WITTGENSTEIN

Stanley Cavell’s Wittgenstein

By James Conant

T
HE AIM OF THIS PAPER IS TO PROVIDE AN OVERVIEW OF ONE ASPECT OF
Stanley Cavell’s reading of Wittgenstein: his interpretation of
Wittgenstein’s teaching concerning the relation between meaning
and use. After furnishing a rough overview of this aspect of Cavell’s
reading of Wittgenstein, I shall attempt to bring into focus what is most
distinctive and elusive about it by illustrating, by means of some examples drawn
from commentators on Cavell’s work, how easy it can be to miss Cavell’s and—
if Cavell is right about who Wittgenstein is—Wittgenstein’s point.
Now Wittgenstein has become quite famous in recent years for putting
forward something that gets called a “use-theory of meaning.” Wittgenstein
writes:

For a large class of cases—though not for all—in which we employ the word
“meaning” it can be defined thus: the meaning of a word is its use in the
language. 1

If you look at commentaries on Wittgenstein’s work, you will find that

commentators have wildly different understandings of what Wittgenstein is
saying in this passage. Here is what Cavell says:

“The meaning is the use” calls attention to the fact that what an expression
means is a function of what it is used to mean or to say on specific occasions
by human beings. . . . [T]o trace the intellectual history of philosophy’s
concentration on the meaning of particular words and sentences, in isolation
from a systematic attention to their concrete uses would be a worthwhile
undertaking. . . . A fitting title for this history would be: Philosophy and the
Rejection of the Human.
Wittgenstein’s motive . . . is to put the human animal back into language
and therewith back into philosophy. . . . He undertook, as I read him, to trace the

James Conant is professor of philosophy at the University of Chicago. He received

both his BA (1982) and PhD (1990) from Harvard University. He taught for
nine years at the University of Pittsburgh before moving to Chicago in 1999.
He has published articles in philosophy of language, philosophy of mind,
aesthetics, among other areas, and on philosophers such as Wittgenstein, Kant,
Nietzsche, Kierkegaard, William James, Frege, Carnap, Putnam, Cavell, Rorty,
and McDowell, among others. He is currently working on three projects: a
monograph on skepticism, a co-authored work (with Cora Diamond) on
Wittgenstein, and a forthcoming collection of essays. He has edited two volumes
of Hilary Putnam's papers and co-edited (with John Haugeland) one volume of
Thomas Kuhn's papers.
vol.XIII no.1 2005 THE HARVARD REVIEW OF PHILOSOPHY
52 James Conant
mechanisms of this rejection in the ways in which in investigating ourselves,
we are led to speak “outside language games,” consider expressions apart
from, and in opposition to, the natural forms of life which give those
expressions the force they have. . . . What is left out of an expression if it is used
“outside its ordinary language game” is not necessarily what the words mean
(they may mean what they always did, what a good dictionary says they mean),
but what we mean in using them when and where we do. The point of saying
them is lost.
How great a loss is that? To show how great is a dominant motive of the
Investigations. What we lose is not the meaning of our words—hence, definitions
to secure or explain their meaning will not replace our loss. What we lose is a
full realization of what we are saying; we no longer know what we mean.2

I want to draw attention to four topics that are touched upon in this
passage: (1) philosophy’s usual way of asking after the meaning of an
expression—“philosophy’s concentration on the meaning of particular
words and sentences, in isolation from a systematic attention to their
concrete uses”; (2) Wittgenstein’s contrasting way of asking after the
meaning of an expression— which takes it to be “a function of what it is
used to mean or to say on specific occasions by human beings”; (3) where
the first way leads us—when we engage in the philosopher’s usual way
of employing language, we are led to speak “outside language games”;
and (4) what the second way hopes to show us—that, when we are led to
speak thus, “we no longer know what we mean.”
Let’s start with (1) and (2). A contrast is drawn here between
two ways of asking after the meaning of an expression:

(1) asking after its meaning in isolation from a systematic

attention to concrete uses (asking after what the word or sentence
means);

(2) asking after its meaning through a consideration of its

concrete uses (asking after what the word or sentence is used to
mean on specific occasions by human beings).

Readers of Wittgenstein assume that what must be at issue in his

invocation of “use” is the importance of asking the second of these
questions over and above merely asking the first; or, to put the point in more
fashionable—and therefore more dangerous—terms: that what must be at issue
is the importance of supplementing our syntactic and semantic theory of
language with a third layer, a theory of the pragmatics of natural language or
a theory of speech-acts. Cavell takes Wittgenstein to want to urge something
potentially far more threatening to traditional analytic philosophy of language
than any such mere call for supplementation; namely, that “for a large class of
cases” there isn’t anything which can count as asking the first question (asking
what a sentence means) apart from a prior consideration of the second question
(apart, that is, from asking: when might it be said? where? by whom? to whom?).
Let us now turn for a moment to (4). Wittgenstein, according to Cavell,
goes on to claim something which has proved even harder for the analytic

THE HARVARD REVIEW OF PHILOSOPHY vol.XIII no.1 2005

 Stanley Cavell’s Wittgenstein 53

philosophical tradition to hear: namely, that when, under the pressure of

philosophy, meaning and use come apart, then we are subject not just to some
degree of ignorance or perplexity as to what we mean; rather, we are subject to
something (which is apt to strike us as) far more paradoxical: we are subject to
an illusion of meaning. The way Cavell puts this is to say that, when the
philosopher reaches such a crossroads in his philosophizing, he is subject to a
“hallucination of meaning”:

[H]e imagines himself to be saying something when he is not, to have

discovered something when he has not. Someone in these particular straits
may be described as hallucinating what he or she means, or as having the
illusion of meaning something. 3

This is paradoxical insofar as it challenges the deeply rooted intuition

that in the realm of meaning esse est percipe: if I seem to mean something by
my words then, by golly, I do. Wittgenstein himself concedes what he takes to
be sound in this intuition when he writes: “one feature of our concept of a
proposition is sounding like a proposition.”4 But it is only one feature of our
concept of a proposition, and it is central to Cavell’s and Wittgenstein’s teachings
that possession of this feature, though a necessary condition on something’s
counting as a proposition, is by no means a sufficient one.
This takes us to Wittgenstein’s numerous thought-experiments, such as
his famous invitation to his reader to try to imagine a private language.5
According to Cavell, Wittgenstein offers these carefully constructed invitations
to the reader to imagine such-and-such, in order to lead the reader up a
dialectical ladder. When the reader has reached the top of the ladder, he is to
throw the ladder away: he is supposed to come to see that he had, all along,
only been under the illusion of imagining something.6 More specifically,
according to Cavell, what Wittgenstein wants us to come to see about such
cases is that such a hallucination of meaning arises because we imagine we
transfer the meaning of an expression where we have failed to transfer the use.
Wittgenstein describes the sort of awkward relation we occupy with respect to
our words in such cases as one in which we are led “to speak outside language
games.”
This takes us back to (3). Cavell’s most sustained discussion of a
Wittgensteinian example of how the philosopher is led to speak “outside
language games” is his discussion of epistemological skepticism. The
skeptic wants to examine the integrity of our claims to knowledge; but
not any sort of claim will suit the purposes of his investigation. It has to,
for the skeptic’s purposes, be (what Cavell calls) “a best case of knowledge”—
the sort of case which, if a ground for doubt can be adduced, will thereby
cast doubt over the possibility of knowledge as such.7 On Cavell’s reading,
Wittgenstein’s aim is to show the skeptic that he hovers between the horns
of a dilemma. Here is how Cavell summarizes the dilemma:

If the epistemologist were not imagining a claim to have been made, his procedure
would be as out of the ordinary as the ordinary language philosopher finds it to
be. But, on the other hand, if he were investing a claim of the sort the coherence

vol.XIII no.1 2005 THE HARVARD REVIEW OF PHILOSOPHY

54 James Conant
of his procedures require . . . then his conclusion would not have the generality
it seems to have.8

Either (this is the first horn of the dilemma) the claim that the skeptic adduces
will not be a (proper) claim or (this is the second horn) the claim is a (proper)
claim to knowledge. If it is the former, Wittgenstein aims to show that an
investigation of its epistemic credentials does not bear on the integrity of our
ordinary claims to knowledge. If the latter, Wittgenstein aims to show that it
will not be the sort of example of knowledge the failure of which can serve the
skeptic’s purposes: namely, the specification of a ground for doubting it will not
cast a shadow over the whole of our knowledge.
The epistemologist may try to seize the first horn of this dilemma and
say that he knows that he is using the word “know” or “claim” in a strained
and unusual way, but not to worry; for he means to use the word in a special
philosophical sense, and it is a relatively technical concept of “knowing” or
“claiming” that he aims to investigate. But then he runs up against the problem
that the conclusion of his investigation cannot yield, as he takes it to, a discovery
concerning the integrity of our ordinary claims to knowledge—it can at most
yield a discovery about so-called “knowing” or so-called “claiming.” If the
epistemologist wants his conclusion to yield a discovery about the nature of
knowledge überhaupt his investigation cannot afford to lose sight of our concept
of knowledge. Cavell remarks:

Let me . . . emphasize what I take the sense of discovery to indicate about the
philosopher ’s conclusion. First, since it is a discovery whose content is that
something we have, supposedly, all believed has been shown to be false or
superstitious or in some respect suspect, its sense of being a discovery
depends upon a sense of its conflicting with what we would all, supposedly,
have said we knew or thought. And this sense of conflict depends upon the
words which express the conclusion meaning or seeming to mean what these
words, as ordinarily used, would express. For that is what the conclusion
conflicts with. When the philosopher concludes that “we don’t really see or
know” something, why would this so much as seem to deny what the ordinary
person means when he says “I see or know” something unless the words meant,
or seemed to mean, the same thing? 9

The sense of having arrived at a philosophical discovery—that is, a

discovery which reveals that things are not at all as we pre-
philosophically suppose them to be—is the hallmark of skepticism. Thus
if the skeptic is to be entitled to his sense of discovery he has no choice
but to accept the second horn of the dilemma, and to agree to the
condition that, when he concludes that “we don’t really see or know”
something, his conclusion will only be able to conflict with what the
ordinary person believes (namely, that he sees or knows something when
he says “I see” or “I know” something) if, when the skeptic and the
ordinary person employ it, the word “see” or “know” signifies the same
concept. Thus the burden of Wittgenstein’s treatment of skepticism, on
Cavell’s reading, is directed towards a discussion of the second horn of
the dilemma: to showing the skeptic that if his words are meant in the

