactual situations. We just say ‘Suppose this man had
lost’. … even if there were a purely qualitative set of necessary and sufficient conditions for being
Nixon, … I … would not demand that we find these conditions before we can ask whether Nixon
might have lost the election, nor does it demand that we restate the question in terms of such
conditions. We can consider Nixon and ask what might have happened to him had various
circumstances been different.
(Kripke 1980: 46–7)
Again:
Don’t ask: how can I identify this table in another possible world except by its properties? I have the
table in my hand, I can point to it, and when I ask whether it might have been in another room, I am
talking, by definition, about it. I don’t have to identify it after seeing through a telescope.
(Kripke 1980: 52)
Two points are being emphasized in these passages. First, possible worlds
are ‘stipulated not discovered’. Second, I can identify an individual about
which I can go on to make a supposition without giving a purely qualitative
necessary and sufficient condition – just as ‘this man’ or ‘Nixon’.
When Kripke says that possible worlds are stipulated not discovered he is
not claiming that I can stipulate that something is possible, as I can stipulate
that I will use a new word, say ‘squelschy’, as synonymous with, say,
‘wise’. Whether there is a possible world at which so-and-so is the case,
whether it is possible that so-and-so is the case, is, in Kripke’s view, an
objective matter of fact, which is so or not independently of any stipulation.
But I can stipulate that my counterfactual supposition is about a particular
individual, say Nixon, that is, that I am to be regarded as speaking correctly
when I say ‘Nixon might have lost’ if and only if Nixon might have lost. I
might say ‘The PM might have been a member of the Conservative Party’
and to ward off misunderstanding say, ‘I mean that man, who is in fact PM,
might have been a member of the Conservative Party’. Here I am
stipulating that my ambiguous utterance is to be read with the description
‘the PM’ given wide scope, and so that I am entertaining the possibility of a
situation in which a particular actual individual has a certain property.
But it does not follow from the possibility of such a stipulation that I can
guarantee, by stipulating that I am considering a possible world in which
Nixon, that very man, lost, that there is such a world. If I stipulate that I am
considering a possible world in which Nixon had different parents I do not
thereby guarantee that there is a possible world in which he had different
parents. In fact, if Kripke’s own ‘necessity of origin’ thesis is correct, there
is no such possible world. This point may be put (Mackie 2008) by saying
that if Kripke’s thesis of the necessity of origin is correct then in stipulating
that I am talking about Nixon in saying that Nixon might have lost, I am
implicitly, whether I know it or not, stipulating that I am talking about
someone with Nixon’s actual parents, and that my supposition is not merely
that Nixon might have lost, but that someone with Nixon’s actual parents
might have lost.
Kripke’s second point is that I can identify someone using a name or
demonstrative expression, as ‘Nixon’ or ‘this man’, and go on to make a
statement about what might have happened to him, without having any
knowledge of necessary or sufficient qualitative conditions for being him.
There is no more difficulty in understanding the statement that there is a
possible world in which Nixon lost the election than there is in
understanding the statement that there is a possible world in which a man of
such and such a qualitative description lost the election. And this is so, at
least, because the former statement means no more than the familiar modal
claim ‘Nixon might have lost the election’ (‘It is better … to avoid
confusion, not to say “In some possible world, Humphrey would have
won”, but rather, simply, “Humphrey might have won”’ (Kripke 1980: 48)).
It is hard to grasp the point of Kripke’s remarks about trans-world
identity because the target is a position no one has ever held. It has two
components. The first is the Lewisean reductionist notion of a possible
world as a maximal sum of spatio-temporally related parts, according to
which other possible worlds are indeed to be thought of as like other
countries. We do not need modal vocabulary to explain this notion of a
possible world. This is the ‘foreign country’ conception Kripke repudiates.
The second component of Kripke’s target is the idea that some individuals
(the same individuals) ‘live simultaneously’ in these different countries.
This is no part of Lewis’s conception – as Kripke emphasizes. Lewis thinks
that no individual is part of several worlds, he merely has counterparts in
several. Kripke’s arguments that the problem of transworld identity is a
pseudo-problem are aimed at showing that if, as he thinks, the first
component of this picture is mistaken, i.e. that the notion of a possible
world is rather the notion of a possible state or history of the one actual
world, there is no genuine question to be asked about the necessary and
sufficient condition for an individual to be an ‘inhabitant’ of two possible
worlds. His remarks seem bafflingly platitudinous because they are
platitudinous – in the same sort of way a mathematical proof can seem
obvious once it is given.
However, there are contentious issues on which Kripke takes a stand
which can be, but need not be, phrased in the language of transworld
identity. In the Preface to Naming and Necessity Kripke indicates one such
issue and his stance on it in a footnote (1980: 18): ‘if there is a
philosophical argument excluding qualitatively identical but distinct worlds,
it cannot be based exclusively on the supposition that worlds must be
stipulated purely qualitatively’. The world in which Adam and Noah have
switched roles and the actual world would be such a pair of indiscernible
but distinct worlds, but the existence of this world can be ruled out given
Kripke’s thesis of the essentiality of biological origin, because such a world
would be one in which Adam and Noah had origins other than their actual
ones. However, Kripke’s thesis of the essentiality of origin does not rule out
another possible world in which someone other than Noah, an actual or
possible individual with no essential property incompatible with Adam’s
origin, plays the Adam role, for, as he emphasizes, his thesis is one about
the necessary conditions of being an individual, not the sufficient
conditions. To rule out such a world it seems that we must suppose that
every individual has a non-trivial individual essence, a property essential to
that individual and necessarily not possessed by any other. Kripke does not
take a stand on whether (all) individuals have individual essences. He does
not assert or deny that some possible or actual individual other than Adam
could have played the Adam role. But he does assert that there is no
argument for denying this based simply on the supposition that worlds, i.e.
world states, must be stipulated purely qualitatively, since they need not be.
As noted, Kripke recognizes that although Lewis adheres to the first
component of his ‘foreign country’ target picture of possible worlds, he
does not think that there are individuals living simultaneously in distinct
worlds, but rather thinks that individuals are world-bound, having only
counterparts in other worlds. In a footnote, bracketing his scepticism about
Lewisean reductionism, Lewis presents his famous ‘Humphrey’ objection
to counterpart theory.
[For Lewis] if we say ‘Humphrey might have won the election’, we are not talking about something
that might have happened to Humphrey, but to someone else, a counterpart. Probably, however,
Humphrey could not have cared less whether someone else, no matter how much resembling him,
would have been victorious in another world. Thus Lewis’s view seems to me even more bizarre than
the usual notion of transworld identification that it replaces.
(Kripke 1980: 45 n. 13)
This argument has generally been found unconvincing. The starting point
is that when we say ‘Humphrey might have won’ we mean that Humphrey
might have won, i.e. that Humphrey himself might have won. We are
talking about Humphrey. But on the counterpart theory we are talking about
someone else, a counterpart to Humphrey. So counterpart theory gets the
truth-conditions wrong, it assigns truth-conditions to our modal statements
which render them irrelevant to our concerns.
The objection to this argument is that according to counterpart theory
‘Humphrey might have won’ is about Humphrey, as well as being about
someone else; it says that Humphrey, the man himself, has a winning
counterpart. Lewis expresses this response to Kripke himself:
Counterpart theory does say that someone else … enters into the story. … Insofar as the intuitive
complaint is that someone else gets into the act, the point is rightly taken. But I do not see that is an
objection, any more than it would be an objection against ersatzism that some abstract whatnot get
into the act. What matters is that someone else … should not crowd Humphrey out. And there all is
well.
(1986: 196)
Having argued against Quine that the distinction between essential and
accidental properties makes objective, context-independent, sense, Kripke
goes on in the third lecture to defend specific non-trivial essentialist claims.
In particular, he defends the necessity of biological origin, the necessity of
material origin for artefacts, the necessity of constitution for material
objects and the necessity of identity and distinctness for everything.
In a famous passage, prompted by an article by Timothy Sprigge, Kripke
affirms that Queen Elizabeth II could not have had different parents, and, in
particular, could not have been the biological daughter of Mr and Mrs Harry
Truman:
They might have had a child resembling her in many respects. Perhaps … even … a child who
actually became the Queen of England. … This still would not be a situation in which this very
woman whom we call ‘Elizabeth II’ was the child of Mr and Mrs Truman. … How could a person
originating from different parents, from a totally different sperm and egg, be this very woman? One
can imagine, given the woman, that various things in her life would have changed. … One is given,
let’s say, a previous history of the world up to a certain time, and from that time it diverges from the
actual course. … And so it is possible that even though she were born of those parents she never
became Queen. … But what is harder to imagine is her being born of different parents. It seems to
me that anything coming from a different origin would not be this object.
(1980: 112–13)
It is the most natural thing in the world to ponder on how things would have
been if one’s parents had been different. ‘If I had been the son of the King
I’d have had more opportunities in life’.
Thinking about origin more generally the same point holds. It is natural
to say in an appropriate context that the same building might have been
built of a different lot of bricks or the same ship of a different batch of steel
(Strawson 1979). We understand Lewis very well when he says, describing
his dishpan/plastic case ‘… suppose the factory had received its order for
plastic dishpans a day later. The same piece of plastic would have been
made. But … it would have constituted a wastebasket. The dishpan would
have been made the next day out of different plastic’ (1986: 252). There is a
clash of intuitions, and such a clash cannot itself constitute a reason to
endorse Kripke’s intuitions rather than those of Dummett, Wiggins,
Strawson and Lewis.
But Kripke also gives or suggests several arguments in support of the
necessity of origin.
The first argument, suggested by Kripke’s remark that when considering
alternative possibilities ‘one is given … a previous history of the world, and
from that time it diverges from the actual course’ is spelled out by
Dummett:
It may truly be said of President Nixon, for instance, that he might never have been a politician,
because there was a time in his life at which it would have been true to say that he might never
become a politician. … But we cannot push back the moment in respect of which a property is to be
characterised as [something it might not come to possess] beyond the point at which the object came
into existence.
(Dummett 1973: 131)
If all possibilities for Elizabeth are limited in this way to divergences from
actuality after she came into existence she could not have had different
parents. But the principle is absurd since it makes all features she possessed
at her origin (however intuitively accidental or relational) essential. Kripke
does not in fact endorse the principle. However, he does tentatively endorse
a slightly weaker one which requires that possibilities for Elizabeth must
involve divergences from the actual course of world history but allows that
such divergences may take place before Elizabeth came into existence:
Ordinarily when we ask intuitively whether something might have happened to a given object, we
ask whether the universe could have gone on as it actually did up to a certain time, but diverged in its
history from that point forward so that the vicissitudes of that object would have been different from
that time forth. Perhaps this feature should be erected into a general feature about essence. Note that
the time at which the divergence from actual history occurs may be some time before the object itself
is actually created.
(1980: 115)
The argument here is brief but appears to rely on an appeal to the necessity
of origin. Cases in which it is plausible to say that an artefact is originally
made of completely different matter in two circumstances also suggest that
it is plausible that an artefact might be originally made of different kinds of
matter in different circumstances. Even if the necessity of origin is allowed
the argument is not convincing unless it is assumed that if a wooden table is
originally made of a block of wood that block is both essentially originally
wooden and could not have changed over time from being wood to being
some other kind of substance. Otherwise, to imagine the table in fact made
of the block originally made of a different kind of substance does not
require going back over the entire history of the universe but only
imagining a change in the block in the course of its existence.
