G. 17. BAKER AND P. M. S. HACKER
CRITICAL STUDY:
ON MISUNDERSTANDING WITTGENSTEIN:
KRIPKE'S PRIVATE LANGUAGE ARGUMENT
]. "WITTGENSTEIN'S ARGUMENT AS IT STRUCK KRIPKE"
After a long period of neglect, Wittgenstein's discussions of rulefollowing
have, in the last few years, received some serious attention.
This has been ,atimulated partly by a growing interest in his philosophy
of mathematics, and partly by the publication of the enlarged edition of
his Remarks on the Foundations of Mathematics, which includes a new
fifty-page section on rule-following. Perhaps the most important stimulus,
however, is the conviction among many philosophers that the
confrontation between realism and anti-realism, between truthconditional
semantics and semantic theories involving the notion of
assertion-conditions, is the fundamental issue in contemporary philosophy.
Accordingly, the early Wittgenstein is strapped to the truthconditions
bandwagon, and the later Wittgenstein, strait-jacketed within
the confines of anti-realism, is harnessed to the assertion-conditions one.
Since his remarks on rules have a clear bearing on issues that interest
participants in this confrontation, they have become the focus of
extensive discussion.
Saul Kripke's essay 'Wittgenstein on Rules and Private Language' 1
applies this currently popular picture of Wittgenstein's early and later
work to a reconsideration of the famous private language argument in
the Philosophical Investigations §§243 ft. The discussion of rule-following
that precedes the private language argument (PI §§143-242) is the
focal point of his examination, and from it he draws a variety of original
and controversial conclusions. Rather disarmingly, he suggests at the
outset that
the presentpapershouldbe thoughtof as expoundingneither'Wittgenstein's'argument
nor 'Kripke's':rattlerWittgenstein'sargumentas it struckKripke.2
To use the writings of a philosopher as a Rorschach spot is perfectly
legitimate. But there is an ever-present danger that one's ruminations
will be taken as descriptions of the Spot. Indeed, in the course of his
Synthese58 (1984)407-450. 0039-7857/84/0583-0407 $04.40
 1984by D. ReidelPublishingCompany
408 G. P. BAKER AND P. M. S. HACKER
reflections Kripke attributes to Wittgenstein a variety of views he never
held, and imposes upon his writings a variety of interpretations for
which there is no licence. In this paper we shall try to differentiate
sharply between Wittgenstein's argument as it struck Kripke and
Wittgenstein's argument, and to demonstrate that on some salient issues
Wittgenstein's argument not only differs from, but actually confutes
Kripke's picture.
More than mere exegetical correctness is involved here. Kripke is
surely right in thinking that sections 134-243 of the Investigations
contain some of the most original and significant philosophical
reflections written this century. If the line of argument pursued in them
is valid, their implications, both within philosophy and without, are
considerable. Modern philosophical logic, theoretical linguistics, as
well as branches of empirical psychology would stand in need of
re-evaluation. So it is important to understand what Wittgenstein was
arguing. Only then can we assess it, and see where we stand.
Kripke's interpretation of the core of the Philosophical Investigations
is as follows. The 'real private language argumenť is not in §§243 ff, but
in §§143-242. Indeed the conclusion of the private language argument
is stated in §202:
And hence also 'obeying a rule' is a practice. And to think one is obeying a rule is not to
obey a rule. Hence ,it is not possible to obey a rule 'privately': otherwise thinking one was
obeying a rule would be the same thing as obeying it.
The problem Wittgenstein confronts in these hundred sections is,
Kripke claims, a sceptical one. The discussion of the rule for forming
the series of even integers is designed to raise the question of how I can
know whether my current use of a word (e.g., 'plus') coheres with what I
previously meant by it, given that my current use is (or can always be
made out to be) a novel application. Nothing in any instructions given
to me (or which I give myself) forces me to go on '1002, 1004', rather
than '1004, 1008'; the instructions I gave, the examples I produced, can
be made out to be consistent with both ways of proceeding. Equally,
nothing in my mind constituted the fact of my meaning myself to go on
thus or otherwise. (I didn't run through an infinite series in an instant
and the formula I had in mind has no magical powers to generate the
answer.) Scepticism about being able to know whether I am using a
word in accord with what I meant by it leads to the paradox stated at
§201:
ON MISUNDERSTANDING WITTGENSTEIN 409
This was our paradox: no course of action could be determined by a rule, because every
course of action can be made out to accord with the rule. The answer was: if everything
can be made out to accord with the rule, then it can also be made out to conflict with it.
And so there would be neither accord nor conflict here.
So epistemological scepticism about applying a word in accord with
what one means by it leads to the conclusion that there can be no
meaning at all, and language is impossible. This paradox is 'perhaps the
central problem of Philosophical Investigations'. 3
Wittgenstein, according to Kripke, gives a 'sceptical solution', of a
Humean form,, to his sceptical problem. This strategy consists in
accepting the sceptics premises but denying that the sceptical conclusion
follows from them. So Wittgenstein agrees that there is no
fact-in-the-world that constitutes meaning something by one's words.
But this annihilates the possibility of meaning only on the assumption
that sense is given by truth-conditions (i.e., correspondence to possible
facts-in-the-world). He allegedly repudiates this theory, replacing it by
the picture of sense as determined by conditions for assertion. The
assertion-conditions for my meaning W by' W' are my being inclined to
apply 'W' thus and so, given that the rest of the community is too. In
these circumstances there is no reason to deny that my current ('noveľ)
application of' W' accords with what I (and others) previously meant by
' W'. Therefore, meaning something by a word requires a community to
supply agreement and to prevent thinking one is following a rule and
following a rule from collapsing into each other. Hence 'it is not
possible to obey a rule "privately"', and the conclusion of the private
language argument is really stated before what goes by the name of 'the
private language argumenť has even begun. This does not, however,
preclude our conceiving of Robinson Crusoe meaning something by his
words as he talks to himself on his desert island. For in so conceiving of
him 'we are taking him into our community and applying our criteria for
rule following to him.' 4 A physically isolated person can follow rules,
but a person considered in isolation cannot.
This sketch provides the bare outline of Kripke's colourful painting.
We shall fill in more detail only where necessary.
2. SCEPTICISM ABOUT WITTGENSTEIN'S 'SCEPTICAL
PROBLEM' AND 'SCEPTICAL SOLUTION'
Although Wittgenstein found some of the problems of scepticism
410 G. P. BAKER AND P. M. S. HACKER
interesting, even revealing (as in On Certainty), anyone familiar with
Wittgenstein's work will feel sceptical about the suggestion that he
made a sceptical problem the centre piece of his chef ďoeuvre. In his
very first philosophical notes he wrote against Russell:
Scepticism is not irrefutable, but obvious nonsense if it tries to doubt where no question
can be asked.
For doubt can only exist where a question exists; a question can only exist where an
answer exists, and this can only exist where something can be said. s
In his last notes on certainty he wrote:
The queer thing is that even though I find it quite correct for someone to say
"Rubbish!" and so brush aside the attempt to confuse him with doubts at bedrock, nevertheless,
I hold it to be incorrect if he seeks to defend himself (using, e.g. the words "I
know"). 6
It would be very surprising to discover that someone who throughout
his life found philosophical scepticism nonsensical, a subtle violation of
the bounds of sense, should actually make a sceptical problem the
pivotal point of his work. It would be even more surprising to find him
accepting the sceptic's premises, the 'doubts at bedrock', rather than
showing that they are 'rubbish'.