THE HARVARD REVIEW OF PHILOSOPHY vol.XIII no.1 2005

 Stanley Cavell’s Wittgenstein 55

ordinary way—if his example of a claim to knowledge is taken to involve an

ordinary claim to knowledge (for example, a claim to know that there is a
goldfinch in the garden)—then the overturning of the claim in question will fail
to generalize in the requisite way for it to disclose a discovery about the nature
of knowledge as such. When a ground for doubt is entered against an ordinary
claim to knowledge, it may show that the person who has made the claim doesn’t
know what he claims to know (for example, that he doesn’t know that the bird
is a goldfinch) but not that the speaker doesn’t know anything.
The skeptic’s twin sense of the peculiarity and the coherence of his
investigation derives from his hovering between the two horns of this dilemma,
without determinately settling on either one. He wants to use words in an
admittedly somewhat extraordinary way and yet mean them in their ordinary
sense. He wants to occupy at the same time a position which is both inside and
outside the ordinary language-game in which his words have their home. He
therefore speaks “outside language games.” This does not just mean that he
speaks outside the ordinary (and therefore inside some relatively esoteric
scientific or literary) language-game; it means that he speaks “outside language
games” altogether, ordinary or extraordinary. Wittgenstein describes what
happens when we “speak outside language games” as cases of language
“idling”10 or being “on holiday.”11 Cavell takes these descriptions seriously.
The words we call upon in such cases are said to be “idle” or “on holiday”
because they fail to be at work: they do not engage any determinate
circumstances of use.
What the skeptic needs, in order to pull off his trick, is to engage
in a performance that qualifies as a speech-act of claiming while
prescinding from all of the messy context-dependent details that come
with any actual concrete situation in which a claim is made. In the
following passage, Cavell discusses a putative example of such an utterly
generic case of claiming of the sort which the skeptic’s inquiry requires
as its point of departure if its conclusion is to have any hope of attaining
the requisite degree of generality:

[N]o one would have said of me, seeing me sitting at my desk with the green
jar out of my range of vision, “He knows there is a green jar of pencils on the
desk,” nor would anyone say of me now, “He (you) knew there was a green
jar…”, apart from some special reason which makes that description of my
“knowledge” relevant to something I did or said or am doing or saying. . . .
Perhaps one feels: “What difference does it make that no one would have
said, without a special reason for saying it, that you knew the green jar was on
the desk? You did know it; it’s true to say that you knew it. Are you suggesting
that one cannot sometimes say what is true?” What I am suggesting is that
“Because it is true” is not a reason or basis for saying something; and I am
suggesting that there must, in grammar, be reasons for what you say, or be a
point in your saying of something, if what you say is to be comprehensible. We
can understand what the words mean apart from understanding why you say
them; but apart from understanding the point of you saying them we cannot
understand what you mean. 12

vol.XIII no.1 2005 THE HARVARD REVIEW OF PHILOSOPHY

56 James Conant

Cavell argues here that there isn’t anything which is just saying (and therewith
meaning) the words “He knew there was a green jar there.” To mean these
words one must mean something by them and this requires that one have a
reason or basis for saying them. In the absence of such a reason or basis, there
isn’t anything which is the thought which I express in uttering these words.
This cuts against the received wisdom of contemporary philosophy of language.
Philosophers are apt to think that it suffices to fulfill the conditions for the
expression of a determinate thought if I utter a sentence which satisfies the
following two conditions: (i) the individual words of which the sentence is
composed are so-called “meaningful” words of the language, and (ii) the words
are combined in accordance with the so-called “rules” of the language.
Wittgenstein contests just such a view in the following passage from On
Certainty:

“I know that that’s a tree”—this may mean all sorts of things: I look at a plant
that I take for a young beech and that someone else thinks is a black-currant.
He says “that is a shrub”; I say it is a tree. —We see something in the mist
which one of us takes for a man, and the other says “I know that that’s a tree.”
Someone wants to test my eyes etc. etc.—etc. etc. Each time the ‘that’ which
I declare to be a tree is a different kind. 13

As we run though these different examples of speaking the sentence “I

know that that’s a tree,” on each occasion of speaking, the sentence—
that is, the string of words—uttered remains the same. In this sense of
“say” (uttering certain words), in each case we say the same thing: “I
know that that’s a tree.” But, in each of these very different cases,
Wittgenstein says, “the ‘that’ which I declare to be a tree is of a different
kind”—in this sense of “say” (and this is the sense that matters for
Wittgenstein), in each case that I say “that’s a tree,” what I say is different.
In each case, the context makes a contribution to what thought it is that
I express by these words. Wittgenstein, according to Cavell, thinks that
that it is a misunderstanding of how language works to think, as
philosophers are prone to think, that the role of a sentence in our language
is to provide something which on its own steam (apart from a context
which confers a fully determinate sense upon it) allows for the expression
of a determinate thought. The role of a sentence, rather, is to provide a
linguistic instrument which is usable in many different circumstances
to express any of many distinct thoughts.
In order better to see the elusiveness of this conception of
meaning, let us look at some recent attempts to explicate it. Here is
Michael Williams on Cavell:

Cavell’s key notion is that of a “concrete” claim. A concrete claim is one that
has a definite point: it is a claim that is made to inform, to warn, to amuse, or
whatever. It is, for Cavell, a matter of “grammar ” that an utterance must enter
a concrete claim if it is to count as an intelligible act of assertion. This means
that truth alone does not guarantee intelligible assertability, for a statement
can be true without its assertion having any point at all. . . . [I]n Cavell’s eyes,
though the skeptic uses meaningful sentences, and even uses them in ways

THE HARVARD REVIEW OF PHILOSOPHY vol.XIII no.1 2005

 Stanley Cavell’s Wittgenstein 57
that are recognizably analogous to familiar uses, an essential feature of his
enterprise forces him to enter his claims in a way that violates the conditions
for fully meaningful speech. Though his words mean what they always mean,
there is nothing he means by them. In this way, the sceptic’s procedure generates
an illusion of meaning. 14

Williams has Cavell here saying that the proposition the skeptic comes
out with may well be true but there is a problem about asserting this
(true) proposition in the context in which the skeptic wants to come out with it.
Now it is a necessary condition on a proposition’s being true that it be
meaningful; that is, it is a necessary condition on our being able to assess the
truth of a proposition that first we be able to understand it. So Williams, in
effect, has Cavell saying that there isn’t any problem about what claim the
skeptic wants to make—there isn’t any problem about what his proposition
means, or alternatively: there isn’t any problem about what the skeptic’s utterance
would mean if it were assertible—the problem just is that the skeptic’s claim
runs into conflict with various, as it were, additional (pragmatic) constraints on
assertibility. So it looks as if the problem lies not with the semantic content of
the assertion, but with arranging a situation in which we could avail ourselves
of its semantic content—a situation in which we could felicitously perform the
relevant speech-act without it misfiring. So it starts to look as if we can think
the (skeptic’s) thought alright, there is just a problem about finding a situation
in which we can express it out loud. It looks as if the problem lies not in what
the skeptic is trying to say, but in an incompatibility between the content of
what he wants to say and the context of utterance. This suggests a certain way
of taking what Wittgenstein means when he says we are led to speak outside
language-games: it looks as if the unintelligibility of certain utterances is to be
traced to a failure of fit between certain kinds of sentences and certain kinds of
situations.
This (mis)reading of Cavell on Wittgenstein is also to be found in
the work of Marie McGinn:

Cavell’s attempt to show that the knowledge claims that that the philosopher
investigates are illegitimate or unintelligible is thus an attempt to show, not
that the words that the philosopher utters in introducing these claims are
themselves meaningless, but that, given the context in which he utters them,
we cannot see the point of his saying them, we cannot see what he means by
them, we cannot construe his utterance of them as an act of intelligible
assertion.
The crucial idea, therefore, is that there are two distinct notions of meaning—
word-meaning and speaker ’s-meaning—that are linked together in a much more
complex way than the traditional philosopher has supposed. . . . Cavell’s view of
the relationship between word-meaning and speaker’s-meaning might . . . be
expressed as the claim that it is a mistake to suppose that the task of interpreting
others can ever be taken over completely by a systematic theory of meaning for
[a language]. . . . In particular, a speaker’s uttering a given sentence, s, to which
a theory of meaning assigns the interpretation, p, is never a guarantee that the
speaker is correctly described as having performed the act of asserting p. . . .
For the interpreter can never put off altogether the need to satisfy himself
that the content-specifying description of the act of assertion that the systematic
theory yields makes this particular utterance, in these particular concrete

vol.XIII no.1 2005 THE HARVARD REVIEW OF PHILOSOPHY

58 James Conant
circumstances, intelligible to him as the act of a human agent participating in a
humanly recognizable form of life. 15

McGinn here attributes to Wittgenstein—and Cavell—an implicit reliance on a

distinction made famous by Paul Grice: the distinction between sentence-
meaning and speaker’s-meaning. 16 The additional, supposedly distinctively
Wittgensteinian, contribution to an understanding of the nature of language,
according to McGinn, lies in helping us to see that word-meaning and speaker’s-
meaning “are linked together in a much more complex way” than the traditional
philosopher has supposed. So, whereas Grice might have supposed that what
the words of a sentence mean very nearly fully specifies what would be said on
any speaking of them, Wittgenstein teaches us that—although the words do
specify what is “meant” in one sense of “meaning”—there are two “distinct
notions of meaning,” and thus what is “meant” (in this second sense of
“meaning”) still remains to be settled. Insofar as we only specify what the
meaning of a sentence is, an important ingredient of what is meant on any
speaking of the sentence is left out—for it turns on something further: on the
point of saying it, on one’s reasons for so speaking. Thus McGinn says that, in
cases of attempted speech-acts of assertion which misfire, the problem lies—not
in our being unable to specify the content of the assertion, but rather—in seeing
what is supposed to have been the point of having uttered this (independently
meaningful) proposition in this (unsuitable) context.
What happens in McGinn’s reading of Cavell on Wittgenstein is, in
effect, that a distinction is introduced between two levels of nonsense:

[The] attempt to show that the knowledge claims that the philosopher
investigates are illegitimate or unintelligible is thus an attempt to show, not
that the words that the philosopher utters in introducing these claims are
themselves meaningless, but that, given the context in which he utters them,
we cannot see the point of his saying them, we cannot see what he means by
them, we cannot construe his utterance of them as an act of intelligible
assertion. 17