At any rate, it is clear that the argument for the necessity of constitution
Kripke suggests requires a substantial essentialist commitment as a premiss.
It cannot, nor of course does Kripke ever suggest it can, itself convince
someone with no prior commitment to essentialism.
The necessity of identity and distinctness figure prominently (as
premisses) in Kripke’s last lecture (and implicitly, as we have seen, in his
arguments for the necessity of origin). The argument for the necessity of
identity is given briefly in the Preface to Naming and Necessity and more
carefully in the much earlier ‘Identity and Necessity’. In the Preface he
writes (p. 3):
Waiving fussy considerations derived from the fact that x need not have necessary existence it was
clear from (x)&(x = x) and Leibniz’s Law that identity is an ‘internal’ relation: (x)(y)(x = y ! &x = y).
(What pairs < x, y > could be counterexamples? Not pairs of distinct objects, for then the antecedent
is false; nor any pair of an object and itself, for then the consequent is true.)
(2) (x)◻(x = x)
But:
is a substitution instance of (1), the substitutivity law. From (2) and (3), we can conclude that, for
every x and y, if x equals y, it is necessary that x equals y:
This is because the clause ◻(x = x) of the conditional drops out because it is known to be true.
(1971: 135–6)
This is an argument for the thesis that an identity pair (any pair of objects x
and y such that x = y) is necessarily an identity pair; similarly Kripke could
argue that a distinctness pair is necessarily a distinctness pair. Nothing
follows about the necessary truth of any statement, much less the necessary
a posteriori truth of any statement. Kripke emphasizes the necessary
distinctions himself:
We must distinguish three distinct theses: (i) that identical objects are necessarily identical; (ii) that
identity statements between rigid designators are necessary; (iii) that identity statements between
what we call ‘names’ in actual language are necessary.
(1980: 4)
Kripke returns to this issue and restates and generalizes this argument in the
third lecture:
Now in spite of the arguments I gave before for the distinction between necessary and a priori truth,
the notion of a posteriori necessary truth may still be somewhat puzzling. Someone may … argue as
follows. You have admitted … that this table might have turned out to be made from ice from water
from the Thames. I gather that Hesperus might have turned out not to be Phosphorus. What can you
mean when you say that such eventualities are impossible? If Hesperus might have turned out not to
be Phosphorus, then Hesperus might not have been Phosphorus.
(1980: 140–1)
Kripke responds to this by invoking the idea of a qualitatively identical
epistemic situation:
The objector is correct … that if I hold that this table could not have been made of ice, … I must …
hold that it could not have turned out to be made of ice; it could have turned out that P entails that P
could have been the case. What, then, does the intuition that the table might have turned out to have
been made of ice … amount to? I think it means simply that … I (or some conscious being) could
have been qualititatively in the same epistemic situation that actually obtains, I could have the same
sensory experience that I in fact have about a table, which was made of ice.
(1980: 141–2)
He sums up:
The general answer to the objector can be stated, then, as follows: any necessary truth, whether a
priori or a posteriori, could not have turned out otherwise. In the case of some necessary a posteriori
truths, however, … under appropriate qualitatively identical evidential situations an appropriate
corresponding qualitative statement might have been false. … The inaccurate statement that
Hesperus might have turned out not to be Phosphorus should be replaced by the true contingency
[that] two distinct bodies might have occupied, in the morning and the evening, respectively, the very
positions actually occupied by Hesperus–Phosphorus–Venus.
(1980: 142–43)
These passages contain Kripke’s response to the sceptic about the necessary
a posteriori who is prepared to accept that the statements in question are
necessary (unlike the Fregean sceptic mentioned earlier) but wonders why
they should be called a posteriori. Kripke’s argument is that we can only
know a posteriori that ‘Hesperus’ and ‘Phosphorus’ are coreferential, and
hence that ‘Hesperus is Phosphorus’ expresses a truth. So we can only
know a posteriori that Hesperus is Phosphorus. Or again, that we can only
know a posteriori that the heavenly body visible in the evening is the
heavenly body visible in the morning. So we can only know a posteriori
that Hesperus is Phosphorus. In the case of the table the argument is that
Kripke can only know a posteriori that the table in front of him is not made
of ice (he can only know a posteriori what he expresses by saying ‘the table
in front of me is not made of ice’) and hence he can only know a posteriori
that what he expresses by saying ‘this table is not made of ice’ (or, if he
introduces ‘T’ as a rigid designator of the table in front of him, ‘T is not
made of ice’) is the case. Thus it is necessary but only knowable a posteriori
that Hesperus is Phosphorus and that this table (T) is not made of ice.
The reason for doubting that this is an adequate explanation of the a
posteriori status of these statements is brought out by Kripke himself in his
paper ‘A Puzzle about Belief’ (1979). If we take seriously Kripke’s thought
that the description originally associated by a speaker with a name is
inessential to his understanding of it, so that once introduced by a
description a name can later be used by the original baptizer or other
speakers in association with no, or quite different, descriptions, it becomes
quite unclear why evidence for the associated contingent statement
(whether this is thought of as the metalinguistic statement that ‘Hesperus’
and ‘Phosphorus’ are coreferential or the statement that the heavenly body
seen in the night sky is the heavenly body seen in the morning sky) is
necessary for justified belief that Hesperus is Phosphorus. Before acquiring
evidence that ‘Hesperus’ is coreferential with ‘Phosphorus’ a competent
user of the names justifiably believes that Hesperus is Hesperus. The
question is why he does not also already justifiably believe that Hesperus is
Phosphorus despite not believing that ‘Hesperus’ and ‘Phosphorus’ are
coreferential. Kripke’s story of puzzling Pierre brings out the urgency of
this question. Pierre, a monolingual French speaker living in Paris believes
that London is pretty (on the basis of the pictures he has seen), a belief he
expresses by saying in French ‘Londres est jolie’. He is then kidnapped and
transported to an ugly area of London, where he has to learn English by the
direct method. Disliking his surroundings and never thinking that the city
he has been taken to is the pretty city in the pictures he does not assent to
‘London is pretty’ and repeatedly assents to ‘London is ugly’, whilst often
repeating to himself ‘Londres est jolie’. As Kripke asserts, it is undeniable
that Pierre continues to believe after his kidnap that London is pretty and is
justified in doing so, although he does not believe that the sentence ‘London
is pretty’ is true and would not be justified in doing so. Nor, of course, does
he believe, nor would he be justified in believing, that the city he is living in
is pretty. Granted that it is a necessary truth that Hesperus is Phosphorus,
therefore, it is unclear how, given his own views, Kripke’s argument in the
quoted passage can establish that it is knowable only a posteriori. What
Kripke appears to be assuming is that in order to be justified in believing
that Hesperus is Phosphorus someone must be justified in believing that the
sentence ‘Hesperus is Phosphorus’ is true, or that the heavenly body in the
evening sky is the heavenly body in the morning sky. But the unfortunate
Pierre is justified in believing that London is pretty despite not justifiably
believing either that the sentence ‘London is pretty’ is true or that the city in
which he is now living is pretty (even though the only reference-fixing
description he has for ‘London’– as opposed to ‘Londres’– is the
description ‘the city in which I am now living’). So why should a justified
belief that Hesperus is Phosphorus have to fulfil these further requirements?
To see the point from a slightly different perspective, consider the
proposition that London is Londres. Before his kidnap Pierre believes that
London is London, a proposition he expresses by saying ‘Londres est
Londres’. After his kidnap he retains this belief. After his kidnap he is not
prepared to say (in an English– French word-salad) ‘Londres is London’
(anymore than he is prepared to say ‘London is jolie’) and in fact would
insist that what this expresses is false. But the necessary truth he
nonetheless expresses with this sentence, according to Kripke, is the
necessary truth that London is London, which he believes. So he believes
that London is Londres, just as he believes that London is pretty. That he is
not prepared to say in the French–English word-salad ‘London is Londres’
is irrelevant. Just so, then, a sceptic about the a posteriori status of
‘Hesperus is Phosphorus’ will say, the fact that someone is not prepared to
say ‘Hesperus is Phosphorus’ is not evidence that he does not believe that
Hesperus is Phosphorus.
The same point can be made about Kripke’s other purported examples of
necessary a posteriori truth. Given that it is a necessary truth that this table
is not made of ice (or, if ‘T’ is a rigid designator whose reference is fixed by
Kripke by the description ‘the table now in front of me’, that it is a
necessary truth that T is not made of ice), Kripke’s argument that it is not
knowable a priori that T is not made of ice (if it exists) is dubious. Given
that the table is not made of ice a correct description of it is ‘the table here
which is not made of ice’. If a rigid designator ‘T*’ is introduced with its
reference fixed by this description it is, according to Kripke’s account,
knowable a priori that T* is not made of ice if it exists. So if it can also be
known a priori that T* is T and if the possibility of a priori knowledge is
closed under logical implication, it can be known a priori that T, if it exists,
is not made of ice. Setting aside this argument, the dubiousness of Kripke’s
claim that it can only be known a posteriori that T (this table) is not made of
ice, if it exists, is once again highlighted by direct comparison with the case
of puzzling Pierre. Although Pierre’s evidential situation after his kidnap is
such that he is not warranted in assenting to the English sentence ‘London
is pretty’, he is still justified in believing that London is pretty since he is
justified in believing to be true the sentence ‘Londres est jolie’, which he
understands and says that London is pretty. Just so, a sceptic about Kripke’s
argument can say, although my evidential situation when I confront the
table does not warrant me in assenting to the English sentence ‘T is not
made of ice if it exists’ I can still be justified without that evidence in
believing that T is not made of ice if it exists since I can be justified in
believing that T* is not made of ice if it exists. I understand the sentence
‘T* is not made of ice if it exists’ and justifiably believe that what it says is
true; since it says that T is not made of ice if it exists I am justified in
believing that T is not made of ice if it exists despite having no justification
for believing that the sentence ‘T is not made of ice if it exists’ is true or
that the table in front of me if it exists is not made of ice. (The same
considerations can be used to cast doubt on the idea that non-identities are
necessary a posteriori truths. The sentence ‘If they exist Venus and Mars are
different entities’ can only be known to be true empirically, but if I fix the
reference of the rigid designator ‘Mars*’ as ‘the planet which is Mars and is
other than Venus’ I can know a priori that if they exist Venus and Mars* are
different entities. The considerations sketched above can now be brought to
bear.)
Consider also the example of the number of planets, which can be
explained more simply since no existential assumption need be made. If I
introduce ‘N’, with its reference fixed by the description ‘the number of
planets’, in ignorance of how many planets there are, according to Kripke
what I express by ‘N is greater than 7’ is a necessary truth. But accepting
this claim the question can be pressed why it is knowable only a posteriori.