Initial qualms may be strengthened by reflection on the oddity of the
so-called scepticism. What is classically known as scepticism typically
involves challenging an apparent evidential nexus. The sceptic agrees
that we do know the truth of statements about subjective experience,
but, since they do not entail statements about objects, he denies that we
really know anything about the material world. In a more obliging
frame of mind, he accepts the possibility of knowledge about the
behaviour of others (or about memories and current evidence, or
singular statements) but denies that it supports cognitive claims about
other minds (the past, inductive generalizations). But Kripke's sceptic,
unlike the classical sceptic, saws off the branch on which he is sitting.
For he is not claiming that certain given knowledge fails to support
other commonly accepted cognitive claims. His conclusion is not that
he certainly means either a or b by ' W', but cannot be sure which; nor
is it that he knows what he now means, but cannot be certain whether it
is the same as what he meant yesterday. Rather, he concludes with 'the
paradox' that there is no such thing as meaning, so language cannot be
possible. But this is not scepticism at all; it is conceptual nihilism, and,
ON MISUNDERSTANDING WITTGENSTEIN 411
unlike classical scepticism, it is manifestly self-refuting. Why his
argument is wrong may be worth investigating (as with any paradox),
but that it is wrong is indubitable. It is not a sceptical problem but an
absurdity.
To defend common sense presumably means to find good reasons
why what we ordinarily and more or less unreflectively believe is true,
and is known to be true. Kripke's Wittgenstein 7 is a common-sense
philosopher, holding that philosophy only states what everyone admits.
He resembles ttume, who wrote "We may well ask, what causes induce
us to believe in the existence of body? but it is vain to ask, whether there
be body or not? That is a point, which we must take for granted in all our
reasonings." 8 The similarity with Hume allegedly reaches deeper, since
Kripke's Wittgenstein pursues a Humean strategy of giving a 'sceptical
solution' to his sceptical problem, i.e., he concedes that the sceptic's
negative assertions are unanswerable, but contends that our ordinary
belief is nevertheless justifiable, because it does not require the
justification the sceptic has shown to be untenable. The switch from
truth-conditional semantics to assertability conditions is argued to effect
just this move.
This analysis is wrong on several counts. First, Hume does not, in his
analysis of the self, objectivity, and causation, defend common sense.
He is not claiming that philosophy never casts doubt on the rational
justification of ordinary beliefs, but that philosophy is impotent to
change them. What Kripke quotes follows this passage:
Nature, by an absolute and uncontrollable necessity has determin'd us to judge as well
as to breathe and feel... Whoever has taken the pains to refute the cavils of this total
scepticism, has really disputed without an antagonist, and endeavour'd by arguments to
establish a faculty, which nature has antecedently implanted in the mind, and rendered
unavoidable...
* . . If belief, therefore, were a simple act of thought without any peculiar manner of
conception, or the addition of a force and vivacity, it must infallibly destroy itself and in
every case terminate in a total suspense of judgment. But as experience will sufficiently
convince anyone who thinks it worthwhile to try, that though he can find no error in the
foregoing arguments, yet he continues to believe and think and reason as usual, he may
safely conclude, that his reasoning and belief is some sensation or peculiar manner of
conception which 'tis impossible for mere ideas and reflections to destroy.
Hume insists that we have no good reason to believe in the existence of
objective particulars (no good reason to believe that our several
perceptions belong to a unitary self; etc.). Far from defending 'common
412 G. P. BAKER AND P. M. S. HACKER
sense', he insists that all the arguments are on the side of the sceptic.
Rational investigation proves that the fundamental beliefs of 'common
sense' are fictions, generated by the workings of the imagination
according to natural laws of mental association.
Second, Hume's 'sceptical solution' does not consist in giving a
justification for our ordinary beliefs, i.e., rational grounds showing them
to be well-founded. Rather, he defends a radical split between theory
and practice, between Reason and Nature, in the following sense.
Though every argument speaks against belief in objectivity, this cannot
affect our beliefs. "Philosophy would render us entirely Pyrrhonian were
not nature too strong for it.''9 He denied the Pyrrhonist thesis that
sceptical arguments will lead to suspension of belief and ataraxia.
Belief is not determined by Reason, but by Nature. Hume was indeed
not trying to subvert our beliefs, but to show that they are determined,
non-rationally, causally, by Nature against Reason. 1°
Third, not only is Hume thus misrepresented, but so also is Wittgenstein.
He insisted that he was not defending any opinions: 'On all
questions we discuss I have no opinion; and if I had, and it disagreed
with one of your opinions, I would at once give it up for the sake of
argument, because it would be of no importance for our discussion.' 11
Nor was he defending common sense, if that means giving reasons for
believing that material objects exist independently of our perceptions of
them, or that other people enjoy experiences, or that we mean things by
our words, etc. Rather the task, in this respect, consists in showing that
the philosophical puzzle (of scepticism, idealism or solipsism) rests on
systematically traversing the limits of sense. His purpose was the
investigation of ordinary concepts which are used in the expression of
common sense beliefs. But he was not concerned to defend those
beliefs, rather to clarify those concepts. For the problems of philosophy
arise through the distortion and misuse of ordinary concepts, and the
way back to sanity consists in obtaining an OYoersicht of the problematic
expressions.
Finally, Kripke contends that in the course of his 'sceptical solution'
to his 'sceptical problem' Wittgenstein in effect denies some of our
ordinary beliefs, contrary to his principle that 'If one tried to advance
theses in philosophy it would never be possible to debate them, because
everyone would agree to them.' 12 Wittgenstein is forced to do this
because he accepts the apparent sceptical denials of our ordinary
assertions, and he only saves himself from blatant inconsistency by
ON MISUNDERSTANDING WITTGENSTEIN 413
'cagily' refusing to state his conclusions in the form of definite theses or
straightforward formulations. It is, in fact, this caginess that explains his
inability to write a work with conventionally organized arguments.
What 'thesis' is Wittgenstein supposed to be propounding here? According
to Kripke, ~Wittgenstein holds, with the sceptic, that there is no
fact as to whether I mean plus or quus.' ~3
This is quite wrong. If Wittgenstein had claimed that when A told B
to expand the series '+2', A did not mean B to go on '1002, 1004...',
he would be denying what we all admit. But he does not claim this:
'Certainly; and you can also say you meant it then; only you should not
let yourself be misled by the grammar of the words "know" and
"mean".' 14 And if A meant such and such, then, Wittgenstein might
add, it is a fact: that he meant such and such (infra). What Wittgenstein
is denying is a philosophical claim, viz., that the 'act of meaning' effects
miracles, such as traversing an infinite series in a flash, and a philosophical
thesis, viz., that my meaning such and such is a fact-in-the-world
(or more specifically, a fact-in-my-mind), and that my justification for
saying that I meant addition by 'plus' is that I have observed this fact in
my mind. To deny this is not to deny what we all admit, but to repudiate a
nonsensical metaphysical theory.
The considerations so far are merely intended to foster initial doubts
about Kripke's interpretation. It is, we suggest, prima facie implausible
(though not impossible) that Wittgenstein should find sceptical problems
and Humean sceptical solutions the fountainhead of philosophical
insight. But there is a far greater likelihood that he would concur with
Dr. Johnson's ironic remark about sceptics:
Truth, sir, is a cow, which will yield such people no more milk, and so they are gone to
milk the bull.