This passage implicitly draws a distinction between a claim’s being

meaningless and its being unintelligible. Meaningfulness has to do with
sentences, and intelligibility has to do with context-embedded speech-
acts. It can be perfectly clear what the meaning of a sentence is; yet a
context-embedded utterance of it can fail to be intelligible because it can
fail “to be intelligible as the act of a human agent participating in a
humanly recognizable form of life.” This allows McGinn to conclude that Cavell’s
Wittgenstein holds (i) that the sentences that the philosopher utters are
themselves perfectly meaningful, and yet (ii) that, given the context in which
he utters them, his utterance can nonetheless be charged with unintelligibility.
“Unintelligible” here means: we understand what his words mean but we cannot
see the point of his saying them, we cannot see what he means by them.
Thus we arrive at a fairly complete misunderstanding of the passages
in Cavell that are the target of Williams’s and McGinn’s exegeses. The whole
point of this region of Wittgenstein’s thought, on Cavell’s reading of him, is to
help us to come to see that meaning and use cannot come apart from one another

THE HARVARD REVIEW OF PHILOSOPHY vol.XIII no.1 2005

 Stanley Cavell’s Wittgenstein 59

in just the manner that Williams and McGinn have Cavell supposing that it
can. Consider the following passage from On Certainty:

Just as the words “I am here” have a meaning only in certain contexts, and not
when I say them to someone who is sitting in front of me and sees me clearly, —
and not because they are superfluous, but because their meaning is not determined
by the situation, yet stands in need of such determination. 18

What Wittgenstein says here is not (as Williams and McGinn propose):
It is clear what the sentence “I am here” means, yet what is meant in
saying it remains less than fully intelligible given the unsuitability of
the context of use. What Wittgenstein says here about the words “I am
here” is precisely the opposite: that “their meaning is not determined by
the situation”—that their meaning still “stands in need of
determination.” In Philosophical Investigations, Wittgenstein employs this
same example (“I am here”) to emphasize that “the meaning of an
expression” is not something which an expression possesses already on
its own and which is subsequently imported into a context of use:

You say to me: “You understand this expression, don’t you? Well, then—I am
using it in the sense you are familiar with.” —As if the sense were an
atmosphere accompanying the word, which is carried into every kind of
application.
If, for example, someone says that the sentence “This is here” (saying which
he points to an object in front of him) makes sense to him, then he should ask
himself in what special circumstances this sentence is actually used. There it
does make sense. 19

What we are tempted to call “the meaning of the sentence” is not a

property the sentence already has in abstraction from any possibility of
use and which it carries with it—like an atmosphere accompanying it—
into each specific occasion of use. It is, as Wittgenstein keeps saying, in
the circumstances in which it is “actually used” that the sentence has
sense. This is why Wittgenstein says in the previous passage from On
Certainty: the words “I am here” have a meaning only in certain contexts—
that is, it is a mistake to think that the words themselves intrinsically
possess some sort of meaning apart from their capacity to express a
meaningful thought when called upon in a context of use. The problem
with the pseudo-employment of “I am here” under consideration in the
passage above is that the meaning of the words “is not determined by the
situation”; that is to say, it is not clear, when these words are called
upon in this context, what is being said—if anything.
The philosopher, Wittgenstein says, tends to think that he understands
“the meaning of a sentence” apart from and prior to any concrete occasion of
use:

A philosopher says that he understands the sentence “I am here,” that he

means something by it, thinks something—even when he doesn’t think at all
how, on what occasions, this sentence is used.20

vol.XIII no.1 2005 THE HARVARD REVIEW OF PHILOSOPHY

60 James Conant

The philosopher takes there to be something which is the thought which the
sentence itself expresses. He takes himself already to know what it means: what
it means is a function of what these words combined mean. To consider the use
of the sentence for such a philosopher, is to consider an additional dimension of
meaning. An investigation of “use,” for such a philosopher, is an investigation
into the relationship between “the meaning of the sentence”—which we are
able to grasp independently of its contexts of use—and the sorts of things this
sentence can express or imply (over and above what it means taken by itself)
when brought into conjunction with the various contexts of use into which it
can be intelligibly imported. Questions can be raised about why what is said is
said and what the point of saying it on a particular occasion of use is. But the
very possibility of asking such questions presupposes that it is already reasonably
clear what thought is expressed, and thus what it would be for the truth to have
been spoken on this occasion of speaking. Cavell’s Wittgenstein is concerned to
contest such a conception of the relation between meaning and use. What your
words say depends upon what they are doing—how they are at work—in a
context of use. Wittgenstein writes:

If someone says, “I know that that’s a tree” I may answer: “Yes, that is a
sentence. An English sentence. And what is it supposed to be doing?” 21

The charge is directed here not against the sentence “I know that that’s a
tree,” but against a failure on the part of a speaker to provide the sentence
with something to do on an occasion of speaking. This is not to say that
the sentence “I know that that’s a tree,” where uttered when a tree stands
in plain and open view, cannot be given a sense. We can, as Wittgenstein
repeatedly emphasizes, find a context of use in which these words would
be doing something under such circumstances. Thus, in the following
passage, Wittgenstein sketches a background that confers upon the
sentence “I know that that’s a tree” a sense, even though the sentence is
uttered while a tree stands in plain and open view:

[S]omeone who was entertaining the idea that he was no use any more might
keep repeating to himself “I can still do this and this and this.” If such thoughts
often possessed him one would not be surprised if he, apparently out of all
context, spoke such a sentence [as “I know this is a tree”] out loud. (But here
I have already sketched a background, a surrounding, for this remark, this is
to say given it a context.) 22

To sketch in this way a background, a surrounding, for a remark is to confer a

determinate meaning on it—it is to enable us to see what is being claimed by
the speaker who claims by means of this remark to know something.
Consider the following sentence: “There is a lot of coffee on the table.”23
This sentence may, on one occasion of speaking or another, say any of indefinitely
many distinct things—it may express indefinitely many distinct thoughts. Each
of these thoughts will be true under different conditions. Wittgenstein points
out that “how a sentence is meant can be expressed by an expansion of it.”24 If
you wished to expand on what you meant in having said, “There is a lot of

THE HARVARD REVIEW OF PHILOSOPHY vol.XIII no.1 2005

 Stanley Cavell’s Wittgenstein 61

coffee on the table,” depending upon what you meant, a different expansion
would be required. In clearing up a certain misunderstanding of what was meant
by your words, you might find yourself saying: “I meant that coffee has been
spilled on the table”; or alternatively: “I meant that there is a huge urn of coffee
on the table in question”; or: “I meant that there are bags of coffee stacked on
the table.” But we cannot account for these differences in what is said (in
expressing each of these different thoughts by, in each case, uttering the words
“There is a lot of coffee on the table”) by supposing that we are drawing on
different meanings of the words “there,” “coffee,” “a lot,” “is,” “on,” “the,” or
“table.” The indeterminacy in what thought the sentence “There is a lot of coffee
on the table” expresses in each of these sayings of it, is not one that turns on
any ambiguity in the meaning of the words of which it is composed. In the
sense in which it makes sense to speak of “the meanings of the words” (that is,
what the dictionary says their meaning is), the same “meaning of the word” is
being drawn on for each of these words (“there,” “coffee,” “a lot,” “is,” “on,”
“the,” “table”) in each of these distinct uses of the sentence. Nevertheless, what
is meant by the sentence, in each case, is not the same. Seeing what words, on a
given occasion of speaking, mean is a matter of appreciating what they can
mean in the circumstances of the speaking. It is a matter of perceiving—of the
various possible contributions which circumstances of use might make—what
sort of contribution the actual circumstances are most reasonably taken to make.
For Cavell’s Wittgenstein, understanding a proposition is a matter of perceiving
a certain physiognomy of meaning in an employment of words. This is not
something you can do apart from a consideration of the context of significant
use.
The following is an example of the sort of passage from Cavell that
Williams and McGinn latch onto and from which they construct their reading
of Cavell:

“Not saying anything” is one way philosophers do not know what they mean.
In this case it is not that they mean something other than they say, but that they
do not see that they mean nothing (that they mean nothing, not that their
statements mean nothing, are nonsense).25

Cavell’s point here is that, in cases where there is a failure of meaning, the failure
is to be traced to a failure on the part of the speaker to project that string into a
new context in a fashion which admits of a stable reading—in a fashion which
admits of our being able to perceive in the sentence, when we view it against the
background of its circumstances of use, a coherent physiognomy of meaning.
The reason for not putting the blame on the linguistic string taken is isolation is
not, pace Williams and McGinn, because it is perfectly clear what the string
taken by itself must mean; but rather because there is no straightforwardly
clear sense to be made of what the speaker wants it to mean in the context in
question. This is, again, not because—as McGinn and Williams have it—this
linguistic string and that context are inherently incompatible; for, as Cavell points
out and Wittgenstein repeatedly emphasizes, we could stipulate a sense for these
words in this context (but then we run against the first horn of the dilemma), or

vol.XIII no.1 2005 THE HARVARD REVIEW OF PHILOSOPHY

62 James Conant

we could find or invent a context of use for this combination of words which is
a natural extension of their ordinary language-game (but then we run up against
the second horn of the dilemma).
For Cavell’s Wittgenstein, there are not two distinct and autonomous
levels of making sense—the semantic and the pragmatic—such that one can
fully meet the first-level demands for propositionhood and yet still fall short of
the conditions for being an assertible proposition. It is just such a view
that is repudiated by Cavell in the following passage:

‘[I]t makes sense’ . . . just means that we can easily imagine circumstances in
which it would make sense to say it. . . . It does not mean that apart from those
circumstances it makes (clear) sense. The point is not that you sometimes cannot
say (or think) what is the case, but that to say (or think) something is the case
you must say or think it, and “saying that” (or “thinking that”) has its conditions. 26