It is true that without empirical justification I cannot know that ‘N is greater
than 7’ expresses a truth or that the number of planets is greater than 7. But
I can know a priori that 9 is greater than 7. If I can also know a priori that 9
= N, I can know a priori that N is greater than 7 if possible a priori
knowledge is closed under logical implication. Independently of this
argument the dubiousness of the Kripkean claim that it can only be known a
posteriori that N is greater than 7 is once again highlighted by direct
comparison with the case of Pierre. Pierre is justified in believing that
London is pretty although he has no evidential warrant for believing to be
true the sentence ‘London is pretty’.Heisso justified since he is justified in
believing to be true the sentence ‘Londres est jolie’, which he understands
and which says that London is pretty. Just so, a sceptic about Kripke’s
argument can say, although my evidential situation when I view the heavens
does not warrant me in assenting to ‘N is greater than 7’ I can still be
justified without that evidence in believing that N is greater than 7 since I
can be justified in believing that 9 is greater than 7.
In short, there is a unresolved tension in Naming and Necessity which
Kripke himself draws attention to in ‘A Puzzle about Belief’. It is tempting
to think that if proper names are, as Kripke argues, de jure rigid designators,
grasp of which is not essentially tied to any descriptive content, they must
be substitutable salva veritate not only in modal contexts but also in
epistemic contexts. As Kripke puts it, adopting some terminology from
Geach (1962), such contexts, it appears, must be ‘Shakespearian’ (alluding
to ‘a rose by any other name’) if proper names are de jure rigid designators
without descriptive content. But if so, his putative examples of the
necessary a posteriori no longer qualify as such. On the other hand, if
proper names are not substitutable salva veritate in epistemic contexts, the
most readily available explanation is some version of the Frege–Russell
view, according to which the examples cease to be examples of the
necessary a posteriori, because they cease to be examples of necessary
truth. (The defender of the Frege–Russell view will accept that it is correct
to say, on the basis of what he says in French, that Pierre believes that
London is pretty but deny that just because I assent to ‘9>7’ and so can be
said to believe that 9 > 7 I can also be said to believe that N is greater than
7; he can explain the difference by the correspondence between the senses
of ‘London’ and ‘Londres’, which is lacking in the case of ‘9’ and ‘N’.)
Defenders of the Kripkean necessary a posteriori need to fight on two
fronts: to deny that epistemic contexts are Shakespearian, whilst
maintaining that modal contexts are.
The challenge the Kripkean faces in explaining the necessary a posteriori
can be put another way. The task for him is to provide an explanation,
without appeal to resources only available to the descriptivist, of the
informativeness of true identity statements of the form ‘a = b’ which cannot
also be deployed to ‘explain’ the informativeness of true, but
uninformative, identity statements of the form ‘a= a’. Suppose I use
‘Cicero’ to refer to Cicero, but without any associated identifying
description and under the misapprehension that Cicero was a Greek
admiral. I am in the same position with respect to ‘Tully’. I am open to the
thought that there is more than one Greek admiral called ‘Cicero’ and
equally open to the thought that there is more than one Greek admiral called
‘Tully’. Kripke needs to explain, without using descriptivist resources, how
in this situation, which according to him is absolutely typical, I can be
informed by the statement ‘Cicero is Tully’, but not by the statement
‘Cicero is Cicero’. Equivalently, he needs to explain how the information
provided to me by a statement of the form ‘Cicero is F’ can be different
from that provided to me by a statement of the form ‘Tully is F’, and to
explain this in a way that does not entail that successive token statements of
the form ‘Cicero is F’ will provide me with different pieces of information.
Kripke’s putative examples of contingent a priori truth also crucially
involve rigid designators. The general form of a candidate statement is
‘aisF’ or ‘if a exists a is F’ where ‘a’ is a rigid designator whose reference is
fixed by the description ‘the F’ and being F is an accidental property of the
thing designated by ‘a’. Kripke’s most discussed example is the length of
the standard metre rod, but he thinks that the same mechanism by which a
priori contingent knowledge can be generated in this case can be used to
generate it in other cases also: for example ‘Jack the Ripper’ (where the
reference is fixed by the description ‘the man, whoever he was, who
committed all those murders, or most of them’) and ‘Neptune’ (‘the planet
if there is one causing such and such perturbations in the orbits of these
other planets’). Other philosophers have introduced their own examples.
One of the most well-known is Gareth Evans’s example of Julius, the
inventor of the zip. Dummett suggests the example of St Anne, the
reference-fixing description being ‘the mother of the Blessed Virgin Mary’.
Kripke argues for the contingent a priori status of his examples as
follows. Suppose we fix the reference of ‘one metre’ by means of the
stipulation: one metre is to be the length of stick S at time t. Stick S is an
ordinary rod chosen as the standard of length. It is not a necessary property
of stick S that it is one metre long at t, since it is an accidental feature of
stick S that it is the length it is at t (equivalently, it is a contingent property
of the length one metre that it is the length of stick S at t). So the
description ‘the length of stick S at t’ is an accidental designator of the
length. The length of stick S at t might not have been the length of stick S at
t. So one metre might not have been the length of stick S at t. So it is not a
necessary truth that one metre is the length of stick S at t, i.e. the statement
‘It is a necessary truth that one metre is the length of stick S at t’ is false if
the description ‘the length of stick S at t’ is read as having narrow scope.
Consequently ‘one metre is the length of stick S at t’ is not a necessary
truth. However, on the basis of the reference-fixing stipulation the stipulator
can know a priori that one metre is the length of stick S at t (if stick S exists
at t), automatically, without further investigation, or at least he can do so if
someone who stipulates that ‘squatchelor’ is to apply to all and only
unmarried men can know a priori that all and only squatchelors are
unmarried men. How can someone who stipulates that ‘squatchelor’ is to
apply to all and only unmarried men come to know that all and only
squatchelors are unmarried men? He can reason as follows: ‘All and only
those things to which “squatchelor” as-I-use-it-at-the-moment applies are
squatchelors. All and only those things to which “squatchelor” as-I-use-it-
at-the-moment applies are unmarried men. So all and only squatchelors are
unmarried men.’ Similarly Kripke’s stipulator can reason: ‘All and only
those things to which “a metre long” as-I-use-it-at-the-moment applies are
things which are a metre long. All and only those things to which “a metre
long” as-I-use-it-at-the-moment applies are the same in length as stick S at
t. So all and only those things which are a metre long are the same in length
as stick S at t.’ If the former reasoning provides a priori justification for the
truth that all and only squatchelors are unmarried men the latter provides a
priori justification for the truth that things a metre long are the same length
as S at t.
The argument generalizes. It applies whenever the reference of a rigid
designator is fixed by a description giving a contingent property of its
bearer. Dummett’s example of St Anne can be used to illustrate. Very little
is known of St Anne, and the reference can be taken to be fixed by the
description ‘the mother of the Virgin Mary’. St Anne might have died in
infancy, so she might not have been the mother of Mary, or a parent at all.
Consequently it is not a necessary truth that St Anne was the mother of
Mary nor a necessary truth that St Anne was a parent. On Kripke’s account,
then, neither ‘St Anne was the mother of Mary’ nor ‘St Anne was a parent’
is a necessary truth. But it is knowable a priori that if she existed St Anne
was the mother of Mary.
In these cases the referents are contingent existents so what Kripke
claims to be knowable a priori are conditional statements. But we can
appeal to the example of the rigid designator ‘N’ of the number of planets.
Since ‘the number of planets’ must have reference it can be known, on
Kripke’s account, completely a priori that the categorical statement that N is
the number of planets is true.
What is puzzling about the contingent a priori, Kripke suggests, can be
brought out as follows:
If someone fixes a meter as ‘the length of stick S at t’, then in some sense he knows a priori that the
length of stick S at t is one meter, even though he uses the statement to express a contingent truth.
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, even though it is undeniably a contingent fact that S is one meter long.
(1980: 63)
Kripke’s views about natural kind terms are opposed to what we can call the
‘cluster of identifying marks’ theory, according to which such terms as
‘water’ are, in their primary usage, predicates whose meaning and extension
are fixed by the cluster of properties we regard as the criteria for their
application. This cluster of identifying marks theory of natural kind terms is
obviously similar to the cluster of identifying descriptions theory of proper
names and was held by much the same philosophers, e.g. Wittgenstein, as
Kripke emphasizes. It is because of this that Kripke thinks it appropriate to
describe his views by saying that he holds, like Mill, that proper names
have no connotation, but goes beyond Mill in extending the same
conclusion to natural kind terms. The question we have to keep carefully in
mind, then, is how far Kripke’s arguments against the cluster of identifying
marks theory of natural kind terms, construed as predicates, also qualify as
arguments against the cluster of identifying descriptions theory of natural
kind terms, construed as proper names of kinds. I shall conclude, not very
far.
Kripke proceeds against the cluster of identifying marks theory by way of
examples.
Consider the term ‘tiger’. According to the cluster of identifying marks
theory ‘x is a tiger’ means ‘x has such and such (superficially observable)
properties’, e.g. is large, striped, ferocious and cat-like in appearance. But,
Kripke argues, even if these were all the known properties of tigers, as
presumably they were for primitive man, it would not be the case that any
possible creature having these properties was a tiger, nor even that any
actual creature having these properties was a tiger. Suppose that in a
counterfactual situation there are creatures having the appearance of tigers
which are, in fact, not mammals at all, but very peculiar reptiles, then these
reptiles are not tigers because they have the wrong internal structure.
Suppose that, on another planet in the actual universe or in an unexplored
part of the globe, such creatures really do exist; still they are not tigers –
their discovery would not be the discovery that there were reptilian tigers as
well as mammalian tigers. If it is accepted that the tiger-like reptiles in the
counterfactual situation are not tigers, then the cluster of identifying marks
story does not give the meaning of ‘x is a tiger’, and if the tiger-like reptiles
would not be tigers even if they existed in the actual world, it cannot be that
this cluster even fixes the extension of ‘x is a tiger’, i.e. the cluster of
identifying marks story must be wrong both as a theory of meaning and as a
theory of reference- or extension-fixing for natural kind terms.
Moreover, Kripke argues, there might have been and might be tigers that
do not have all the properties involved in the cluster of identifying marks,
or indeed any of the properties in the cluster. A three-legged tiger is not a
contradictio in adjecto. It might be that no tigers have four legs (we are
deceived by an optical illusion). Indeed it might have been and might be
that tigers have none of the properties by which we originally identified
them.
Perhaps none are quadrupedal, none tawny yellow, none carniverous, and so on; all of these
properties turn out to be based on optical illusion or other errors … So the term ‘tiger’… does not
mark out a cluster concept in which most, but not all of the properties used to identify the kind must
be satisfied. On the contrary, possession of most of these properties need not be a necessary condition
for membership of the kind, nor need it be a sufficient condition.
(1980: 121)
What fills the seas and rivers of this planet is not water even if its
inhabitants speak English and call it ‘water’ and despite the fact that any
non-scientist from Earth, if he were to visit Twin Earth, would be unable to
tell that the liquid in its seas and rivers was not water. Moreover, the stuff
on Twin Earth did not belong to the extension of the predicate ‘is water’ as
used on Earth even, say, 700 years ago, when the properties the average
non-scientist associates with that predicate now were the only properties
associated with it by anyone – non-scientist or scientist (alche-mist) alike –
even though, if anyone living on Earth 700 years ago had been carried off
by the Martians to Twin Earth, he would have found the stuff in the seas
and rivers to be exactly as he would have expected water to be.
Again, if Kripke is right about all this, the cluster of identifying marks
theory of natural kind terms is mistaken.
Now Kripke’s view of what is going on in these cases is as follows.