3. THE 'PARADOX'
The interpretation of Investigations.§§201-202 is crucial for Kripke's
case. For the 'paradox' of §201 is 'perhaps the central problem' of the
whole book, and the conclusion of §202 'it is not possible to obey a rule
"privately" ' is not an anticipation of the private language argument, ~s
but a statement of its conclusion, based on arguments already given. By
§243 the impossibility of a private language has already been proved.
The core problem is, according to Kripke, a normative version of
Goodman's 'new riddle of induction'. No past fact about my mind or my
414 G. P. BAKER AND P. M. S. HACKER
behaviour constituted my meaning W by' W', so nothing in my present
use of ' W' can constitute accord (or conflict) with what I meant by ' W'
(the meaning I assigned to 'W'). So I cannot know that in my current
use I am still using ' W' with the same meaning. But not even God, were
He to peer into my mind could know this. So there is no such thing as
using a word in accord with a rule (with the meaning one gave it), no
such thing as meaning something by a word, no such thing as a
meaningful language. This is the 'paradox'.
There are reasons for doubting whether §§201-2 are the pivotal
remarks of the book. Also reasons for doubting that they are the
culmination of a sceptical argument, or that they incorporate a refutation
of the possibility of a private language. And there are alternative
interpretations of the remarks which rest on better exegetical evidence.
Finally, there are doubts whether Kripke's 'problem' is coherently
stateable. These considerations strengthen the doubts sown in the last
section.
The history of §§201-2: The Philosophical Investigations Part 1 went
through four typescript stages, dated 1938 (TS 220 in von Wrighťs
catalogue), 1942-43 (TS 239), January 1945 (the 'Intermediate Version'),
and 1945-46 (the final version, TS 227)J 6 The Intermediate
Version is almost identical with the final version in respect of argument
from §1 to §217, but it does not contain §§201-3. So although the
Intermediate Version contains the whole of the argument prior to
§§201-3, it does not incorporate the 'conclusion' of the argument
which, according to Kripke, is the pivotal point of the book. Moreover,
the Intermediate Version contains the bulk of the private language
argument,17 and if §§201-3 already establish the conclusion of the
private language argument, and §243 ft. only examine a potential
counter example to the argument concerning sensations, it is odd, to say
the least, that Wittgenstein omitted the puncture saliens in this draft.
Further light is shed on the matter by the immediate MS source, MS
129, pp. 119tt., which was written from 17 August 1944 onwards.
Equally illuminating is the more remote MS source, the pocket notebook
MS 180(a), which contains material transcribed into MS 129.
In MS 129 the context of the occurrence of §§201-3 is a discussion of
how I know that this is red, i.e., what is now PhilosophicalInvestigations
§§377-81 (which does not occur in the Intermediate Version). The
argument is concerned to establish that judgments such as 'These
images are identicaľ or 'This is reď do not rest upon recognition. It runs
ON MISUNDERSTANDING WITTGENSTEIN 415
as follows: How do I know that this is red? One is inclined to say - I look
at it and see that it is red. But how does this wordless seeing-that-it-isred
help me iif I do not know what to say, how to express this
'recognition' in words? And sooner or later I must make the transition
to using an expression. And at this point rules leave me in the lurch.
What does that mean? It means 18 that rules are not self-applying. That
sooner or later I have to apply a rule (by themselves as it were, they
hang in the air). All good teaching, at the end of the day will not help
me in this respect, for it cannot relieve me of the task of applying the
rule, of 'making a leap' and saying 'This is reď, or acting in a certain
manner, which comes to the same thing. It is of no avail to try to
interpose, between looking at an object and saying that it is red, a bogus
recognitional stage of 'seeing that it is this', for now one would need a,
rule to effect the transition from this 'seeing that it is this' to saying it is
red. 19But this transition would be a 'private' one and the only rule which
could guide it would be a private ostensive definition determining what
this is. Yet justification must be public (cf. PI §378).
At this point in MS 129, 119 we have what is now §201 ('This was our
paradox...'), which originally began 2° 'I found myself in the difficulty
that it seemed...', which clause Wittgenstein crossed out and replaced
by 'This was our paradox.' This clearly refers back to what is now §198,
which occurs earlier in the MS (p. 25). Why this allusion to these earlier
difficulties at this point? Because the issue is similar, and the conclusion
of §198 can be brought to bear on the present problem. §198 argued
that the connection between the expression of a rule and the act that
counts as accord with the rule is forged, inter alia, by training. We are
taught that acting thus in response to such-and-such is correct, and
anything else incorrect. We explain and justify this, but not another
thing, by reference to the rule; and so on. To follow a rule is a custom; it
involves a regular use of the expression of rules in training, teaching,
explaining, and in giving reasons. Now Wittgenstein embroiders on
these observations in the new context of a discussion of recognition and
the unmediated application of 'reď and 'same' (as well as 'same image'
and 'red image'). The 'earlier difficulty' that a rule could not determine
what to do in accord with it, the 'paradox', was evidently a misunderstanding.
This is shown by the fact that no interpretation, no rule for the
application of a rule can satisfy us, can definitively fix, by itself, what
counts as accord. For each interpretation generates the same problem,
viz., how is it to be applied. 21 Now Wittgenstein adds a new point to the
416 G. P. BAKER AND P. M. S. HACKER
argument of §198, deepening the implications of the claim that what
counts as following a rule is fixed by a normative regularity. What the
absurd paradox that rules cannot guide one shows is that how one
understands a rule need not be an interpretation, but is manifest in
acting, in what we call 'following the rule'. (The draft of §201 here is
much the same as the final version.)22 That we have 'understood a rule
in a certain way', Wittgenstein continues, ('Dass wir eine Regel
"aufgefasst" haben') is shown inter alia in the certainty, the absence of
fumbling, in its application. Then follows Investigations §§202-3, i.e.,
that is why 'following the rule' is a practice. How does this observation
bear on the discussion of colour recognition? It is simple: there can be
no rule guiding the transition from a seeing-that-this-is-so to saying
'This is reď because there could be no technique of application for such
a 'rule', no normative regularity in its employment, no custom of
applying the word always in the same way, no practice (infra). The
'private' following of the (apparent) rule, a mental ostensive definition
presupposed by the supposition that recognition intervenes between
seeing and saying, is a sham, in which following a rule and thinking one
is following a rule collapse into each other. The sequel continues to
probe what is wrong with talking about judging an object to be red on
the grounds of recognition, and develops into a discussion of seeing
aspects.
MS 180(a) is the source underlying MS 129 on this theme. The longer
discussion there begins with an examination of aspect-seeing (pp. 52ff.):
do I interpret the figure now thus, now otherwise? Do I see everything
always as something? Do I need words for such 'visual interpretation'
or are words only necessary to communicate what I see? This evolves
into a discussion of my judgments about my visual images, and thence
into a long investigation into how I know that this is red? The key
theme here too is to repudiate the suggestion that any recognitional
process mediates between looking, and saying that this is red. Here
occur early drafts of Investigations §§378-81 (in a different order). The
draft of §201 (on p. 72) occurs after an examination of the following
sequence of propositions: I see that it is red, but don't know what it is
called! I see that it is this colour, and I know that this colour is called
such and such. Which colour? I recognise that it is this. But now I must
make the transition to words or deeds! After a draft of §201 Wittgenstein
elaborates: the rules here leave us in the lurch because there is
no (genuine) transition from seeing that this is this to seeing that it is
ON MISUNDERSTANDING WITTGENSTEIN 417
red, there is no technique of going by a rule here. The 'rules' are
free-floating. For this (pseudo-) transition is a private one. If the
transition from looking at an object to applying the word 'reď cannot
be made without mediation (by a recognitional process, or a private
ostensive definition) then it cannot be made by means of a rule either.
