Kripke: “Naming and Necessity”

A RETURN TO MILL

Kripke proposes a Millian theory of names, according to which proper names don’t
have senses. He claims there is an inherent problem in the Fregean notion of sense.

A CONFLATION IN FREGE’S NOTION OF SENSE

Kripke’s response to Fregean theories: we must distinguish sharply between two things
that Fregean theories tend to conflate. In a part of Lecture I not reprinted in Martinich,
he says:

“Frege should be criticized for using the term ‘sense’ in two senses. For he takes the sense
of a designator to be its meaning; and he also takes it to be the way its reference is
determined. Identifying the two, he supposes that both are given by definite descriptions”
(Naming and Necessity, p. 59).

On p. 291, Kripke explicitly draws the distinction between giving the meaning
(=semantic content) and determining the reference.

Giving the semantic content

This is one facet of a Fregean sense. A sense is supposed to be that which is grasped
by the mind, a “meaning” in some non-technical sense. This way of looking at the
description (or cluster of descriptions) associated with a name is to say that the
description (or cluster) gives what Mill would call the connotation of the name.

On this view, a name like ‘Aristotle’, while it refers to Aristotle, means (has as its
semantic content) ‘the tutor of Alexander the Great’, or ‘the most famous pupil of
Plato’, or ‘the author of the Nicomachean Ethics’, or ‘the person who has all or most
of the following properties …’.

Fixing the reference

This is another facet of a Fregean sense, but it is weaker than meaning. Here the
point is that the description is that which picks out the thing that the name is being
used to refer to. On this view, a description (such as ‘the most famous pupil of
Plato’) simply picks out the person to whom the name ‘Aristotle’ is being applied.
This facet of sense does not require that the name ‘Aristotle’ means ‘the most
famous pupil of Plato’. It is a route to a referent (denotation), but not a synonymy.

Kripke’s idea is that we can, and should, separate these two facets. Why?

Copyright © 2008, S. Marc Cohen 1 Last modified 10/20/2008 3:22 PM

 Why fix the reference without giving the semantic content?

If the Fregean view of names were correct, some claims that seem to be contingent
would turn out to state necessary truths, and would be knowable a priori.

Example:

(1) Aristotle was a student of Plato.

is a contingent truth. But if ‘Aristotle’ simply means ‘the most famous student of
Plato’, then (1) is synonymous with:

(2) The most famous student of Plato was a student of Plato.

But (2) is a necessary truth, and can be known a priori to be true. Given that Plato
was a teacher, we don’t have to conduct an empirical study to determine whether
Plato taught Plato’s students. But anything synonymous with a necessary truth is
itself a necessary truth. So, (2) is necessary, so is (1).

But if a name’s associated description(s) only fix the reference, but don’t provide the
semantic content, of a name, we will be able to resist this unpalatable conclusion.
For if ‘Aristotle’ and ‘the most famous pupil of Plato’ don’t have the same meaning,
then (1) and (2) are not synonymous, and may well express different propositions

(2) may be necessary, but (1) is contingent.

A PRIORITY VS. NECESSITY

Kripke also sharply distinguishes between two notions that are often run together.

A Priori

This is an epistemological concept: what we can know independent of experience

(vs. a posteriori, what can be known only empirically, through experience.)

Necessary

This is a metaphysical concept: what is true in every possible world (vs.

contingent: what could have been otherwise, true in some, but not all, possible
worlds.)

It has been a commonplace in philosophy to think these two notions coincide, even if
they are not synonymous. That is, it is a commonplace to think that if it can be known a

Copyright © 2008, S. Marc Cohen 2 Last modified 10/20/2008 3:22 PM

priori that p, then it is necessary that p, and that if it is necessary that p, then it can be
known a priori that p.

The traditional picture looks like this:

Necessary a priori Necessary a posteriori

Contingent a priori Contingent a posteriori

where the two shaded boxes are empty.