Apart from specifying circumstances in which it makes sense to say a

certain sequence of words we have failed determinately to specify what
it is that the sequence of words says. Nonsense, according to the view
that Cavell attributes to Wittgenstein, arises when there is an absence
of sense. The view that Williams and McGinn attribute to Cavell takes
philosophical nonsense to be due to an inappropriate kind of presence of
sense; it takes philosophical nonsense be one of two sorts of cases
described by Cavell in the following parenthetical remark:
We are, one might say, asked to step back from our conviction that this must
be an assertion . . . and incline ourselves to suppose that someone has here been
prompted to insistent emptiness, to mean something incoherently.
(This is not the same as trying to mean something incoherent. Wittgenstein
alludes to this possibility in saying “When a sentence is called senseless, it is
not as it were its sense that is senseless.” 27 Nor is it the same as meaning
something other than you think. This would describe cases in which your
words make sense, and they are put together correctly, but you are as it were
meaning them in the wrong place.) 28

Cavell parenthetically mentions both of these sorts of cases precisely in order to

distinguish them from the sort of case that he takes to be at issue in Wittgenstein’s
employment of “nonsense” as a term of philosophical criticism. Let’s first
consider the second of the cases mentioned here. This is not a case of sheer
unintelligibility. The intelligibility of such a case, however, depends upon our
seeing the utterer as imagining himself to be in circumstances other than those
in which he actually is. The first sort of case—the one mentioned in the preceding
portion of the passage—is the sort Wittgenstein himself attempts to characterize
when he says:

When a sentence is called senseless, it is not as it were its sense that is

senseless. But a combination of words is being excluded from the language,
withdrawn from calculation.29

Wittgenstein’s saying that a certain combination of words is to be

“excluded from the language” or “withdrawn from circulation” might
suggest that he takes the combination of words in question to have an

THE HARVARD REVIEW OF PHILOSOPHY vol.XIII no.1 2005

 Stanley Cavell’s Wittgenstein 63

intrinsically flawed sense—to be nonsense not because of an absence of sense,

but because of the flawed presence of sense. The preceding section of Philosophical
Investigations begins:

To say “This combination of words makes no sense” excludes it from the

sphere of language and thereby bounds the domain of language. But when one
draws a boundary it may be for various kinds of reason.30

This raises the question: what are Wittgenstein’s reasons for proposing that we
exclude particular combinations of words from the language? In Philosophical
Grammar, we find this:

How strange that one should be able to say that such and such a state of affairs
is inconceivable! If we regard a thought as an accompaniment going with an
expression, the words in the statement that specify the inconceivable state of
affairs must be unaccompanied. So what sort of sense is it to have? Unless it
says these words are senseless. But it isn’t as it were their sense that is senseless;
they are to be excluded from our language as if they were some arbitrary
noise, and the reason for their explicit exclusion can only be that we are
tempted to confuse them with a proposition of our language.31

Wittgenstein’s reasons for proposing that we explicitly exclude an

expression from the language are—not because it is as it were the sense
of expression which is senseless, but—because “we are tempted to confuse”
sentences in which it figures senselessly with meaningful propositions
of our language. Saying that “it is as it were its sense that is senseless” is
not meant to be a description of a possible sort of case of nonsense. Rather
it is meant as a description of a sort of case that we imagine we come
upon in philosophizing. Wittgenstein thinks there is a conception of
nonsense that, in philosophizing, we find it all but impossible to avoid
falling for. One way of falling for this conception is to think that a
proposition is nonsensical because its parts are illegitimately combined—
another way of falling for it is to think that a con tent and a context cannot be
combined (one “cannot” utter these words in this context). Already in Tractatus
Logico-Philosophicus, Wittgenstein set his face against such a conception of, as
it were, substantial nonsense. It is à propos such a conception of nonsense that
he writes in the Tractatus: “we cannot give a sign the wrong sense.” 32
Most commentators on Wittgenstein’s work (early and late) take his
deployment of “nonsense” as a term of philosophical criticism to represent the
conclusion of an argument to the effect that certain combinations of
expressions—or the employments of certain combinations of expressions in
certain contexts—are inherently nonsensical. If you are a scholar of
Wittgenstein’s early work, you are likely to think that the trouble is to be traced
to so-called “violations of logical syntax” (that is, to the logical incompatibility
of the parts of the proposition). If you are a scholar of his later work, you are
likely to think it is to be traced to so-called “violations of grammar” (which
sometimes means the same thing as “violations of logical syntax” and sometimes
means “the incompatibility of certain meanings with certain contexts of use”).
But, according to Cavell, what the early Wittgenstein calls the logic of our