Consequently, 'to follow a rule' designates a practice, which cannot be
replaced by the bogus appearance of a practice (cf. §202). This seems to
abolish logic, he wrote in an early version of §242, but it does not. It is
one thing to lay down methods of measurement, another to obtain
results of measurement. But what we call 'measuring' is partly determined
by a certain constancy in results of measurement.
What tentative conclusions can we draw from these data? First, the
history of §§201-3 suggests that these remarks are not the pivot of the
whole book. They were not incorporated in the Intermediate Version,
which contained in completed form both the argument preceding
§§201-3, and the bulk of the private-language argument. Secondly, in
their original context they quite explicitly build upon §198 and upon
the senselessness of private ostensive definition, as established by the
private-language argument. Thirdly, their original purpose was to
deepen the insight of §198 and to bring it to bear on fallacies
concerning recognition as mediating between saying and seeing. It was
not to defend a new paradox, viz., that there can be no such thing as
following a rule. Fourthly, the MS context has nothing to do with
scepticism in any shape or form, neither with sceptical problems nor
with sceptical conclusions. (No where is it suggested that I do not know
that poppies are red!) There is no evidence to suggest that Wittgenstein
was concerned with a normative version of Goodman's new problem of
induction. Fifthly, in its original contexts in both MSS the remark that
following a rule is a practice has nothing directly to do with social
practices. Its exclusive concern is with the fact that rule-following is an
activity, a normative regularity of conduct which exhibits one's
Auffassung of a rule, manifests how one understands a rule.
Do these sections incorporate the conclusion of the private language
argument? That depends upon what one takes this conclusion to be! If it
is held to be the contention that it is not possible to obey a rule
'privately', then, of course, it does. But that is unsurprising. More
interestingly, in context, these sections do incorporate at least some of
the conclusions of the private language argument, since they explicitly
presuppose them as having already been established. And rightly so!
418 G. P. BAKER AND P. M. S. HACKER
For the moment we claim not that this informationis decisive, only that
it is suggestive. It is possible, but improbable, that the repositioning of
these remarks betokens a total reorientation of their sense. Perhaps, after
composing the Intermediate Version, Wittgenstein suddenly realized
that these two MS remarks, embedded in a discussion of knowing that
this is red, concerned with dissolving confusions about recognition, in
fact contained in crystallized form the core of his book.
The interpretationof §§201-2: A closer look at these crucial remarks
in their final, different, context may strengthen our doubts. §201
paragraph (a) is clearly concerned with the question raised in §§198(a),
viz., '"But how can a rule show me what I have to do at this point.
Whatever I do can, on some interpretation, be brought into accord with
the rule."' The problem here posed belongs to a series of closely
interwoven questions running through §§139-242. If the first part of the
book can be said to be concerned with uprooting the Bedeutungsk6rper
conception of meaning in all its forms, this part is concerned with the
more subtle Regelsk6rper conception. It aims to break the hold of a
misguided conception of rules as mysteriously, magically, determining
or constituting the meanings of expressions, and of understanding as a
grasping of rules which then guide us along predetermined rails. The
problem setting context is as follows. Earlier, Wittgenstein has argued
that the meaning of an expression is its use. Equally the meaning of an
expression is what is given by an explanation of meaning (§75); and an
explanation of meaning is a rule for the use of the expression. But how
can such an explanation, e.g., an ostensive definition or a series of
examples or a general formula, determine the complex use of an
expression? For any rule can be variously interpreted. The statement of
a rule is not a repository from which the use unfolds, or a logical
machine that generates applications of its own accord.
With respect to understanding, which is the correlate of explanation,
the problem presents itself thus. We typically understand an expression
in an instant. We can say what we understand by giving an explanation.
But how can what is understood, grasped in an instant, be something
like the use, the pattern of application of an expression, that is spread
out in time (cf. §138)? And if what is understood is expressed by an
explanation, a rule, how does that explanation, which can be variously
projected, guide one in how one uses an expression? The apparent
tension within these claims is the concern of §§143-242. This
discussion falls into two main parts. The first (§§143-84) 23 explores the
ON MISUNDERSTANDING WITTGENSTEIN 419
concept of understanding (with a long digression on reading)24 and
establishes that understanding is not a mental event, state or process.
Rather, to say of a person that he understands a word is to characterize
him as having, at a particular time, a capacity, a mastery of a technique.
Understanding is akin to an ability. The second part (§§185-242)
clarifies the notions of an acťs conforming (or conflicting) with a rule,
and of an agenťs following a rule. By particularization, this elucidates
what it is for the use of a word to be correct (to conform to its
explanation), and what it is to mean something and to understand an
expression.
After exple,ring two suggestions of how accord with a rule is
determined, each of which was unsatisfactory for quite different
reasons, Wittgenstein turns to face this issue head-on in §198. If
whatever one does can be brought into accord with the rule on some
interpretation, 25 how on earth can a rule guide one? (Note that this is
the identical problem of §201(a)). The answer is given immediately,
with no suggestion of an irresoluble paradox that needs to be sidestepped.
We ought not to say that because whatever we do can be brought
into accord with the rule on some interpretation, therefore the rule
cannot guide us. That would be absurd. Rather
any interpretation still hangs in the air along with what it interprets, and cannot give it any
support. Interpretations by themselves do not determine meaning.
Only in a context in which there is an established technique of
application of a rule, in which the rule is standardly involved in
explanation and justification, in teaching and training, can questions of
giving interpretations arise. For only then is the expression used, and an
internal relation established between act and rule. Only if there are
genuine rules, only if something does actually count as following (and
everything else as going against), is there room for interpreting a rule
correctly or incorrectly. But then what counts as accord with the rule is
fixed independently of interpretations.
§§199-200 emphasize that normative behaviour requires a multiplicity
of occasions as a context, an evident regularity of point and
purpose. §201, removed from its MS context, is stripped of any
connection with the problem of colour judgment, image-identification,
interpreting double-aspect figures. Whereas it originally applied the
resolution of tile question of §198 to puzzles about recognition, in the
course of which it deepened the argument of §198, now its sole role is
420 G. P. BAKER AND P. M. S. HACKER
just the latter. §§198-9 answered the question: 'How can a rule
determine what counts as accord with it?' by reference to the existence
of a normative regularity of conduct. §201 adds a crucial point about
understanding a rule (an 'Auttassung' of a rule):
how one understands a rule need not be an interpretation, but may be exhibited in what
we call 'following the rule' and 'going against iť from case to case of its application, z6
Understanding is mastery of a technique, and how one understands a
rule is manifest in the exercise of that technique in practice, in what
one does in various cases. Far from §201 accepting a paradox and
by-passing it by means of a 'sceptical solution', Wittgenstein shows that
here, as elsewhere, a paradox is a paradox only in a defective surrounding.