Against this, Kripke argues that there can be necessary a posteriori truths, and that
there can be contingent a priori truths. Examples:

Necessary a posteriori

Goldbach’s conjecture: every even number greater than 2 is the sum of two primes.

Fermat’s theorem: ‘xk + yk = zk’ has no solution in the domain of integers for any k
greater than 2.

[To these, Kripke will add, in lecture III (not reprinted in Martinich), the
following examples:

‘Gold has atomic number 79’

‘Water = H2O’
‘This table is made of wood’.

Kripke’s position is that every identity statement whose terms are proper
names is a necessary truth if it is true at all, even though most such statements
are known empirically, a posteriori.]

Contingent a priori (examples omitted in our excerpts)

‘The standard meter stick is one meter long’ (Naming and Necessity, p. 54).
‘Water boils at 100º C’ (Naming and Necessity, p. 56).

These claims may seem bizarre. To understand them, we must get clear on Kripke’s
concept of a rigid designator.

Copyright © 2008, S. Marc Cohen 3 Last modified 10/20/2008 3:22 PM

THESIS: NAMES ARE RIGID DESIGNATORS

Kripke calls both names and descriptions ‘designators’. And he makes clear (p. 290)
that he is considering descriptions as used attributively (as Donnellan would put it). A
rigid designator, he tells us, designates the same object in every possible world. A
nonrigid designator designates different objects in different possible worlds.

Possible worlds

A possible world is simply a way things might have been. It is not something that
can be seen through a telescope or visited on a space ship. The terminology, Kripke
admits, is misleading. It might better be called a possible state of the world or a way
things might have been or a counterfactual situation.

Kripke makes this clear in the preface to the monograph version of Naming and
Necessity, p. 15:

I will say something briefly about ‘possible worlds’. … In the present monograph I
argued against those misuses of the concept that regard possible worlds as something like
distant planets, like our own surroundings but somehow existing in a different dimension,
or that lead to spurious problems of ‘transworld identification’. Further, if one wishes to
avoid the Weltangst and philosophical confusions that many philosophers have associated
with the ‘worlds’ terminology, I recommended that ‘possible state (or history) of the
world’, or ‘counterfactual situation’ might be better. One should even remind oneself that
the ‘worlds’ terminology can often be replaced by modal talk—‘It is possible that …’

Rigid designators

‘The inventor of bifocals’ is a nonrigid designator. In the actual world, it designates

Benjamin Franklin. But, things might have gone differently, and Franklin might not
invented bifocals. They might have been invented by someone else. In this case, the
description ‘the inventor of bifocals’ would designate someone other than Franklin.
(Of course, the description would still have exactly the same meaning — it would
just designate a different person.)

But ‘Benjamin Franklin’ is (according to Kripke) a rigid designator. There is no way

things might have gone according to which Franklin would not have been Franklin.
He might have been a very different sort of person, but he wouldn’t have been
someone else.

Kripke introduces two important caveats (see p. 296) that tend to make the notion of a
rigid designator a little more complicated:

Copyright © 2008, S. Marc Cohen 4 Last modified 10/20/2008 3:22 PM

 1. The object denoted by a rigid designator does not have to exist in every possible
world. (It does not have to be a necessary object.) Benjamin Franklin might
never have existed, but still the name ‘Benjamin Franklin’ is a rigid designator; it
designates the same object in every possible world. (There are some possible
worlds — ones in which Franklin does not exist — in which the name ‘Benjamin
Franklin’ simply does not designate anything that exists in that world.)

2. The claim that a designator is rigid does not mean that we can’t imagine a
possible world in which that designator is used differently from the way it’s
actually used. We can imagine a world in which a different set of parents named
a different person ‘Benjamin Franklin’. But that would not be a world in which
someone else was (as we use the name) Benjamin Franklin.

As Kripke stresses, it’s the way we use the designator in question that
determines which object, in a possible world, it designates. The way we use the
expression ‘the inventor of bifocals’ (attributively) it designates, with respect to
a given possible world, whoever invented bifocals in that world. So in the actual
world, it designates Franklin. In another possible world, it designates Spinoza.
In yet another possible world, it designates no one. (Imagine a possible world in
which there were no bifocals, or in which they were found growing on trees.)