When we introduce a term like ‘water’ or ‘gold’ or ‘tiger’ into the language
we introduce it as a proper name of a kind. For example, ‘tiger’ is first
introduced by a reference-fixing definition of the following type: ‘Tigers
are the kind of fierce, cat-like, yellow and black striped creatures which live
in the jungle over there and occasionally steal and eat our goats and
infants’. This is analogous to fixing the reference of a man’s name by the
definition: ‘Tom Jones is the man presently living in the old house by the
river over there, who is always so rude and unsociable’. Kripke emphasizes
the likeness to the case of proper names of individuals:
In the case of proper names, the reference can be fixed in various ways. In an initial baptism it is
typically fixed by an ostension or description. Otherwise the reference is usually determined by a
chain, passing the name from link to link. The same observations hold for such a general term as
‘gold’. If we imagine a hypothetical baptism of the substance, we must imagine it picked out by some
such definition as ‘gold is the substance instantiated by the items over there, or at any rate, by almost
all of them’. … I believe that in general names for natural kinds … get their reference fixed in this
way, the substance is defined as the kind instantiated by (almost all) of the given sample. The
‘almost’ all qualification allows that some fool’s gold may be present in the sample. If the original
sample has a small number of deviant items, they will be rejected as not really gold. If, on the other
hand, the supposition that there is some uniform substance or kind in the initial sample proves more
radically in error, reactions can vary; sometimes we can declare that there are two kinds of gold,
sometimes we may drop the terms ‘gold’… the original samples get augmented by the discovery of
new items. … More important, the term may be passed from link to link exactly as in the case of
proper names, so that many who have seen little or no gold can still use the term. Their reference is
determined by a causal (historical) link.
(1980: 135–6)
On the Kripkean picture the term ‘gold’ acquired its reference in the way
we imagined above the name ‘Tom Jones’ to acquire its reference, or, to use
Kripke’s own example, in the way ‘Jack the Ripper’ acquired its reference.
‘Jack the Ripper’ was introduced by fixing its reference as: the man who
committed those murders, or at any rate, most of them. If evidence had
emerged that a few of the murders were committed by a different person (a
copycat killer) the reference would have remained the man who committed
most of the murders (or maybe the most gory, or those committed or
discovered first). The victims of the copycat killer would not have been
included in the history books as victims of Jack the Ripper. But reactions
might have varied if honours, as it were, had been divided. Be that as it
may, the name’s original reference was: that man, the man who committed
those murders, whoever he is – and it was understood that it was the job of
the experts (the detectives) to discover the identity. In the same way, ‘gold’,
on Kripke’s picture, was originally introduced just as a name for: that kind
of substance, the kind instantiated by those samples, whatever it is – and it
was understood that its nature was a matter for future discovery. So it was
entirely comprehensible to its first users that a substance absolutely
indistinguishable from gold by all the tests available to them for being gold
might not be gold but a different kind of stuff.
Moreover, since reference-fixing definitions do not give meaning, once
‘Tom Jones’, ‘Jack the Ripper’, ‘the tiger’ and ‘gold’ have been introduced
in this way, it is perfectly sensible to speculate about, say, a possible world
in which not Tom Jones, but someone else is the unsociable old man
presently living in the old house by the river or in which not tigers, but
another kind of creature are the fierce, cat-like, black and yellow striped
animals that steal and eat our goats and infants. And just as in each case
what is being named is not something that can only exist where and when it
is named, it is perfectly sensible to speculate about actual places and times
where it is considerably different from how it is here and now: places and
times where Tom Jones is not rude and unsociable and tigers are not fierce,
black and yellow striped carnivores. Similarly, it is possible to speculate
that there are, or might be, other rude and unsociable old men around or
other kinds of black and yellow striped carnivores.
It is with this kind of comparison of names for natural kinds with proper
names of individuals that Kripke suggests that we can make sense of the
counterexamples which refute the cluster of identifying marks theory of
natural kind terms. To be Tom Jones is to be the same man as the man so
baptized; to be a tiger to be the same kind of thing as the kind so baptized.
Being the same kind of thing is no more a matter of known superficial
appearance than being the same man is. What it is a matter of is somewhat
vague; all the same, according to Kripke it is a notion of which we all have
some intuitive grasp and which is essential to our understanding of natural
kind terms. According to the cluster of identifying marks theory someone
confronted by a superficially tiger-like creature ought to be able, if he
understands the term ‘tiger’, to determine whether it is really a tiger by
examining it. On this theory it makes no sense for him to say, after
examining it, that it might be a tiger but that he remains uncertain. But,
according to Kripke’s account, such a remark is perfectly comprehensible.
Someone for whom the reference of tiger is fixed by the description ‘the
kind of ferocious striped carnivore of which my village lives in terror’ will
know that there could be many kinds of creature of the same superficial
appearance as tigers and so, without a knowledge of the internal structure of
tigers – say, whether they are mammalian or reptilian – ought not to be
prepared to say with certainty of any particular creature of such an
appearance that it must be a tiger. Similarly, someone transported to Twin
Earth and confronted with a sample of XYZ can perfectly intelligibly, on
Kripke’s account, withhold judgement whether the stuff is water even
whilst acknowledging that it has all the properties he knows water on Earth
to possess. For he can point out that the stuff might not be of the same kind
as the samples on Earth despite sharing these properties. Another example.
Suppose a chemist comes across a previously unknown substance
somewhere on Earth which he calls ‘nemonite’; he has just begun to
explore its properties when he is abducted by the Martians and transported
to Twin Earth. He is then confronted by a substance of chemical formula
UVW and asked whether it is nemonite. He carries out the same few tests
on it that he was able to carry out on nemonite on Earth and it behaves in
the same way. Obviously, it would be absurd for him to announce on this
basis that the stuff on Twin Earth is nemonite, for he knows that it might
turn out, on further investigation, that nemonite and the stuff on Twin Earth
are quite different kinds of stuff. Equally obviously, he will no more be able
to say whether the stuff on Twin Earth is nemonite even after he has
become fully familiar with its properties, since however extensive his
knowledge of it, it cannot possibly enable him to say whether it is of the
same kind as the stuff on Earth which he was only beginning to investigate.
Kripke’s account of how natural kind terms function can make sense of
all this; the cluster of identifying marks theory cannot. But it is important to
see what feature of Kripke’s account makes the stories intelligible. It is not
that natural kind terms are construed as rigid designators of kinds. This can
play no role in explaining the intelligibility of the native’s refusal to say
whether the creature he has examined is a tiger, nor the refusal of the
chemist abducted by the Martians to say whether the stuff on distant Twin
Earth is nemonite. This is obviously so, since the distinction between rigid
and non-rigid designators is a matter of how the reference is determined in
counterfactual, not actual, situations. What seems to make these cases
intelligible is the thesis that natural kind terms are names of kinds together
with our knowledge, which is part of our grasp of the concept of a kind, that
super-ficial similarity in external appearance is not a sufficient condition of
sameness of biological kind and that substances which to the inexperienced
eye are very similar can be of distinct material kinds or be of the same kind
despite being superficially very different. (It might be, as Dummett (1973:
143) says, that in the case of animal kinds descent from a common stock is
also important.)
Let us recap. Kripke writes as if his account of how natural kind terms
are introduced is inconsistent with Frege–Russell descriptivism and takes
the examples he gives in support of his position as more grist to his ‘more
Millian than Mill’ account of naming. But although these cases are
problematic for the cluster of identifying marks account they are not
problematic for Frege–Russell descriptivism. The commitments of the
cluster of identifying marks account go far beyond the commitments of
descriptivism. If it were applied to proper names of individuals the cluster
of identifying marks account would amount to the claim that the reference
of a name of an individual is determined solely by qualitative features of the
individual available to inspection by anyone who uses the name. But this is
not entailed by descriptivism. If I introduce a name of a man by such a
description as ‘the tall fat man over there now’, I am not committed to the
claim that there are no circumstances in which that man is short and thin or
no circumstances in which other men are tall and fat. What I intend when I
fix the reference of a name in this way is that it is henceforth used correctly
only if it is used to refer to the same man, and it is part of our understanding
of the notion of the same man that the same man may be superficially
different at different times and places and different men may be
superficially the same. Similarly, as Kripke’s work, along with Putnam’s,
has brought out, the notion of sameness of kind allows that superficially
similar things may be of different kinds and superficially different things of
the same kind. It is for this reason that the cluster of identifying marks
account is mistaken. But it does not follow that descriptivism is incorrect.
An account which is compatible with Kripke’s discussion (even if it is
not compatible with Kripke’s intention) is that when natural kind terms are
introduced they are introduced by descriptions of the form: the kind of thing
of which these (ostensively introduced) samples are instances (as Kripke
says). Subsequently, users intend to refer to what the first users referred to –
they use the term deferentially as descriptivism allows. Kripke does not, in
fact, attempt to sketch out in the case of natural kind terms, as he does in
the case of proper names of individuals, a counterexample to the
descriptivist view. Rather, he writes as if he has already done so in his
account of how natural kind terms are introduced. But this account only
shows that natural kind terms are not equivalent to descriptions free of
names, demonstratives and other indexicals, just as proper names of people
are not. Kripke’s story in fact requires, in the case both of natural kind
terms and proper names of people, that when the term is first introduced an
identifying description must be available to the baptizer, in conformity with
descriptivism. In the case of proper names of people Kripke goes on to
argue (a) that even as used by the initial baptizer whilst he retains the
identifying description in question the proper name is not equivalent to any
description because the name is a rigid designator and the description is not
and (b) that subsequently users of the name will lack any identifying
description of the sort descriptivism requires. In the case of natural kind
terms the corresponding arguments are simply not given, so it is open to the
descriptivist to accept the conclusion Kripke explicitly draws, that the
cluster of identifying marks theory is incorrect, whilst maintaining his
position.
In short, it appears that the main thrust of Kripke’s explicit arguments
about natural kind terms is that they cannot be adequately defined in terms
of indexical-free descriptions known to their users – i.e. clusters of
identifying marks – an important thesis but not one inconsistent with
descriptivism. To drive this point home recall that if Kripke’s attack on the
description theory of proper names, as defined by the theses (1)–(6) he lists,
is successful, then it establishes externalism in the sense defined by Putnam
–‘meanings just ain’t in the head’. But Kripke’s arguments against the
cluster of identifying marks theory of natural kind terms leave internalism
unscathed. Of course, the same is also true of Putnam’s own Twin Earth
arguments. As arguments for externalism they are simply unsound. Even we
accept, with Putnam, that ‘water’ on Earth denotes H2O and ‘water’ on
Twin Earth denotes XYZ, externalism is not established since it is just false
that Oscar and Twin Oscar are narrowly psychologically identical in the
sense of ‘narrow psychology’ Putnam himself defines (and needs, if his
argument is to have the revolutionary implications he perceives). Oscar
believes, say, of himself, that is, of Oscar, that he has a headache. Twin
Oscar does not. This is a narrow psychological state in Putnam’s sense since
Oscar can be in it even if nothing else exists (though no one else can). So
Oscar and Twin Oscar differ in their narrow psychological states (Noonan
1981, 1985). Similarly, Oscar intends to use ‘water’ to speak of the
substance he believes that he, that is, Oscar, bathes in. Twin Oscar does not.