If this is remedied the appearance of paradox will vanish,z7For every
paradox is disguised nonsense (and this one is barely even disguised!).
Hence it may never be accepted and by-passed by other arguments. It
must be dissolved by clarification of concepts.
What has been rejected in §201 is not the truism that rules guide
action (or that we know that our use of an expression conforms with its
meaning, or that we are actually applying expressions in accord with
their explanations, i.e., the rules for their use). Rather, what is repudiated
is the suggestion that a rule determines an action as being in
accord with it only in virtue of an interpretation.
The first sentence of §202 merely repeats the penultimate point of
§201, viz., how I understand a rule (racine Auffassung) is ultimately
exhibited not by an interpretation (the substitution of one expression of
a rule for another), but in what we call 'following the rule', i.e., in what I
do in applying the rule. Hence following a rule is an activity, a Praxis. It
is a misinterpretation to take 'Praxis' here to signify a social practice.
The contrast here is not between an aria and a chorus, but between looking
at a score and singing. The term 'practice' is used here in a similar
sense to that in the phrase 'in theory and in practice'. The point is not to
establish that language necessarily involves a community (infra), but
that 'words are deeds'. But a practice is not mere action, it is regular
action in accord with a rule, 'not something that happens once, no
matter of what kinď.28 Note that nothing in this discussion involves any
commitment to a multiplicity of agents. All the emphasis is on the
regularity, the multiple occasions, of action (cf. §199). What is crucial
for the concept of following a rule is recurrent action in appropriate
contexts, action which counts as following the rule. 29 Whether others
ON MISUNDERSTANDING WITTGENSTEIN 421
are involved is a further question. Of course, with us social creatures
rule-following is generally a social practice. But the point of the
argument was not to establish this (obvious) fact, but rather to show that
rule-following, and hence a language, is a kind of customary behaviour,
a form of action, not of thought. The 'foundations' of language are not
in private experience, the 'given' indefinables, but in normative
regularities of conduct.
The remainder of §202, is, in this context, puzzling. For Wittgenstein
has not yet explained what following a rule 'privately' means. The
passage derives from MS 129, 121, where it occurs after the exposition
of the private language argument (50 of the 74 remarks constituting the
private language argument in the Investigations occur in MS 129, only
two of which come after p. 121). There the allusion to 'following a rule
"privately"' is perspicuously a back-reference to the private-language
argument. In transposing this remark it has become, perhaps inad-,
vertently, an anticipation of §258 (cf. MS 129, 43) of the private
language argument.
4. THE PRIVATE LANGUAGE ARGUMENT
The foregoing 'archaeologicaľ investigations provide reasons for
doubting whether the discussion up till §§201-2 contains the whole of
the private language argument proper. We have not proved that the
core question of the book is not a sceptical one, nor that the solution is
not a 'sceptical solution'. But if it is, it does not lie in the passages of
§§198-202. Before going on to provide further argument, however, we
should look forward from §202 to what is normally conceived as the
private language argument in order to see whether Kripke's claims
about it rest on firmer foundations.
Kripke contends that the 'real private language argumenť occurs
prior to §243, and that the crucial considerations for that argument are
contained in the discussion leading up to §202 which states its conclusions.
What is commonly called 'the private language argumenť
deals with the application of the general conclusions about language
drawn in §§138-242 to the problem of sensations.3° Inner experience,
like mathematics, allegedly seemed to Wittgenstein to be a counterexample
to his view of rules, hence he treats it in detail. The connection
between the discussion of sensations and the mathematical reflections is
422 G. P. BAKER AND P. M. S. HACKER
shown in Remarks on Foundations of Mathematics I-§3:
How do I know that in working out the series +2 I must write "20004, 20006", and not
"20004, 20008"? - (The question: "How do I know that this colour is 'reď?" is similar.)
This passage, Kripke contends, 31 illustrates that Wittgenstein regards
the fundamental problems of the philosophy of mathematics and of the
private language argument (i.e., the 'problem' of sensation language) as
at root identical, stemming from his paradox. The impossibility of a
private language (of sensations) follows from the incorrectness of the
private model for language and rules, which is established in §202.
This interpretation of the concerns of §§243 ft. is perverse. The
private language argument is not about 'the problem of sensations',
which constitutes a prima facie counterexample to a thesis about rules.
It is concerned with establishing the nonprimacy of the mental, the
'inner', the subjective. In this enterprise Wittgenstein is stalking a much
larger quarry than a potential counterexample to one of his own 'theses'
(what theses?), namely the conception of the mental underlying the
mainstream of European philosophy since Descartes. It is noteworthy
that in his 'Notes for Lectures on "Private Experience" and "Sense
Data" ' he worked with the examples of 'seeing reď and 'having a red
visual impression' (i.e. perception) no less than with that of toothache (a
sensation). And his concern is explicitly with the refutation of idealism
and solipsism: the original fly in the fly-bottle is the solipsist!32
The private language argument is indeed built on the previous
discussions, not only of rule-following but also of ostensive definitions,
samples, meaning, understanding, and explanation. So too Kanťs
'Dialectic' is built on the 'Analytic', but that does not mean that the
'real "Dialectic" ' is the 'Analytic'. What is new in the private language
argument is the question of whether a 'private' sample can be employed
to give meaning to a word, whether a mental paradigm can be
employed, via a stipulation or explanation to oneself, to constitute a
norm of correct use.33 Do the foundations of language lie in mental
ostensive definitions of simple 'indefinable' perceptual predicates?
Certainly a great tradition of European philosophy embraced such a
conception, not only in the remote past, but in the writings of Russell,
Carnap (at one stage) and other logical positivists, and indeed, for a
brief time, in Wittgenstein's own work.
Nor is it true that the question raised has implicitly been answered by
the antecedent discussion of rule-following. After all, it may be argued
ON MISUNDERSTANDING WITTGENSTEIN 423
(and, alas, often is) that given biological nature,34 you must have just
what I have when we both look at tomatoes, hit our shins, etc. So you
know what 'pain' or 'looks reď means from your exemplar, just as I
know from my exemplar. And evolution, or a good angel, has so
arranged matters that our exemplars are qualitatively identical. So our
public language with its vast network of regularities of action is the
confluence, or congruence, of our private languages. So thought and
language rest firmly on the bedrock of the subjective.
It is, of course, true that in MSS 129 and 180(a) this argument is ruled
out on the grounds that there can be no possibility of establishing an
internal relation between a private sample, or 'subjective interpretation',
and an action determined in a practice as being in accord
with such a 'rule'. Moreover, if something can be a justification for me it
must also be capable of functioning as a justification for others. So my
'seeing that this is so' cannot function as a justification. There can be no
technique of applying a 'private rule'. Such a rule really would 'hang in
the air', and there would indeed be no distinction between thinking one
is following a rule and actually following it. Here one would have only a
'Schein-Praxis'. But the argument to establish these conclusions is the
private language argument! By transposing §202 from the vicinity of
§§377-81 it has been deprived of its argumentative support. As it
stands, the last sentence does indeed state a conclusion of the private
language argument, but now it 'hangs in the air'.