It is certainly possible that Benjamin Franklin might have had a different name.
Suppose his parents had decided to name him ‘Richard’; and suppose that there
was also at the same time a potato farmer in Maine named ‘Benjamin Franklin’.
In this situation, the name ‘Benjamin Franklin’ designates a potato farmer in
Maine. So isn’t this a counterfactual situation in which Benjamin Franklin was a
potato farmer in Maine, and not a statesman living in Philadelphia?

No. This response misunderstands Kripke’s idea. It is not at issue how people,
in the counterfactual situation we are imagining, used the name ‘Benjamin
Franklin’. We use the name ‘Benjamin Franklin’ to pick out the object we are
placing in a counterfactual situation, and then attempt to answer questions about
that object. Hence, with respect to the possible world under discussion, we
would evaluate these propositions as follows:

1. Benjamin Franklin is named ‘Benjamin Franklin’. F

2. Benjamin Franklin is named ‘Richard Franklin’. T

3. Benjamin Franklin is a potato farmer in Maine. F

4. A man named ‘Benjamin Franklin’ is a potato farmer in Maine. T

Copyright © 2008, S. Marc Cohen 5 Last modified 10/20/2008 3:22 PM

AN AMBIGUITY IN ‘RIGID DESIGNATOR’?

Kripke seems to use the term ‘rigid designator’ ambiguously. Compare his formulation
on p. 293, right, middle, with the one on p. 293, right, bottom. This gives us two
possible definitions:

1. α is rigid1 iff α designates the same object in every possible world.

2. α is rigid2 iff α designates the same object in every possible world in which that
object exists.

One might consider a third possibility as giving the idea that Kripke is getting at (cf.
Putnam, p. 311):

3. α is rigid3 iff α designates the same object in every possible world in which α
designates anything at all.

A denoting phrase may be rigid in one of these senses but not in another. Consider a
denoting phrase constructed out of a proper name and a description, e.g., ‘the politician
Nixon’. This is to be understood to mean: ( x)(x is a politician ∧ x = Nixon).

This is obviously not rigid1, since there are worlds in which it does not denote
anything at all.

And it is not rigid2, since it does not designate Nixon in worlds in which Nixon
exists but is not political. (Imagine a world in which Nixon became a druggist and
never went into politics at all.)

But it is rigid3, since it designates Nixon in any world in which it designates

anything. (It designates Nixon in the worlds in which Nixon is a politician, and has
no designation in worlds in which Nixon is not a politician.)

Here’s a “rigidity chart”; see whether your results match mine.

Designator rigid1 rigid2 rigid3

‘The inventor of bifocals’ No No No
‘Richard Milhous Nixon’ No* Yes Yes
‘17’ Yes Yes Yes
‘The politician Richard Milhous Nixon’ No No Yes
‘The positive square root of 16’ Yes Yes Yes

Copyright © 2008, S. Marc Cohen 6 Last modified 10/20/2008 3:22 PM

* The reason for the ‘No’ here is that there are worlds in which Nixon does not exist,
and one might well assume that x must exist in W in order for a designator α to
designate x in W.

But it is likely that Kripke did not assume this. For he seems to have thought that a
designator could designate an object even with respect to a world in which that object
doesn’t exist. Consider, for example, the claim “If Hitler had not existed, WWII would
never have occurred.” Here we are using the name ‘Hitler’ to designate the fanatical
Fascist leader, and evaluating a claim about him with respect to a world in which he
does not exist. It would seem that in order for us to make sense of such counterfactuals
(‘if n had not existed …’, where n is a name) we have to allow names to designate an
object with respect to a world in which that object does not exist.

So this ambiguity in ‘rigid designator’ may not affect Kripke’s claim about names, for
it seems plausible to maintain that proper names are rigid in all three senses, so long as
we allow a designator to designate an object with respect to a world in which that
object does not exist. Given this understanding, Kripke probably intended rigid1 to be
the sense he intended.