Again, this intention is a narrow psychological state, so again Oscar and
Twin Oscar differ in their narrow psychological states. Hence Putnam’s
argument does not establish that meanings are not, in the sense intended, ‘in
the head’. Of course, it does establish that there is an indexical component
in our concept of water – for me ‘water’ is defined as ‘the liquid stuff I am
in contact with, in the seas and rivers near to me’ (or perhaps I need the past
tense ‘water is the liquid stuff I have been in contact with throughout my
life, in the seas and rivers near to me’ so that when I am kidnapped by the
Martians and transported to Twin Earth I speak falsely when I say ‘this
liquid I am now drinking is water’). But this is not something inconsistent
with internalism, i.e. descriptivism as Kripke understands it, but a form of
it.
This is not to say that Kripke’s alternative to the cluster of identifying
marks account, even when not viewed as a rival to descriptivism, is
unproblematic. The intuition that ‘gold’ or ‘tiger’ might have been
introduced in a primitive community to name a kind that members of the
community could not distinguish, given their limited knowledge, from a
kind instantiated elsewhere is highly plausible. In fact, there are two
plausible intuitions here. The first is that a primitive community in Bengal
might introduce the term ‘tiger’ for the kind of large, dangerous, striped
creature that comes in the night and attacks them, whilst recognizing that
there might well be other kinds of large, dangerous, striped creatures that
they could not distinguish from those they call ‘tigers’– since they have
never caught and killed a tiger or even got close to one. The second is that
the community could employ a notion of sameness of kind whilst
recognizing that their knowledge of and tests for sameness of kind are so
limited that pairs of things they could not at all tell apart even on the closest
examination might be of different kinds, i.e. might not fall in the extension
of the relational predicate they employ which we translate as ‘is of the same
kind as’. However, what seems most readily to make sense of these
intuitions is the thought that the term could have been introduced to name
the narrowest, the most uniform, kind instantiated by the samples, so that
even if the stuff (thinking of the ‘gold’ example) elsewhere (on Twin Earth)
were indistinguishable from that on Earth by all the means available to the
community, they could still make sense of the idea of differences they could
not discern which made the difference. And, indeed, it is only charitable to
assume that some such restriction must be intended by Kripke since kinds
form a hierarchy, so to talk of ‘the kind’ instantiated by the samples, unless
elliptical, is to fail to refer. If Kripke does intend such a restriction it fits
well with our intuitions about XYZ on Twin Earth. However, it does not fit
well with all our intuitions. It is, of course, particularly difficult to see how
the restriction can give the right results in the case of biological kinds, since
every individual organism is different, genetically and in its history, from
every other. Yet the members of a primitive community could presumably,
on Kripke’s story, pick out the kind Tiger by reference to a single tiger: ‘the
kind of thing (animal?) of which that is an example’, or by reference to a
female tiger and its cubs. In the case of chemical kinds the problem is still
present, if less obvious. Gold naturally occurs in one isotope, so any
baptismal sample for ‘gold’ on Earth consists entirely of one isotope, but it
seems that the term ‘gold’ could have been introduced in the way Kripke
sketches by a primitive community without its extension being composed
only of samples of the naturally occurring isotope, as a restriction to the
narrowest kind instantiated by the samples would require. Similarly, if all
our baptismal samples for ‘water’ had happened to be H2O, then, given the
restriction, the term ‘water’ could not have been introduced in the way
suggested, and at the same time have been given an extension including
samples of heavy water. The restriction does not even allow for all of
Kripke’s own intuitions. He remarks that if this substance (H2O) can take
another form – such as the polywater allegedly discovered in the Soviet
Union with very different identifying marks from that of what we now call
water – this is a form of water because it is the same substance (1980: 129).
An alternative way of elaborating, or modifying, Kripke’s proposal,
which might give intuitively acceptable results, would be to propose that
the introducers of a natural kind term might themselves use it deferentially
– deferentially to future users in their community. This is suggested by
Putnam’s statement that ‘the key point is that the relation sameLis a
theoretical relation: whether something is or is not the same liquid as this
may take an indeterminate amount of scientific investigation to determine’
(1973: 702). Putnam’s account, in outline, is that the term ‘water’ gets its
reference fixed as follows: speakers latch on to certain actual samples and
apply the term ‘water’ to them and their use of ‘water’ accords with the
following rule: ‘for all possible x, x is a sample of water if and only if x is
of the same kind as the actual samples’– where the same kind relation
(sameL) istobe determined by scientific theory. The contrast here is with,
e.g., the definition of a ‘paediatrician’ as a doctor specializing in the
treatment of children. A tempting way to understand this proposal and
contrast is that users of ‘water’ use it deferentially to future users because
they fix its reference as ‘stuff of the same kind as these samples’, and they
use the relational expression ‘of the same kind as’ deferentially to future
users in the community. So the communities on Earth and distant Twin
Earth might refer respectively to H2O and XYZ when they use the term
‘water’ even before any chemistry is known, because the introducers of the
term on Earth intend its reference to be that which future users in their
community would pick out as ‘water’– and, by contrast, the users of the
term on Twin Earth intend exactly the same. This proposal at least provides
a response to the prima facie puzzle about how the chemically ignorant
speakers on Earth and Twin Earth, who are unable to distinguish H2O and
XYZ by any means available to them, can grasp a notion of sameness of
kind which distinguishes them, and it is a proposal in accord with Putnam’s
notion of deferential reference, or division of linguistic labour – thought of
as extending across generations. The suggestion is that the supposition that
the chemically ignorant users of ‘water’ on Earth and Twin Earth must
make if they are to be referring to different kinds, is that they have a
possibly imperfect grasp of a notion, i.e. that of sameness of kind, of which
other future users will know more.
This is vague. But the point I wish to emphasize is that it is not merely if
Kripke’s account of natural kind terms is interpreted within a descriptivist
framework that it is necessary to explain how the initial ignorant users can
secure reference to a unique kind; the same is true if Kripke’s account is
read, as he intends, as opposed to descriptivism. Kripke’s proposal that
natural kind terms be understood as analogous to proper names of
individuals requires that the initial baptism contains an identifying
reference to a kind, just as an introduction of a proper name requires an
identifying reference to an individual.
So far I have written as if the intuitions to which Kripke appeals in his
attack on the cluster of identifying marks theory cannot be contested. Of
course, this is not so. If we in fact discovered a Twin Earth on which there
was a substance super-ficially like water with a different chemical
composition, what would we say? What would we say about the extension
of ‘water’ as used by our chemically ignorant ancestors if such a discovery
were now made? Kripke and Putnam write as if it is clear what our
reactions would be. But others are not convinced. Lewis writes:
Like any up-to-date philosopher of 1955, I think ‘water’ is a cluster concept. Among the conditions
in the cluster are: it is liquid, it is colourless, it is odourless, it supports life. But, pace the philosopher
of 1955, there is a lot more to the cluster than that. Another condition in the cluster is: it is a natural
kind. Another is indexical: it is abundant hereabouts. Another is metalinguistic: many call it ‘water’.
When we hear that XYZ off on Twin Earth fits many of the conditions but not all, we are in a state of
semantic indecision about whether it deserves the name ‘water’. … When in a state of semantic
indecision, we are often glad to go either way and accommodate our own usage to the whim of our
conversational partners. … So if some philosopher, call him Schmutnam, invites us to join him in
saying that the water on Twin Earth differs in chemical composition from the water here, we will
happily follow his lead. And if another philosopher, Putnam … invites us to say that the stuff on
Twin Earth is not water … we will just as happily follow his lead. We should have followed Putnam’s
lead only for the duration of that conversation, then lapsed back into our accommodating state of
indecision. But, sad to say, we thought that instead of playing along with a whim, we were settling a
question once and for all.
(1994: 424)
If Lewis is right Twin Earth cases pose no threat to internalism or
descriptivism (though the mere fact that they are unclear suffices to refute
the cluster of identifying marks theory of the 1950s philosopher). But, as
we have seen, they pose no threat to these positions anyway. It is an
empirical question what our linguistic dispositions are, but whether they
incline us to withhold ‘water’ from XYZ, or apply it, or to dither, the
sophisticated Frege– Russell internalist descriptivist Kripke intends to
refute in Naming and Necessity need not worry.
The Necessity of Theoretical Identifications
As already noted, the general form of argument by which Kripke argues for
the necessary truth of ‘Hesperus is Phosphorus’ is a special case of this
reasoning. I discover empirically, what I can only know a posteriori, that
Hesperus is Phosphorus. I acquire by philosophical analysis the knowledge
that if Hesperus is Phosphorus it is a necessary truth that it is and I infer that
it is a necessary truth. The knowledge that, if Hesperus is Phosphorus, it is a
necessary truth that it is results from universal instantiation of the principle
of the necessity of identity. Thus knowledge of a general essentialist
principle is required for knowledge of a particular necessary a posteriori
truth of identity that it is necessary.
Entirely in parallel, Kripke thinks, if being made of molecules is an
essential property of Kripke I can come to know that it is a necessary truth
that Kripke, if he exists, is made of molecules by ascertaining a posteriori
that he is, and hence that if he exists, he is – something I can only know a
posteriori, inferring the conclusion by modus ponens using the conditional,
which I can know a priori to be true, that if Kripke is made of molecules if
he exists it is a necessary truth that he is if he exists. And I can know this
conditional about Kripke in particular to be true a priori, if it is, because I
can know a priori that, say, any material object made up of molecules is
essentially so. Thus, knowledge of a general essentialist principle about
individuals is required for knowledge of a particular necessary a posteriori
property ascription that it is necessary. The route to the conclusion involves
universal instantiation followed by modus ponens. The indispensability of
universal instantiation makes it clear that what primarily comes to be
known is the possession of an essential property by an object. The
reasoning remains valid if ‘Kripke’ is replaced by ‘the author of NN’, but
the conclusion then inferred would not then be transformable into ‘it is a
necessary truth that if he exists the author of NN is made of molecules’.
Similarly, in the case in which I fix the reference of ‘N’ by ‘the number of
planets’, I can infer that it is a necessary truth that N is odd, appealing to the
general principle that if any number is odd it is necessarily odd. Thus the
inference remains valid if ‘N’ is replaced by ‘the number of planets’, but in
this case we cannot arrive at the conclusion that it is a necessary truth that
the number of planets is odd.
In the same way, Kripke thinks that I can come to know that all tigers, i.e.
all instances of the kind Tiger, are mammals. I can know by a posteriori
investigation that all actual instances of the kind are mammals. And I can
know by philosophical analysis that if all instances of the kind are
mammals the kind is such that necessarily all instances of it are mammals.
Thus I can infer that the kind Tiger is such that necessarily all instances of it
are mammals. Given that ‘the kind Tiger’ is a rigid designator this can be
transformed into ‘It is a necessary truth that all instances of the kind Tiger
are mammals’. Given that ‘is a tiger’ means ‘is an instance of the kind
Tiger’, it follows that it is a necessary truth that all tigers are mammals. But
I can know by philosophical analysis that if all instances of the kind Tiger
are mammals then the kind Tiger is necessarily such that all instances of it
are mammals only because I can know by philosophical analysis, for
example, the general truth that any biological kind is such that if every
instance of it is a mammal that kind is necessarily such that every instance
of it is a mammal. Again, this is a general essentialist principle without
which my reasoning cannot proceed, but this time it is one about kinds.