Even if we follow Kripke's interpretation, nothing significant is
altered. 'Any individual who claims to have mastered the concept
of .... will be judged by the community to have done so if his particular
responses agree with those of the community in enough cases, especially
the simple ones.'35 So we may accept, e.g., that we need community
support to indulge in, say, colour predication (or any other conceptusing
game), but given that we agree in judgments, agreement in
private definitions is ensured (by a good angel; or the 'argument to the
best explanation'). Hence it is not the case that the conclusion of the
private language argument has been proven before it is raised in
§§243 ft.
It is no less misleading to point at Remarks on the Foundations of
Mathematics I §3 to establish a connection between considerations of
rule-following and the private language argument proper. Of course
there are deep affinities between questions in philosophy of mathematics
and questions in philosophy of mind. But this passage does not point
424 G. P. BAKER AND P. M. S. HACKER
to one. Here there is a connection between expanding the series '+2',
and applying the predicate 'reď. But the concern of the private
language argument is not with objective predications, but with subjective
ones. Wittgenstein states that the question 'How do I know this colour is
"red"?' is similar to the arithmetical question, not that the question
'How do I know that this experience is "seeing red"?' is similar. In fact
the two pairs of questions are very different. In the first two cases we are
concerned with the bedrock of rules. No further rule mediates between
the rule 'Add 2' and saying '20004, 20006', or between an ostensive
definition of 'reď and a judgment that this is red. But the private
language argument focusses on the case of 'I am in pain' or 'I am seeing
reď where the issue is not bridging the gulf between a genuine rule (an
ostensive definition) and its use or application, but whether there is any
such thing as a bogus 'private' rule, viz., a private ostensive definition.
The argument turns on such matters as the lst/3rd person asymmetry
of psychological predicates, criterionless self-ascription and behavioural
criteria for third-person ascription, the noncognitive status of
avowals, etc.
Finally, it is misleading to represent the Remarks on the Foundations
of Mathematics as an examination of an apparent counterexample to
the discussion of rules (as the private language argument examines
sensations) - on the contrary, the discussion of rules, of rule-following
and of application of rules is used to shed light upon the central
questions of the philosophy of mathematics, viz., the nature of mathematical
necessity, the status of proof, and the relation of mathematics
to logic.
5. KRIPKE'S SCEPTICAL PROBLEM
So far we have compared Kripke's discussion to Wittgenstein's, and
found prima facie reasons for hesitating to embrace his interpretation.
We now turn to an examination of the cogency of some of his
arguments.
Kripke casts the central problem of the Investigations in terms of a
'sceptical hypothesis about a change in my usage'. 36 How can I know
that my understanding of 'W', what I mean by 'W', determines the
correctness of applying 'W' to this? The sceptic doubts whether any
instruction I gave myself in the past compels or justifies this answer
(rather than an absurd one). Kripke in effect shifts Wittgenstein's
ON MISUNDERSTANDING WITTGENSTEIN 425
problem of how, in what sense, a rule determines its application, to a
problem of the relation between my past and present intentions, my
meaning addition by 'plus' (and not a different arithmetical operation
christened 'quaddition').
This shift is one from an altogether natural belief to a bizarre one. For
it is natural to think that given that e.g., 'reď or 'plus' means what it
does, it follows that this is red, or that 68 +57 = 125. But it is not
obviously plausible or intuitively evident that in answering the questions
'What colour is this?' or 'What is 68 + 57?' one conceives of
oneself as following an instruction that one gave oneself in the past. A
fortiori in remarking 'that is a splendid reď (looking at a field of
poppies) one does not conceive oneself as obeying one's past instructions
about the meaning of 'reď, as opposed to using the word 'reď
in accord with its meaning. 37
Why does Kripke's predicament lead first to scepticism and ultimately
to conceptual nihilism? The reason offered is that we are
naturally inclined to say that we 'know directly and with a fair degree of
certainty'3~ that we mean W by 'W'. But Wittgenstein shows that no
mental event, act, activity, or process that may occur when we mean or
understand something constitutes the meaning or understanding. This
leaves us, apparently, only one move, viz., to claim that meaning W by
'W' is an irreducible experience known directly by introspection.
Against this Wittgenstein argues at length, and persuasively. We might
try a last stand, contending that meaning is a primitive, sui generis,
state:
Such a move may in a sense be irrefutable, and taken in an appropriate way
Wittgenstein may even accept it. But it seems desperate: it leaves the nature of this
postulated primitive state.., completely mysterious. It is not supposed to be an introspectible
state, yet we supposedly are aware of it with some fair degree of certainty
whenever it occurs. For how else can each of us be confident that he does, at present,
mean addition by 'plus'? Even more important is the logical difficulty implicit in
Wittgenstein's sceptical argument. I think that Wittgenstein argues, not merely as we
have said hitherto, that introspection shows that the alleged primitive state of understanding
is a chimera but also that it is logically impossible (or at least that there is
considerable logicat difficulty) for there to be a state of 'meaning addition by "plus" ' at
all.39
This reasoning betokens misunderstandings of Wittgenstein's
argument. Hence we will interrupt exposition of Kripke's sceptical
problem to show where it went off Wittgenstein's rails.
426 G. P. BAKER AND P. M. S. HACKER
First, according to Wittgenstein, 'I know that I mean...' is either an
emphatic manner of saying that I mean such-and-such, or it is nonsense.4°
There is no distinction between my meaning W by' W' and my
knowing that I mean W by' W' (unlike, say, my knowing that A is dead
and A's being dead). So there is no question here of 'knowing with a fair
degree of certainty' or of 'knowing directly'. My (appropriately)
confident assertion that I mean addition by 'plus' does not rest on
evidence of any kind.
Second, according to Wittgenstein, meaning and intending are not
experiences at all, afortiori not irreducible experiences known introspectively.
But it does not follow at all, nor did Wittgenstein suggest
that it followed, that I do not mean or intend the things I take myself to
mean or intend, the things I sincerely say that I mean or intend.41 All
that follows is that this philosophical picture of meaning and intending,
of self-knowledge and privileged access, is wrong. What Wittgenstein is
concerned with is not scepticism about our right to say that we mean
this or that, that we intend so and so, but extirpation of philosophical
confusion.
Third, Wittgenstein does not argue that introspection, i.e., an
'experimental methoď, reveals that as a matter of fact there is no
'primitive state of understanding'. He does indeed argue that it is
logically impossible for there to be a state of meaning W by 'W', but
not in the manner Kripke suggests. And this argument does not even
suggest that I do not mean what I normally say that I mean. The
argument is that understanding, meaning, and intending are not states
of any kind. 42 Mental states are such things as being nervous, excited,
exhilarated. These obtain for a time and can be clocked, interrupted
and resumed. They run a course, are subject to degrees of intensity and
lapse during sleep. Not so are understanding, meaning and intending.
But it does not follow, according to Wittgenstein, that one does not
mean addition by 'plus', or that one does not intend to play chess when
one sincerely says 'Leťs play chess'.
Fourth, even if, per impossibile, meaning were a state, this would not
solve the apparent problem. For how could any state bridge the gap,
the logical gulf, between a rule and its application? What feature of a
'state of meaning' could make it right to apply the rule thus or
otherwise? Is it not obvious that this hypothetical state of meaning
would be tantamount to another interpretation of the rule? The jump to
its application would still have to be made.