Finally, it seems pretty clear that Kripke did not intend every designator that is rigid3 to
count as a rigid designator. For a rigid designator should follow its designatum into
every possible world in which that object exists. That is, to rigidly designate Nixon, a
designator should designate him in every counterfactual situation. But ‘the politician
Nixon’ does not designate Nixon in a world in which Nixon exists but never goes into
politics.

Soames (Beyond Rigidity) calls such designators partially descriptive names, and
claims that they are not rigid designators. His examples include: Princeton University,
Professor Saul Kripke, Justice Antonin Scalia, Miss Ruth Barcan, New York City,
Mount Rainier, Puget Sound, Whidbey Island, The Empire State Building, Yankee
Stadium, etc.

Copyright © 2008, S. Marc Cohen 7 Last modified 10/20/2008 3:22 PM

THE CLUSTER THEORY

Kripke couches his critique as being against what he calls the “cluster theory.” It
consists of the following theses:

1. To every name or designator ‘X’, there corresponds a cluster of properties … φ

such that A believes ‘φ X’.

2. One of the properties, or some conjointly, are believed by A to pick out some
individual uniquely.

3. If most, or a weighted most, of the ϕ’s are satisfied by one unique object y, then y
is the referent of ‘X’.

4. If the vote yields no unique object, ‘X’ does not refer.

5. The statement, ‘If X exists, then X has most of the ϕ’s’ is known a priori by the
speaker.

6. The statement, ‘If X exists, then X has most of the ϕ’s’ expresses a necessary
truth.

But notice that Kripke’s objections apply to any theory that tries to account for the
meaning of proper names in terms of descriptions. Thus, it is directed against Frege
(where a description provides the sense of a name) and Russell (where most names are
thought of as abbreviated descriptions) as well as a theory that tries to account for the
meaning of a name in terms of clusters of descriptions. So it is really a critique of
descriptivism — any theory that attempts to provide a semantic content for names by
means of descriptions.

CRITIQUE OF DESCRIPTIVISM

Classifying Kripke’s Arguments

Following Soames (Beyond Rigidity), we can distinguish between Kripke’s

semantic, epistemic, and modal arguments against descriptivism. Suppose that n is a
name and the D is a description (or cluster thereof) that is supposed to give the
semantic content of n.

Semantic

The referent of n is not linguistically determined by the D (or indeed by any

description or set of descriptions). (vs. 2, 3, 4)

Copyright © 2008, S. Marc Cohen 8 Last modified 10/20/2008 3:22 PM

 Epistemic

What is known or believed by a speaker who says ‘n is F’ is different from what

is known or believed by a speaker who says ‘the D is F’. (vs. 5)

Modal

Sentences like ‘n is F’ behave differently from sentences like ‘the D is F’ when

placed in modal contexts. (vs. 6)

The Semantic Arguments

Against thesis (2)

(2) “One of the properties, or some conjointly, are believed by A to pick out
some individual uniquely.”

Rebuttal: The Feynman–Gell-Mann example (p. 297). Most people who use the
names ‘Feynman’ and ‘Gell-Mann’ cannot supply descriptions that individuate
these men. The typical user of the name ‘Feynman’ can only say ‘Feynman is a
famous physicist’; still, says Kripke, he uses ‘Feynman’ as a proper name of
Feynman.

Against thesis (3)

(3) “If most, or a weighted most, of the ϕ’s are satisfied by one unique object y,
then y is the referent of ‘X’.”

Rebuttal: the Gödel–Schmidt case (p. 298). Suppose Gödel had not proved the
incompleteness of arithmetic, but that the work had been done by a different
man named ‘Schmidt’. Gödel managed to get a hold of the manuscript, and
passed the work off as his own. In this situation, the description ‘the man who
discovered the incompleteness of arithmetic’ would refer to Schmidt, not Gödel.
But, Kripke maintains, ‘Gödel’ still refers to Gödel — for we are imagining a
situation in which Gödel did not discover the incompleteness of arithmetic. We
are not imagining a situation in which Schmidt was Gödel!