Consequently, just as in the case of knowledge of necessary a posteriori
truths about individuals, the knowledge of necessary a posteriori truths
about kinds in which Kripke is interested can be achieved via the
mechanism he suggests only if we have a priori knowledge of other general
essentialist principles, in addition to a priori knowledge of the general
principle of the necessity of identity. But Kripke never argues in detail, or
indeed at all, except in the case of the necessity of individual origin, for any
such general essentialist principles either governing individuals or kinds.
He simply appeals to our intuitions.
Nevertheless, in the case of kinds, it is at least suggested by the text how
an argument for the general essentialist principles Kripke employs might be
constructed. At one point Kripke recalls the example of the standard metre:
The definition [‘gold is the substance instantiated by the items over there, or at any rate, by almost all
of them’] … express[es] an a priori truth, in the same sense as (and with the same qualifications as)
‘one meter = the length of S’:it fixes a reference. I believe that, in general, terms for natural kind …
get their reference fixed in this way …
(1980: 135–6)
In all these cases, then, just as in the case of the standard metre, if Kripke is
right, there are not only necessary a posteriori truths, but contingent a priori
truths, about the kind.
But what are the necessary a posteriori truths about one metre,
comparable to ‘heat is molecular motion’ or ‘the tiger is a mammal’? It
appears plausible that one such necessary a posteriori truth is ‘A metre is a
greater length than a yard’ (Kripke (1980: 76) suggests that the reference-
fixing description for ‘one yard’ is ‘the distance when the arm of King
Henry I of England was outstretched from the tip of his finger to his nose’).
Given that I have fixed the reference of ‘one metre’ as the length of stick S
at t I can certainly not know this without knowing the length of stick S at t,
which I can do only by measuring it. But once I have measured stick S, I
can infer that it is a necessary truth that one metre is a greater length than a
yard, and thus that anything, in any possible circumstance, which is one
metre is longer than anything, in that or any other circumstance, which is a
yard. It appears that this is analogous to the reasoning by which Kripke
thinks that I can come to know, but only a posteriori, that it is a necessary
truth that the Tiger is a mammal, and thus that necessarily all tigers are
mammals. I can come to know by empirical investigation that the instance
of the kind (tiger Tim, say) by which I have fixed the reference of the kind
term ‘the Tiger’ is a mammal. Consequently I can infer that it is a necessary
truth that the Tiger is a mammalian kind and thus that any creature in any
possible circumstance which is a tiger is a mammal.
My inference that it is a necessary truth that one metre is a greater length
than a yard appeals to two propositions which I know a priori. First, that
any possible or actual things that are the same in length are such that if
either is greater than a yard in length the other is also. That is, that if
something in one possible circumstance is the same in length as something
(else) in the same or a different possible circumstance then, if the first thing,
in the first circumstance, is longer than a yard, the same is true of the
second in the second circumstance. This proposition cannot be stated
without quantifying over merely possible things or making reference to
possible circumstances. It has a counter-factual content greater than the
content of the claim that in fact things that are the same in length are such
that if one is greater than a yard in length the other is also, and its content is
not captured by the modal claim that it is necessarily true that if things are
the same in length and one is longer than a yard then the second is also.
Nevertheless, it does appear to be something knowable a priori, because
analytic, by anyone who grasps the concept of sameness of length and the
Imperial system of length measurement. The second proposition my
inference rests on is that two actual or possible concrete objects are the
same in length just in case the length of the first is the length of the second.
This is an instance of an abstraction principle, comparable to the criterion of
identity for numbers Frege gives: the number of Fs is the number of Gs if
and only if there is a one–one correspondence between the Fs and the Gs, or
the principle by which Frege introduces the concept of a direction in order
to provide a simpler analogue of this: the direction of line a is the direction
of line b if and only if a is parallel to b (Grundlagen 1968: 74). It introduces
discourse about lengths conceived of as abstract objects. And, like Frege’s
criterion of identity for numbers and the direction principle, on its natural
understanding it too, despite having counterfactual content, is knowable a
priori. Putting these propositions together I can infer that the abstract object,
the length one metre, is such that if some concrete object (a fortiori, every
concrete object) the length of which it is is longer than a yard, then that
length is such that necessarily every concrete object of which it is the length
is longer than a yard. This is an instance of the general essentialist principle
about lengths that any length is such that if some concrete object (a fortiori
every concrete object) the length of which it is is longer than a yard, then
that length is such that necessarily, every concrete object of which it is the
length is longer than a yard. From this, given that ‘one metre’ is a rigid
designator of the length of stick S at t, and that stick S at t is longer than a
yard, one can infer that it is a necessary truth that any concrete object which
has a length of one metre is longer than a yard, which is a necessary a
posteriori truth about an abstract object, a length, on a par with Kripke’s
proposed necessary a posteriori truth about the natural kind, the Tiger, that
everything which belongs to it is a mammal.
The mechanism by which such necessary a posteriori truths about
abstract objects can be generated can be illustrated in another example
(Donnellan’s, given by Salmon (2005: 165)).
I understand what it is for two people to be of-the-same-marital-kind,
where this is defined as meaning that, i.e. as an abbreviation for the claim
that, they are both bachelors, both spinsters, both married men, both
married women, both widows, or both widowers. Given this definition I
know that if two people are of-the-same-marital-kind and one is unmarried
the other is also (just as I know that if two things are the same in length and
one is longer than a yard, the other is also). And I know not only that any
actual people who are so related are such that if either is unmarried, both
are, but that any actual or possible such people are such that if either is
unmarried both are, that is that if anyone in any actual or possible
circumstance were to be of-the-same-marital-kind-as anyone else in any
other actual or possible circumstance is or was then the former would be
unmarried if the latter is or was. This is trivially so, because it is analytic
that if two people are both bachelors or both spinsters or … (spelling out
the full definition of same-marital-kind), both are unmarried if either is. So
far, however, no discourse about marital kinds as objects of reference has
been justified. I have simply defined by abbreviative definition ‘x is-of-the-
same-marital-kind-as y’ to mean ‘x and y are both bachelors, both spinsters,
both unmarried men, both unmarried women, both widows or both
widowers’. However, having given this definition I can now introduce talk
of marital kinds as objects by a Fregean definition by abstraction: the
marital kind to which x belongs is the marital kind to which y belongs if
and only if x is-of-the-same-marital-kind-as y. On its natural understanding
this definition has counterfactual force. If I then fix the reference of a
proper name, ‘B’, to be the marital kind to which my (reclusive) neighbour,
Hilary, belongs, then when I make the a posteriori discovery that he is a
bachelor, I can infer, by the reasoning sketched in the example of one metre,
that it is necessary that the marital kind B is such that anyone belonging to
it is unmarried. If I now define the predicate ‘x is a squatchelor’ to mean
‘the marital kind to which x belongs is identical with the marital kind B’, I
can infer that it is a necessary truth that anyone who is a squatchelor is
unmarried; something I can know only a posteriori.
The reasoning by which Kripke suggests that one can infer that it is a
necessary truth that all tigers are mammals, or that it is a necessary truth
that all samples of water contain atoms of oxygen and hydrogen, appears to
be entirely similar. The difference is that it is transparent in the cases just
considered how I can know a priori the crucial relevant proposition linking
sameness of kind with the sharing of the relevant property. It is simply
analytic, given how sameness of marital kind is defined, that if two actual
or possible people are of-the-same-marital-kind then both are unmarried if
either is, and it is, if not analytic, an obvious conceptual truth that if two
actual or possible things are the same in length both are longer than a yard
if either is (just as, in the Fregean examples, it is an obvious conceptual
truth that if two lines are parallel and one lies at 90 degrees to a third so
does the second, and an obvious conceptual truth that if the Fs are one– one
correlated with the Gs then if there are more than three Fs there are more
than three Gs). But the proposition that, if two actual or possible things are
of the same natural kind, both are mammals if either is, is not similarly
transparently knowable a priori. Nor is this true of the proposition that if
two actual or possible samples are samples of the same natural kind of stuff
both are compounds of hydrogen and oxygen if either is. As we have seen
Kripke emphasizes that the notion of sameness of natural kind he is
operating with is a primitive one, available to biologically and chemically
uninformed communities of speakers. Thus it is left unexplained how such
propositions as these can be known a priori. Nevertheless, it does appear
likely that the best account of the mechanism by which Kripke thinks
necessary a posteriori truths about natural kinds can be generated is the one
sketched above, in which such propositions play a crucial role. Whether this
means that Kripke is presupposing knowledge of essentialist principles
other than those he explicitly relies on is moot (see Salmon 1985: ch. 6). Is
the proposition that if two actual or possible things are of the same natural
biological kind either is a mammal if the other is itself an essentialist
principle? The argument for saying so is that it entails (together with the
definition by abstraction by which reference to natural biological kinds as
abstract objects can be introduced) that any natural kind is such that if
something which is a member of that kind is a mammal it is a necessary
truth that anything which is a member of it is a mammal, which is a general
essentialist principle about natural kinds.
But on the same basis it can be argued that the proposition that if two
actual or possible things are the same in length either is longer than a yard if
the other is is itself an essentialist principle on the ground that, together
with the abstraction principle for lengths, it entails that every length is such
that if something of which it is the length is more than a yard long it is
necessarily true that everything of which it is the length is more than a yard
long. Or again, that the proposition that if two actual or possible people are
of-the-same-marital-kind either is male if the other is is itself an essentialist
principle on the ground that, together with the abstraction principle for
marital kinds, it entails that every marital kind is such that if someone of
whom it is the marital kind is male it is necessarily true that everyone of
whom it is the marital kind is male.
So perhaps Kripke cannot be charged with presupposing any general
essentialist principles additional to those to which he explicitly appeals.
Nevertheless, it is important to see that the general essentialist principles
about kinds conceived as abstract objects on which he does rely can only be
derived, in the manner sketched above, from the a priori knowable general
propositions about individuals, by appeal to the abstraction principles by
which talk of the kinds as abstract objects is introduced. And these, on the
understanding of them which allows the derivations, are not expressible in
the language of quantified modal logic. The abstraction principle for marital
kinds, for example, cannot be understood merely as saying that it is a
necessary truth that for all persons x and y, the marital kind to which x
belongs is identical with the marital kind to which y belongs if and only if x
is of-the-same-marital-kind-as y. To allow the derivation of the general
essentialist principle that any marital kind is such that if any person
belonging to that marital kind is unmarried necessarily any person
belonging to that marital kind is unmarried, its content has to be that for all
possible worlds W and W* and possible persons x and y the marital kind to
which x belongs in W is identical with the marital kind to which y belongs
in W* if and only if x in W is of-the-same-marital-kind-as y in W* –
something that cannot be expressed in the language of quantified modal
logic, but requires for its expression an explicit quantification over possible
worlds (Salmon 1985). (Alternatively, assuming Lewis’s modal realism and
the world-boundness of possibilia, its content can be captured without
quantification over possible worlds, in the statement: for all possible
persons x and y, the marital kind to which x belongs is the marital kind to
which y belongs if and only if x is-of-the-same-marital-kind-as y. The
content of the proposition that necessarily all squatchelors are unmarried is
thus that all possibilia of-the-same-marital-kind-as reclusive Hilary are
unmarried. But this, from the modal realist viewpoint, is not a de re modal
statement about an abstract object at all (it does not say anything about all
or some counterparts of such an abstract object), but a de re modal
statement about Hilary, to the effect that he is essentially unmarried (taking
being-of-the-same-marital-kind-as as the relevant counterpart relation).