vol.XIII no.1 2005 THE HARVARD REVIEW OF PHILOSOPHY

64 James Conant

language and what the later Wittgenstein calls grammar is not the name for a
grid of rules we lay over language in order to point out where one or another of
its prescriptions are violated. A grammatical investigation is a convening of our
criteria for the employment of a particular concept. But the way an appeal to
criteria comes to bear on a philosophical problem, such as that of skepticism, is
not by showing the philosopher that he has “violated the rules for the use of an
expression,”33 and therefore that there is something determinate that he wants
to mean that he cannot mean by his words.
With respect to the skeptic, the point of (what Wittgenstein calls) a
“grammatical investigation” is to show him that he is faced with a dilemma:
either he stays within our language-games and his words express a doubt but
not the sort of super-doubt that he is after (his doubt will not generalize in the
way that he needs it to in order to bring the possibility of knowledge as such in
doubt), or he will be led to speak “outside language-games,” stripping his
putative context of use of the concrete specificity (and hence the foothold for
our criteria) which permits us to mean and what we do on the occasions on
which we ordinarily employ the word ‘doubt’ to express the concept of doubt.
No rule of grammar is adduced to exhibit ineradicable (logical or grammatical)
flaws in the skeptic’s utterances. Rather the grammar of our various language-
games is exhibited to the skeptic, in order to present him with an Übersicht of
the various possibilities of meaning his words that are available to him. He is to
find, once presented with an übersichtliche Darstellung of the grammar, either
that he is making perfect sense but failing to ask the question he wants, or that
it remains unclear which of the many things he can mean by his words he
wants to mean. Wittgenstein’s aim, in assembling these reminders, is not to
refute the skeptic (that is, to establish the truth of the negation of what he claims),
but to query the sense of his claim: to force on him the question, given what his
words can mean, what he means by them. The problem with his words thus lies
neither in the words themselves nor in some inherent incompatibility between
his words and a determinate context of use, but in his confused relation with
respect to his words. The aim of a Wittgensteinian grammatical investigation,
according to Cavell, is to furnish the philosopher with a perspicuous
representation of the various things he might mean by his words in order to
show him that, in wanting to occupy more than one of the available alternatives
at once and yet none in particular at a time, he is possessed of an incoherent
desire with respect to his words. ϕ

Notes
1
Ludwig Wittgenstein, Philosophical Investigations, trans. G. E. M. Anscombe (Oxford:
Blackwell,1952), §43.
2
Stanley Cavell, The Claim of Reason (Oxford: Oxford University Press, 1979), pp. 206-
207. All subsequent references to Cavell are to this book.
3
Cavell, p. 221.
4
Philosophical Investigations, §134.
5
Philosophical Investigations, §243.
6
“Wittgenstein does not say that there can be no private language. He introduces his
sequential discussion of the topic, at §243, by asking: ‘Could we also imagine…’ The upshot

THE HARVARD REVIEW OF PHILOSOPHY vol.XIII no.1 2005

 Stanley Cavell’s Wittgenstein 65
of this question turns out to be that we cannot really imagine this, or rather that there is
nothing of the sort to imagine, or rather that when we as it were try to imagine this we
are imagining something other than we think.” Cavell, p. 344.
7
See Cavell, pp. 133-145.
8
Cavell, p. 218.
9
Cavell, pp. 164-5.
10
“The confusions which occupy us arise when language is like an engine idling, not
when it is doing work.” Philosophical Investigations, §132.
11
“[P]hilosophical problems arise when language goes on holiday.” Philosophical
Investigations, §38.
12
Cavell, pp. 205-6.
13
Ludwig Wittgenstein, On Certainty, trans. G. E. M. Anscombe (Oxford: Blackwell,
1969), §349.
14
Michael Williams, Unnatural Doubts: Epistemological Realism and the Basis of Scepticism
(Cambridge, MA: Blackwell, 1991), pp. 152, 151.
15
Marie McGinn, Sense and Certainty: A Dissolution of Scepticism (New York: Blackwell,
1989), pp. 85-6. All subsequent references to McGinn are to this book.
16
Paul Grice, Studies in the Way of Words (Cambridge, MA: Harvard University Press,
1989). For an excellent discussion of the differences between Wittgenstein and Grice, to
which the ensuing discussion is indebted, see Charles Travis, “Annals of Analysis,” Mind,
vol. C, no. 2 (April 1991): pp. 237-264.
17
McGinn, p. 85.
18
On Certainty, §348.
19
Philosophical Investigations, §117.
20
Philosophical Investigations, §514.
21
On Certainty, §352.
22
On Certainty, §350.
23
This example is due to Charles Travis.
24
On Certainty, §349.
25
Cavell, p. 210.
26
Cavell, p. 215.
27
Cavell is here quoting Philosophical Investigations, §500.
28
Cavell, p. 336.
29
Philosophical Investigations, §500.
30
Philosophical Investigations, §499.
31
Ludwig Wittgenstein, Philosophical Grammar, ed. Rush Rhees, trans. Anthony Kenny
(Oxford: Blackwell, 1974), p. 130.
32
Tractatus Logico-Philosophicus, §5.4732.
33
I am quoting here from the following passage from Baker and Hacker which identifies
as a central continuity in doctrine in the early and later Wittgenstein what Cavell takes
to be a central continuity in target:

Wittgenstein had, in the Tractatus, seen that philosophical or conceptual investigation

moves in the domain of rules. An important point of continuity was the insight
that philosophy is not concerned with what it true and what is false, but rather
with what makes sense and what traverses the bounds of sense. . . . [W]hat he
called ‘rules of grammar’ . . . are the direct descendants of the ‘rules of logical
syntax’ of the Tractatus. Like rules of logical syntax, rules of grammar determine the
bounds of sense. They distinguish sense from nonsense. . . . Grammar, as
Wittgenstein understood the term, is the account book of language. Its rules
determine the limits of sense, and by carefully scrutinizing them the philosopher
may determine at what point he has drawn an overdraft on Reason, violated the rules
for the use of an expression, and so, in subtle and not readily identifiable ways, traversed
the bounds of sense. (G. P. Baker and P. M. S. Hacker, Wittgenstein; Rules, Grammar and
Necessity [New York: Blackwell, 1985], pp. 39-40, 55 [their emphasis])

vol.XIII no.1 2005 THE HARVARD REVIEW OF PHILOSOPHY