Against thesis (4)

(4) “If the vote yields no unique object, ‘X’ does not refer.”

Rebuttal: the cases already described cover this (p. 299):

Copyright © 2008, S. Marc Cohen 9 Last modified 10/20/2008 3:22 PM

 Feynman case:

Most speakers’ descriptive backing for ‘Feynman’ (e.g., ‘a physicist’ or ‘a

leading contemporary theoretical physicist’) does not yield a unique object.

Gödel case:

The speaker’s descriptive backing may even yield no object. E.g., suppose
no one had discovered the incompleteness of arithmetic, but “the proof
simply materialized by a random scattering of atoms on a piece of paper”
(p. 299). Then the descriptive backing for ‘Gödel’ would yield no object,
but the name would still refer to Gödel.

The Epistemic Arguments

Against thesis (5)

(5) “The statement, ‘If X exists, then X has most of the ϕ’s’ is known a priori
by the speaker.”

Rebuttal: the Gödel–Schmidt case again. Even if I am right in thinking that

Gödel is in fact the discoverer of incompleteness, and the ‘Schmidt’ story just a
fantasy, I still can’t know this a priori.

The Modal Arguments

Against thesis (6)

(6) “The statement, ‘If X exists, then X has most of the ϕ’s’ expresses a
necessary truth.”

Rebuttal: The Aristotle example. “It just is not, in any intuitive sense of
necessity, a necessary truth that Aristotle had the properties commonly
attributed to him … It would seem that it’s a contingent fact that Aristotle ever
did any of the things commonly attributed to him today …” (pp. 295-96).

Copyright © 2008, S. Marc Cohen 10 Last modified 10/20/2008 3:22 PM

 Here a thought experiment helps to establish this. Fix on Aristotle.
Suppose his parents had lived in a different town, not Stagira. Now,
imagine that as a youth he never goes to Plato’s academy. He stays at
home, in Sparta, perhaps, where he becomes a physician, like his father.
The rest of his life is virtually entirely different from the one we read about
in the history books. What we have described is a possible world in which
Aristotle has (virtually) none of the ϕ’s (rather than a world in which
Aristotle does not exist).

KRIPKE’S ALTERNATIVE: “INITIAL BAPTISMS” AND CAUSAL CHAINS

Not an alternative theory, but a “better picture.” Beginning on p. 299, left bottom:

“Someone, let’s say, a baby, is born; his parents call him by a 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 this chain, who has heard about, say Richard
Feynman, in the market place or elsewhere, may be referring to Richard
Feynman even though he can’t remember from whom he first heard of
Feynman or from whom he ever heard of Feynman.”

The idea is that there is a causal link between an initial use of a name (an “initial
baptism”) and subsequent uses by later speakers. It is this causal link, not the speaker’s
grasp of the descriptive content of the name, that determines which thing his use of the
name refers to. Kripke calls this a “better picture” than the description theory, but not
yet quite a theory (p. 300):

“A rough statement of a theory might be the following: An initial ‘baptism’

takes place. Here the object may be named by ostension, or the reference of
the name may be fixed by a description. When the name is ‘passed from link
to link’, the receiver of the name must, I think, intend when he learns it to use
it with the same reference as the man from whom he heard it.”

Kripke does not claim that this is an analysis of the notion of reference: “it takes the
notion of intending to use the same reference as a given” (p. 300).

Kripke’s picture stresses the importance of the notion of intending to refer in

describing the causal chain. Each user of the name implicitly intends to refer, when he
uses it, to the object to which those from whom he learned it intended to refer. And this
chain must stretch back to an initial baptism — an event in which the object named
itself figures causally.

Copyright © 2008, S. Marc Cohen 11 Last modified 10/20/2008 3:22 PM

 So it’s not quite right to suppose that Kripke’s “picture” takes the mental component
out of the theory of reference just because it replaces the description theory with a
theory of causal communicative links. For the nodes in the causal chain are intentions
to refer (and, of course, the overt acts of dubbing, naming, etc. that manifest those
intentions).