Similarly, from the modal realist viewpoint, the proposition that it is
necessarily true that all tigers are mammals is equivalent to the proposition
that all possibilia of the same natural kind as paradigm tiger Tim are
mammals, which is not a de re statement about a kind, the Tiger, but a de re
modal statement about Tim, to the effect that, under a certain counterpart
relation, he is essentially a mammal. It is therefore, from this viewpoint,
entirely unsurprising that it can only be known a posteriori that all tigers are
mammals, but the fact that it can only be so known does not establish that
any proposition is a necessary a posteriori truth, since the proposition that
necessarily all tigers are mammals, i.e. that Tim is essentially a mammal, is
not equivalent to any purely de dicto modal truth.)
The Illusion of Contingency and Mind–Brain
Identity
Kripke’s response to the puzzle is the denial of the first premiss of this
argument. Another response would be to deny the second premiss. Kripke
does not consider this alternative, but as we shall see, it would undermine
his argument against mind–brain identity theories.
According to Kripke the necessary a posteriori status, and hence the
appearance of contingency, of his necessary a posteriori truths is to be
explained by their association with contingent a posteriori truths which
must be known if they are to be known. He illustrates with the case of the
table in front of him, which is known to be wooden only a posteriori but, he
thinks, is essentially (originally) wooden. In this case he suggests that we
have the intuition that the table could have turned out to be made of ice. But
what this means is just that he could have been in qualitatively the same
situation with just the same evidence even though an ice table was present.
So although the statement ‘this table is made of wood if it exists’ states a
necessary truth, there is an associated contingency, ‘the table in front of me
is made of wood’ in which the actual wooden table is picked out by a
contingent property. So, according to Kripke:
Any necessary truth could not have turned out otherwise. In the case of some necessary a posteriori
truths, however, we can say that under appropriate qualitatively identical evidential situations, an
appropriate corresponding statement may have been false. The loose and inaccurate statement that
gold might have turned out to be a compound should be replaced by the statement that it is logically
possible that there should have been a compound with all the properties originally known to hold of
gold.
(1980: 143)
In the case of identities using rigid designators he says, ‘there is a simpler
paradigm. … The references of [the rigid designators “R1” and “R2”] may
well be fixed by non-rigid designators “D1” and “D2”… Then …“D1=D2”
may well be contingent and that is often what leads to the erroneous view
that “R1=R2” might have turned out otherwise’ (1980: 143–4).
The reason we are thus led into error, into an illusion of contingency,
Kripke thinks, is that there are two senses in which ‘could have turned out
otherwise’ can be used. In one sense ‘it could have turned out that P’ entails
that it could have been the case that P, in the sense that it is metaphysically
possible that P. In another, epistemic, sense it does not. When I say, in this
second sense, that it could have turned out that Hesperus was distinct from
Phosphorus what I say could be accurately expressed by saying that it could
have turned out in the first sense that the body visible in the evening was
distinct from the body visible in the morning, which does entail that the
body visible in the evening could have been distinct from the body visible
in the morning in the sense of a metaphysical possibility. Confusing the two
senses, I am disposed, Kripke thinks, to infer that Hesperus could have been
distinct from Phosphorus, in the sense of metaphysical possibility. So I am
disposed to be led into error because I can only come to know that Hesperus
is Phosphorus, a necessary a posteriori truth, via knowing the associated
contingent truth (which is why the necessary a posteriori truth is a
posteriori), and I am inclined to confuse two senses (epistemic and
metaphysical) of ‘it could have turned out that P’.
At the end of Naming and Necessity Kripke brings these considerations
to bear against the mind–brain identity theory. He distinguishes three
different identifications: of a person with his body, of a particular token
sensation (or the event or state of having that sensation) with a particular
token brain state, and of types of mental event with types of physical event.
His aim in each case is to show at least that the only option for the identity
theorist is to embrace a necessitarian version of his position: to say that the
identity is both true and necessary since contingent identity is a confusion.
Thus he argues in the first case that someone who accepts, with
Descartes, that he could exist without his body must also accept that he is in
fact distinct from his body since, if ‘B’ is a rigid designator of my body,
because ‘I’ is also a rigid designator, ‘Iam B’ must be necessary if true at
all. He notes that a precisely parallel argument can be used to conclude,
from the premiss that a statue and the all-times-coincident hunk of matter
from which it is composed could exist apart, that they are not in fact
identical. Many do accept the parallel argument, as Kripke himself does;
others, as we have seen, follow Lewis in embracing the inconstancy of
modal predication, which allows them to accept the premiss in either case,
and reject the conclusion. At any rate Kripke does not think the argument
establishes Cartesian dualism (the identity of the self with a Cartesian
immaterial soul), with which, as he notes, his own thesis of the essentiality
of material origin is inconsistent. The anti-materialist arguments Kripke
takes seriously are the arguments against the token–token identity theory
and the type–type identity theory.
The argument against the token–token identity theory runs as follows.
Let ‘A’ name a certain particular sensation of pain and ‘B’ name a particular
brain state occurring in a man Jones. If ‘A = B’ is true it is necessarily true.
But it is possible for A to exist without B and B without A. So it is not
necessarily true, hence it is not true. Since ‘A’ and ‘B’ are rigid designators
the point of contention can only be the premiss that A could exist without B
and B without A. Kripke’s argument that B could exist without A is that B
is only contingently a pain if it is a pain at all, whereas A is necessarily a
pain: ‘it is at least logically possible that B should have existed without
Jones feeling any pain at all’ (1980: 146); on the other hand, ‘can any case
of essence be more obvious than the fact that being a pain is a necessary
property of each pain?’ (1980: 146). His argument that A could exist
without B is that A is only accidentally a brain state if at all, but B is
essentially a brain state: ‘the pain could have existed without the
corresponding brain state [but] being a brain state is evidently an essential
property of B’. Thus A and B differ in their essential properties and are
distinct.
The token–token identity theorist can respond by denying that A is
essentially a pain. He can say that ‘being a pain, as a property of a physical
state, is to be analyzed in terms of the “causal role” of the state, in terms of
the characteristic stimuli which cause it and the characteristic behaviour it
causes’ and that the causal role of a state is a contingent feature of it (1980:
147). Against this Kripke says only that it strikes him as self-evidently
absurd, since it amounts to the view that this very pain I now have could
have existed without being a mental state at all.
Whether or not this is right, it is clear that Kripke’s argument against the
token–token identity theory needs more than the supposition that ‘A’ and
‘B’ are rigid designators: it also needs the essentialist premiss that pain
tokens are necessarily pains.
In fact, it requires more, because the token–token identity theorist might
allow this but still resist the argument by denying that it is possible for B to
exist without pain. He might say that it is an illusion that the token event B
could exist without pain being felt, just as it is an illusion that the particular
wooden table Kripke uses as an example could have existed without being
made of wood. According to Kripke a table made of ice rather than wood
might have existed and looked exactly the same as the wooden one and
been exactly where the wooden one was, but it would not have been the
wooden one but merely an epistemic counterpart of it. Similarly, the token–
token identity theorist might say, a brain state token could have existed
which looked exactly like B (when Jones’s brain was examined), but was
not a pain. But it would not have been the very token brain state B, which is
essentially a pain. So the token–token identity theorist could accept that all
actual brain state tokens which are pains are essentially so, and explain the
appearance of contingency by appeal to the possibility of brain states which
are not pains but can serve as epistemic qualitative counterparts of the
actual brain states which are pains.
Against this Kripke can only appeal to an (anti-) essentialist intuition: ‘it
would seem logically possible that B could exist without any sensation with
which it might plausibly be identified’ (1980: 146).
However, even if the token–token identity theorist can defend himself
against Kripke’s argument that B might have existed without A, either by
insisting that A is accidentally a pain or by insisting that B is essentially a
pain, he also needs to respond to Kripke’s argument that A might have
existed without B, since A is only accidentally a brain state and B is
essentially a brain state. But, for the token–token identity theorist who says
that A is only accidentally a pain (since it only contingently occupies the
causal role of pain) it is not a great stretch to say that what A essentially is
is a brain state (just as on Kripke’s account a token of heat which causes a
sensation of heat in a perceiver is only accidentally a cause of a heat
sensation, but essentially an occurrence of molecular motion). Finally, to
complete the catalogue of possibilities, the token–token identity theorist
might accept that A is only accidentally a brain state, but say exactly the
same of B. B, he might say, qualifies as a brain state because of its
neurological and physiological role, but it is therefore not essentially a brain
state and is in fact essentially a pain. This is a non-materialist version of the
token–token identity theory.
Thus the token–token identity theorist can resist Kripke’s argument by
rejecting one or more of his intuitions about the (differing) essential and
accidental properties of A and B. But another possibility for him, as in the
case of mind–body identity, is to accept Kripke’s intuitions but deny the
inference to non-identity by rejecting Kripke’s assumption that modal
predication is constant in denotation. Thus, he can say that A is essentially a
pain and B is not and that A is B just as Goliath is essentially a statue and
Lumpl is not though Goliath is Lumpl.
Kripke’s most developed argument in the final pages of Naming and
Necessity is against the type–type identity theorist, who holds not only that
each mental state token is identical with some physical state token, but also
that any two token events of the same psychological type are events of the
same physical type, e.g. that pain is the firing of C-fibres. So type–type
identity theory is put forward as precisely analogous to such theoretical
identifications as that of water with H2O and heat with molecular motion.
Kripke’s argument is that the analogy does not obtain. The argument has
two parts. First, he argues that in the case of pain and C-fibre stimulation, as
in the other cases, the identity, if it holds, must hold necessarily since ‘pain’
and ‘C-fibre stimulation’ are rigid designators. But second, he argues that
the a posteriori status and thus apparent contingency of ‘pain is C-fibre
stimulation’ cannot be explained in the way that the a posteriori status and
apparent contingency of ‘heat is molecular motion’ can be explained. The
reason is that we do not fix the reference of the kind term ‘pain’ by a
flexible designator in the way in which we fix the reference of ‘heat’ by a
flexible designator and so the a posteriori status of the identity statement
cannot be explained away by appeal to an associated genuine contingency.
(Note that Kripke’s challenge to the identity theorist is not merely to
explain the epistemic possibility of the falsehood of ‘pain is C-fibre
stimulation’, but to explain its a posteriori status. Unobvious unproved
mathematical truths share with ‘pain is C-fibre stimulation’ the property
that their denials are epistemically possible, but they are not a posteriori.)
One way the type–type theorist can resist this argument is simply by
denying that ‘pain’ is a rigid designator. This is what Lewis does, defending
the standard functionalist position:
If the state of having neurons hooked up in a certain way and firing in a certain pattern is the state
properly apt for causing and being caused, as we materialists think, then that neural state is pain. But
the concept of pain is not the concept of that neural state. The concept of pain, unlike the concept of
the neural state which is in fact pain, would have applied to some different state if the relevant causal
relations had been different. Pain might not have been pain.