ON MISUNDERSTANDING WITTGENSTEIN 427
Of course, according to Wittgenstein, I may confidently say that I
mean addition by 'plus', but not because I am introspectively aware of
my inner state of meaning. Rather, as I am confident that I intend to
play chess (and not some other game), or confident that what I now
expect is John to come (and not, James to go or the pound sterling to
fluctuate). If my confidence rested on an inner awareness, it would be
inductive. I would have to reason that whenever I have in the past had
this inner state, then I have gone on to... But then I ought to say 'I
think I intend to play chess. Leťs see!', which is absurd. With these
rectifications behind us, we may resume exposition of Kripke.
The claim that there is no such thing as the 'state of meaning addition
by "plus"' leads directly to the culmination of Kripke's sceptical
argument. The sceptic holds, according to Kripke's view, 'that no fact
about my past history, nothing that was ever in my mind, or in my
external behaviour- establishes that I meant plus rather than qUUS'43
and, in view of the introspective argument, 'it appears to follow that
there was no fact about me that constituted my having meant plus'.44
This argument, in particular its phrasing in terms of facts, is crucial for
Kripke's interpretation. For, he claims, the target of Investigations
§§1-133 is the truth-conditional theory of meaning as propounded in
the Tractatus. According to that theory, what makes propositions true
are corresponding facts-in-the-world (hence what would make 'I mean
W by "W" true would be some fact-in-my-mind). And what gives
sentences their meaning are their correlations with possible facts or
conditions in the world, viz., their truth-conditions.45The repudiation of
this truth-conditional theory of meaning is the pivotal point of Wittgenstein's
alleged 'Sceptical Solution' to his alleged sceptical problem.
He is held to agree with the sceptic that there is no fact-in-the-world that
constitutes my meaning W by 'W', but also to repudiate the truthconditional
theory of meaning that requires the obtaining of such facts to
make true claims about meaning.
This is off-target. It misinterprets the Tractatus conception of
truth-conditions (infra). It distorts the structure of the argument of the
Investigations. And, by Wittgenstein's lights, it is plainly wrong.
First, no facts are in the world. It is a fact that Oxford is in England,
but that fact is not in England, nor yet in France - for facts are not in
space. Nor are they temporal entities: the fact that Hastings was fought
in 1066 did not occur in 1066, since it is Hastings - the battle - that
occurred then, not the fact. The fact that it occurred in 1066 did not
428 G. P. BAKER AND P. M. S. HACKER
cease to be a fact in 1067, nor was it a mere proto-fact in the womb of
History in 1065. Hence the fact that I yesterday meant W by ' W' is not
a fact-that-was-in-the-world-yesterday. And the fact that I now mean
W by ' W' is not a fact-in-my-mind-now. But if I did mean W by ' W',
then it is a fact that I so meant, and if I now mean W by ' W' then it is a
fact that I so mean.
Second, if Kripke wishes, one can say that the fact about me that
constituted my having meant addition by 'plus' is the fact that I so
meant. For if I did so mean, it is a fact that I so meant (and I will tell you
if you ask me). Of course, the stick won't move. But that is because one
has got hold of the wrong end of it.
Third, to be sure, when I tell you that I meant W by 'W', or that I
meant you to go '1002, 1004' or that I intend to visit London tomorrow,
I do not read these statements off the 'facts-in-the-worlď. Kripke
intimates that Wittgenstein's only alternative to the 'Sceptical Solution'
is a picture according to which my confidently saying what I mean must
result from my reading off what I mean from a fact-in-the-world (in my
mind). This is precisely what Wittgenstein denies. The picture of
facts-in-the-world is a muddle. Its solution does not consist in denying
that there are any facts concerning my meaning things by words, but in
sorting out the muddle.
Fourth, if the repudiation of his earlier truth-conditional theory of
meaning (understood as involving correspondence with facts-in-theworld)
is the key issue in Investigations §§1-133 and is crucial to the
solution of the sceptical paradox, it cannot but be surprising that
Wittgenstein has no discussion of facts in the whole book. Nowhere
does he examine the concept of a fact, nor suggest that it is not facts
that make propositions true. Is this because he had nothing to say? Or
because he was so cagy?
As a matter of fact, Wittgenstein had a very straightforward way with
facts - and stated it clearly in what is now published as Philosophical
Grammar. 46 That he did not incorporate these remarks into the
Investigations strongly suggests that he was not aiming at this target at
all. His moves are simple. He does not deny that what makes the
proposition that p true is the fact that p. He does not repudiate the
claim that the proposition determines in advance what will make it true
(what fact must obtain to make it true). Rather he rejects the metaphysical
picture that goes with these claims. For these are grammatical
statements, not metaphysical profundities. They concern intra-lin-
ON MISUNDERSTANDING WITTGENSTEIN 429
guistic articulations, not the ultimate connections between language
and reality. It is a convention of grammar that 'The proposition that p'
= 'the proposition that the fact that p makes true'. And so too 'The fact
that p' = 'The fact that makes the proposition that p true'. Like
everything metaphysical the harmony between thought and reality is to
be found in the grammar of the language. 47
Finally, it is noteworthy that the way Kripke sets up his sceptical
paradox initially parallels traditional scepticism in assuming that only
entailment will license a cognitive claim. Nothing in the past or in the
present, in my mind or my behaviour, entails that I now mean by 'plus'
what I previously meant by 'plus'. And so on. This observation parallels
defences of scepticism about the past, about other minds or about
induction. And, of course, we need not accept any such arguments.
What shows that I meant green by 'green' is the way I explained 'green',
and what shows that I meant grue is giving a quite different explanation.
That I gave such and such an explanation does not entail that in
applying 'green' to this object I am using 'green' in accord with what I
meant by it hitherto, but it provides perfectly adequate grounds for that
judgment (if anyone is interested in such a bizarre question). How I use
an expression, how I explain an expression, how I use the explanation as
a norm of correctness (in indefinitely many cases, new and old), what I
count as applying the expression in accord with its explanation (its
meaning) shows what I understand by it. This sceptical problem can be
side-stepped. But we may doubt whether this was ever Wittgenstein's
problem!
6. KRIPKE'S SCEPTICAL SOLUTION
Initial scepticism, according to Kripke, leads us to doubt whether we
are applying words in accord with how we have in the past meant them.
This led, by a quick route, to conceptual nihilism that denies that there
is any such thing as meaning, and faces us with a paradox: language is
impossible. Wittgenstein's solution, Kripke claims, consists in rejecting
a truth-conditional theory of meaning in favour of assertion-conditions.
Within this framework of thought, he can accept the sceptic's premises
that there is no fact-in-the-world constituting my meaning W by 'W',
hence (sic !) that 'Jones means plus by "plus"' has no truth-conditions, 48
yet deny the sceptics paradox. What he does is to describe the assertion
conditions of such statements viz.:
430 G. P. BAKER AND P. M. S. HACKER
Jones is entitled, subject to correction by others, provisionaUy to say, "I mean addition by
'plus'" whenever he has the feeling of confidence - "now I can go on!" - that he can give
'correcť responses in new cases; and he is entitled, again provisionally and subject to
correction by others, to judge a new response to be 'correcť simply because it is the
response he is inclined to give .... Smith will judge Jones to mean addition by 'plus' only if
he judges that Jones's answers to particular addition problems agree with those he is
inclined to give...
ˇ.. Any individual who claims to have mastered the concept of addition will be judged
by the community to have done so if his particular responses agree with those of the
community in enough cases...49
It is difficult here to find any similitude to Wittgenstein's arguments.