IDENTITY STATEMENTS BETWEEN NAMES

Kripke claims that when an identity statement involves a rigid and a nonrigid
designator (flanking the identity sign), the statement is not necessary. (Cf. ‘Franklin =
the inventor of bifocals’. It is only a contingent fact that Franklin was the unique
individual who invented bifocals.)

In general, any true statement of the form ‘N = the ϕ’, where ‘N’ is a rigid designator
and ‘the ϕ’ is a nonrigid designator, will be only contingently true. For if ‘N’
designates the same object in all possible worlds, but ‘the ϕ’ designates different
objects in different possible worlds, there will be at least one possible world in which
‘N = the ϕ’ is not true. Hence, it’s not necessary (i.e., it is not true in all possible
worlds).

Note: some true statements of the form ‘N = the ϕ’ will be necessarily true, but that
will require that ‘the ϕ’ be a rigid designator. This is the right result, since it is
intuitively clear that ‘4 = the positive square root of 16’ is necessarily true.

Any statement of the form ‘N = M’, where ‘N’ and ‘M’ are rigid designators, will be a
necessary truth if it is true at all. (This follows from the definitions of ‘rigid
designator’ and ‘necessary truth’.)

So, since Kripke is committed to the thesis that proper names are rigid designators, he
must also hold that every identity statement involving proper names is either
necessarily true or necessarily false.

To try to make this somewhat startling thesis seem more plausible, Kripke defends it in
the case of ‘Hesperus = Phosphorus’. ‘Hesperus’ is a name that has been given to the
evening star. ‘Phosphorus’ is a name of the morning star. (That is, both names refer to
Venus.)

Now it may seem that Hesperus might not have been Phosphorus (i.e., that ‘Hesperus =
Phosphorus’ is not a necessary truth’). For we can imagine a world in which it is not
the same planet that is the last visible in the morning and the first visible in the
evening. In such a world, we might have named two different planets with the names
‘Hesperus’ and ‘Phosphorus’.

Copyright © 2008, S. Marc Cohen 12 Last modified 10/20/2008 3:22 PM

Kripke’s response is that this is not a world in which Hesperus isn’t Phosphorus.
Rather, it’s a world in which we might have used one of the names ‘Hesperus’ or
‘Phosphorus’ to name a different object from the one we actually use it to name.

So although ‘Hesperus = Phosphorus’ is a necessary truth, “The object named

‘Hesperus’ = the object named ‘Phosphorus’” is not. That’s because, on Kripke’s
view, the description “the object named ‘Hesperus’” is a nonrigid designator.
Nothing but Hesperus could have been Hesperus; but something other than Hesperus
could have been named ‘Hesperus’.

And, of course, Kripke rejects the view (espoused by W. Kneale) that the meaning
of a proper name, e.g., ‘Socrates’, is “the man named ‘Socrates’”.

So, given our usage of the names ‘Hesperus’ and ‘Phosphorus’, there is no possible
world in which Hesperus is not Phosphorus — for there’s no possible world in which
Hesperus is not itself.

This does not mean, of course, that one can know a priori that Hesperus is Phosphorus.
One cannot: it took an astronomical discovery to find out that Hesperus is Phosphorus.
So here we have a necessary truth (that Hesperus is Phosphorus) that cannot be
known a priori.

The appearance that it is only contingently true that Hesperus is Phosphorus is due to
a conflation of epistemic with metaphysical possibility. I may be in a position where it
is possible, for all I know, that Hesperus is not Phosphorus. That is, I may not realize
that the object named ‘Hesperus’ is identical to the object named ‘Phosphorus’. That’s
because the evidence in the two cases (one where the two names name the same object,
the other where they don’t) can be qualitatively indistinguishable.

Copyright © 2008, S. Marc Cohen 13 Last modified 10/20/2008 3:22 PM