(Lewis 1983: 124)
a posteriori knowledge (known on the basis of experience) 2, 39, 220, 221; contingency illusion and
mind-brain identity 210, 211; Naming and Necessity lectures, main contentions 14, 15, 16;
necessary 117, 158–80; philosophical development of Kripke 9–10; theoretical identifications,
necessity of 197, 201, 203
a priori knowledge (knowability independent of experience) 2, 3, 4, 9, 14, 201, 210, 221; contingent
117, 158–80; naming 56, 64, 103
‘A Puzzle about Belief’ (Kripke) 10–11, 166, 170, 178
abstract object/concrete object 5, 131, 221
abstraction principle 204, 221
absurdity, and truth 60
accidental designators 65–66; see also rigid designation
accidental properties of individuals 17, 81, 88, 135; versus essential, 4, 10, 50, 89, 137
actual world 9
Adam (Old Testament) 136, 137, 140, 141
ambiguity 60, 61, 68, 70, 75–76
analogous principle 80
analytic philosophy 2–5, 39, 221
analytic truth 4, 124, 221
Aristotle/Aristotle example 3, 13, 24–25, 37, 64, 83, 84, 96, 97; essentialism, 4, 40, 124, 125, 131
assertoric content 67, 75, 77, 222
attributive use 10, 222
baptism 36, 119, 193; initial 8, 103, 104, 117, 118, 186, 195
Barrett, R. 52
Bible 56, 57
British Empiricism 3, 5
Burge, T. 67, 89, 90
Burgess, J. 98
Carnap, R. 2
Cartesian dualism 213
causal-historical theory of reference 10, 11, 12, 36
C-fibre stimulation 209, 216, 217, 219, 220
Chalmers, D. 94–95
Chisholm, R. 136, 137
Cicero example 11, 84, 85, 90, 91, 93, 103, 104, 108–11, 170, 171, 178
circularity constraint 56, 57–58, 111–12, 114
cluster of identifying marks theory 20, 183–86, 188–92, 195, 196
cluster theory 37, 56, 58, 63, 222; arguments against 102–16; counterexamples 110; and reference-
fixing 102–16
co-denoting descriptions 83
cognitive value 23, 222
compositionality principle 21, 22, 222
concrete objects 5, 131, 204–5, 221
connotation 15, 222
context-dependence 26–27, 133
context-relativity, of essence 128, 134
contingency, illusion of 209–20
‘The Contingent A Priori and Rigid Designators’ (Donnellan) 72
contingent properties 17
coreferential singular terms 25
counterfactual situation 9
counterpart theory 157, 222; critique 141–43
Davies, M. 95
de dicto (of the saying) 4, 222–23; propositional attitude 49, 51
de facto rigidity 62, 63
de jure rigidity 62, 63, 94, 223
de re (of the thing) 222–23; de re modality 4, 40, 42, 46, 50, 53, 123, 222; opacity 128; predications
130; propositional attitudes 49, 50–51, 53
deferential reference 57
demonstrative elements 38, 223
denotation 15, 223
descriptions 6, 24, 29, 83; abbreviations for 72–73; cluster theory see cluster theory; Frege–Russell
theory 6, 10–11, 13, 14, 18, 19, 36, 119, 190; indicative versus subjunctive 67–68, 98–99, 101;
and names 27–36, 68; primary and secondary occurrences 32, 130, 225
descriptivism 15, 94, 190, 191, 192, 195, 196; rigidified 95, 98, 114
direct quotation 26
direct reference theorists 14–15, 15
distinctness pairs 156
Donnellan, K. 6, 10, 11, 38, 106, 175, 176, 222; ‘The Contingent A Priori and Rigid Designators,’ 72
Dummett, M. 67, 68–78, 80, 90, 100, 115, 121, 144, 145, 159, 171, 172, 189
empiricism 2, 3, 223
empty singular terms 28, 30
epistemic necessity 179, 223
epistemological arguments 11
epistemology 34
error 8, 11, 13, 19
essence 128, 134, 223
essence-revealing terms 47, 52, 87, 88, 89
essential properties of individuals 135, 143–58, 223; versus accidental 4, 10, 50, 89, 137
essentialism 17, 39, 136, 148, 155, 157, 200, 207, 223; Aristotelian 4, 40, 124, 125, 131;
intelligibility of 123–35
Evans, G. 10, 66, 95, 120, 121, 122
exportation 49, 50, 53, 86, 87, 223
extensionalism/extensionality 40–41, 43, 44
extension/intension 6, 7, 181–220, 223–24
externalism/internalism 6, 8, 11, 224
Fitch, G. W. 59
flexible (non-rigid) designators 64, 73–74, 118, 131, 189
Franklin, B/Franklin as inventor of bifocals example 60, 61, 65, 66, 67, 69, 70, 74, 78, 79, 80, 87, 88,
90, 95
Frege, G. 2, 3, 20, 85, 88, 226;
Frege–Russell description theory 6, 10–11, 13, 14, 18, 19, 28, 36, 119, 190; Frege–Russell logistic
programme (mathematics) 3; Frege–Russell synthesizers 37–39; Frege–Russell theory of proper
names 8; ‘King of France is bald’ example 29, 30, 31; and Quine 46; on sense 5, 21–27; ‘On
Sense and Reference’ 22, 24
From Stimulus to Science (Quine) 52
functionalism 216
functionality principle 21
Geach, P. 86
Gibbard, A. 128, 131 ‘
Giorgione’/‘Barbarelli’ example (Quine) 127–28
Goliath see Lumpl (lump of clay) and Goliath example
Great Western Railway (GWR) example 129–30 Grundgesetze (Frege) 3 Grundlagen (Frege) 3, 204
Hale, R. V. 89, 90
heat 196–98, 203, 209, 215–19, 220
‘here’ and ‘where I am’ distinction 76–77
‘Hesperus is Phosphorus’ example 14, 15, 160, 162–64, 166, 167, 179, 180, 182, 197–200, 210, 212
historical chain picture 116–202
Hobbes, T. 149, 150, 153, 154
Hughes, C. 177, 219
Humberstone, I. L. 95
Hume, D. 3
identity: identity-revealing aspects 87; mind-brain 18, 209–20; ‘misleading metaphor’ of 59; names
of individuals 14, 15; necessity of 225; puzzle of 26, 31, 42; theoretical identifications, necessity
of 196–209; token-token identity theory 213, 214, 215, 226; transworld see transworld identity;
type-type identity theory 213, 216, 220, 226
‘Identity and Necessity’ (Kripke) 9, 10, 141–42, 156, 160, 196–97
identity statements 9, 24, 26, 27, 196
ignorance 8, 11, 13, 19
incomplete symbols 35
indexical expressions 13, 38, 54, 176, 224
indicative descriptions 67–68, 98–99, 101
Individuals (Strawson) 6, 38
ingredient sense 67, 75, 77, 222
initial baptism 8, 103, 104, 117, 118, 186, 195
intension 47
internalism 8, 11, 192, 224
intuition 126
island names 120, 121
‘Jack the Ripper’ example 66, 70–73, 102, 110, 171, 187
Jackson, F. 94
John Locke lectures ‘Reference and Existence’ 1973, 16
Jonah (Old Testament) 109
Julius (inventor of the zip) 66, 171
Kafka, F. 144
Kaplan, D. 50, 53, 87, 94
knowledge: logical construction of 2–3; a posteriori see a posteriori knowledge (known on the basis
of experience); a priori see a priori knowledge (knowability independent of experience)
Kripke, S.: ‘Identity and Necessity’ 9, 10, 141–42, 156, 160, 196–97; philosophical development 8–
11; ‘A Puzzle about Belief’ 10–11, 166, 170, 178; ‘Speaker’s Reference and Semantic Reference’
10, 107; see also Naming and Necessity lectures, 1970
Kroon, F. 114
naming: ambiguity 60, 61, 68, 70, 75–76; descriptions and names 27–36, 68; fiction versus
substantially false account of real person 56; Franklin as inventor of bifocals example 60, 61, 65,
66, 67, 69, 70, 74, 78, 79, 80, 87, 88, 90, 95; ‘here’ and ‘where I am’ distinction 76–77; modal
argument 65–102; names as rigid designators or abbreviations for descriptions 72–73; rigid
designation 58–65, 223; target 54–58; ‘Thatcherabouts’ example 77, 78, 79–80, 82, 96; thesis 58,
68; vivid names 50, 52, 87; see also proper names
Naming and Necessity lectures 1970, 1, 2, 6, 8, 9, 164; Addenda 16, 17, 18; Appendix 200; main
contentions 11–20; Preface 15, 59, 62, 67, 73, 77–78, 80, 91, 100, 140, 155; supplementary
publications following 10; see also Kripke, S.
narrow psychological state 7, 8, 224
natural and biological sciences 2, 3
natural kind terms 2, 12–13, 224–25; as proper names of kinds 181–96; as rigid designators 13;
subject–predicate sentences 15
natural phenomena 181
‘nec,’ as sentence operator 42, 44, 45, 48
necessity 17, 39, 40, 123–35, 223, 225; and analyticity 3, 4; metaphysical 2, 224; of origin 139, 151,
225; of theoretical identifications 196–209
Nelson, M. 80
Neptune/Uranus example 72–73, 176, 177
Nixon, R./Nixon example 126, 137, 139–40, 142
Noah (Old Testament) 136–37, 140
non-connotative terms 222
non-extensionality 43, 45
non-modal sentences/statements 68, 74–75
non-trivial sense 4, 123, 126, 127, 134, 141, 143, 217; essential and accidental properties, non-trivial
distinction 49–50; and Quine 39, 45, 46, 49–50, 126
Noonan, H. 147, 153
reference 1, 4, 11–12; causal-historical theory 10, 11, 12, 36; dependence on context 26–27, 133;
sense distinguished 5, 21, 25; see also sense/reference
‘Reference and Definite Descriptions’ (Donnellan) 10
reference shift 120, 121
reference-determination 115
reference-fixing 58, 63, 94, 102–16, 117, 118, 186; and cluster theory 102–16; versus synonym-
giving 64; see also cluster theory
referential opacity/transparency 39–40, 43, 225
referential use 10, 222
rigid designation 17, 53, 212, 223, 226; concept of rigidity 16, 18; naming 58–65, 89–90, 93;
necessity 131, 157, 162; and non-rigid (flexible) designators 64, 73–74, 118, 131, 189; proper
names as rigid designators 12, 13, 15, 20, 67, 75; strongly rigid designators 61–62; see also de
facto rigidity; de jure rigidity
rigidified descriptivism 95, 98, 114
Russell, B. 5, 20, 27–36, 101, 130;‘On Denoting’ 3, 28; Frege–Russell description theory 6, 10–11,
13, 14, 18, 19, 28, 36, 119, 190; Frege–Russell logistic programme (mathematics) 3; Frege–
Russell synthesizers 37–39; Frege–Russell theory of proper names 8; Lectures on Logical
Atomism33–34; logical construction of knowledge 2–3; public language, rejection of 33, 34
unicorns 17
verifiability principle 2
Vienna Circle 2, 3
vivid names 50, 52, 87
‘water is H2O’ example 8, 15, 16, 18, 181, 182, 185, 196, 197, 198, 199, 209, 210, 216
Whitehead, A.N. 3
widescopism: and naming 68, 69, 73, 74, 80–85, 90, 91; and necessity 159, 161–62
Wiggins, D. 145
Wittgenstein, L. 2, 3, 56, 63–64, 183
Word and Object (Quine) 125
worldmates 131
world-user’s beliefs and intentions 11, 12
Wright, C. 220