Indeed, it is difficult to see any plausibility in the argument, irrespective
of whether it is Wittgenstein's.
We are asked to accept as plausible the following exchange: we ask
Jones 'Do you mean addition by "plus" (or red by "red", bachelor by
"bachelor" etc.)?' And he is then supposed to answer: 'Yes, I do, and I
know that I do because I feel confident I can answer such questions as
"What is 68 + 57?" correctly (or "What colour is this?", or "Is John a
bachelor?").' But this is bizarre. That one feels confident is not what
entitles one to say that one means W by 'W'. And that one is inclined
to answer thus and so is not what entitles one to judge one's answer to
be correct.5°
First, other things being equal, I will always claim to mean W by
'W'. The question is what I mean by 'W', i.e., whether I actually
understand this word, whether I know what it means. The issue is not
whether I am now using it in accord with what I previously meant by it,
but whether I am now using it in accord with its meaning. The answer
to these various questions is not to insist on my confidence, but to say
what I mean, i.e., to explain what 'W' means, e.g., that, 'bachelor'
means an unmarried male, 'reď means this colour (pointing at a ripe
tomato), and 'plus' means that function which when 2 and 3 are its
arguments, yields 5 as its value, and which, when .... Of course, these
explanations do not bridge the gap between meaning and use, rule and
application. But no explanation of any kind could do that, since the
'gap' is categorial. But giving these explanations does, ceteris paribus,
provide adequate grounds for judging what a person means, and
whether he knows what an expression means.
Second, the very idea that I could have an entitlement to say 'I mean
W by "W"' is odd. For it suggests that there are grounds, assertability
ON MISUNDERSTANDING WITTGENSTEIN 431
conditions, the obtaining of which I must establish before 1 may say that
I mean such and such. But there are no more grounds for my saying 'I
mean W by "IV"' than there are grounds, assertability conditions, for
saying 'I intend to go to London tomorrow' or 'I want a drink'. So too, it
is misleading to suggest that there is here a question of my knowing that
I mean W by "W" (viz., if I have a title to assert, then ceterisparibus, I
know...). But 'I know that I mean W by "W"' is just an emphatic
insistence that I do mean W by "W".
Kripke's description of the assertion-conditions for a third-person
ascription is ne less strange. We must imagine the following exchange:
we ask Smith 'Does Jones mean addition by "plus"?' He replies, "Yes,
because whenever he is asked "What are a plus b" (for any a and b), he
always gives the same answer as I give'. This is awry.
First, whether Jones gives the same answer as Smith is beside the
point. The question is whether he gives the correct answer, 51 i.e., what
counts, in such a case, as the correct answer. In some cases, even this
procedure would be absurd. For example, does Jones mean bachelor by
'bachelor'? Are we to say: 'Yes, whenever he is asked whether Mr. A is
a bachelor he gives the same answer I am inclined to give'? But in most
cases the answer I would be inclined to give would be 'I don't know'!
Would this response satisfy the assertion conditions for 'Jones means
bachelor by "bachelor"'9
Second, the analysis seems to suggest that Smith cannot judge that
Jones understands 'W' (means W by 'W') unless (a) he knows how
Jones applies ' IV' to new instances, and (b) he takes a given application
of ' W' to be correct simply because it is the one he himself is inclined to
give. But it is a conceptual truth that I am entitled to judge someone to
mean W by' W' (to understanď W') on the grounds of the explanations
of' W' he gives'? If he says 'By "bachelor" I mean an unmarried male', is
that not enough? Must he answer the question of whether Genghis Khan
at the age of 22 was a bachelor? And must he answer it the way I am
inclined to? 52
Putting together the first and third person assertion-conditions as a
general account, two points still stand forth as baffling.
First, truth-conditions were meant to provide an account of the
meanings of sentences. Giving the truth-conditions is generally supposed
to be a way of giving or explaining the meaning of a sentence.
Assertion conditions are presumably to inherit this role from truthconditions
in Kripke's account. So the meaning of 'A means addition
432 G. P. BAKER AND P. M. S. HACKER
(or whatever) by "plus" (or whatever)' is supposed to be given by
Kripke's specification of its assertion conditions. But would anyone
thus explain to someone who genuinely wanted to know what 'A means
W by "W"' means? And would such an explanation provide a norm for
the correct use of 'means'?
Second, has the original sceptical question really been answered?
The problem was set up by arguing that we can never have adequate
grounds to judge that anyone now uses an expression in accord with its
previous meaning (with what he, or we, previously meant by it). The
sceptical solution is that no fact about past or present can tell us that our
current use coheres with our past use. But if we all share common
inclinations to apply the term thus-and-so, then we all now mean W by
'W', or, more carefully, as long as a speaker does not apply 'W'
differently from the way the rest of the community is inclined to apply
it, then he will be accepted as following the rule for ' W', as meaning by
'W' what everyone else means.
But does this really resolve the sceptical question? Given that no one
previously ever added 57 and 68, how do we know that our present
community-wide inclination to answer '125' accords with what we
previously meant by 'plus', i.e., with what we would have been inclined
to say, had we previously been asked what 57 + 68 is? Like Kripke, we
put the challenge 'in terms of a sceptical hypothesis about a change in
[our] usage'. 53 For there was yesterday no satisfied assertion condition
for our meaning either plus or quus, since we had no inclination to
answer '57 + 68' either way, as the question, ex hypothesi, had never
occurred to anyone hitherto.
7. ROBINSON CRUSOE RIDES AGAIN
Kripke's assertion conditions account is part of his picture of Wittgenstein's
argument, but not the whole. In addition, he contends,
Wittgenstein describes the role and utility in our lives of assertions that
someone means such-and-such by his words, or that his present use of a
word accords with what he previously meant by it. It turns out,
however, that this role and the conditions of assertion, are inapplicable
to a single person considered in isolation.54 If one person is considered
in isolation, the notion of his following a rule can have no substantive
content. For as long as we regard him as following a rule 'privately', i.e.,
merely following his inclinations (sic!), then there is no difference
ON MISUNDERSTANDING WITTGENSTEIN 433
between his thinking he is following a rule and his following one. It is
this argument which, by §202, rules out as incoherent the private
language whic,h is introduced only in §§243 ft.
What then of Robinson Crusoe on his desert island? It does not
follow, Kripke claims, that he cannot be said to follow rules. 'What does
follow is that if we think of Crusoe as following rules, we are taking him
into our community and applying our criteria for rule-following to him.
The falsity of the private model need not mean that a physically isolated
individual cannot be said to follow rules; rather that an individual
considered in isolation (whether or not he is physically isolated) cannot
be said to do so.'5s
This is muddled. In the first place, it is quite wrong to suppose that
distinctions between appearance and reality are inapplicable to an individual
in isolation, and are ones which that individual cannot employ.
In the particular case of rule-following, there is no reason why Crusoe
should not follow a pattern or paradigm, making occasional mistakes
perhaps, and occasionally (but maybe not always) noticing and
correcting his mistakes. That he is following a rule will show itself in the
manner in which he uses the formulation of the rule as a canon or norm
of correctness. Hence, to take a simple example, he might use the
pattern ............ as a rule or pattern to follow in decorating the
walls of his house; when he notices four dots in a sequence he manifests
annoyance with himself. He carefully goes back and rubs one out, and
perhaps checks carefully adjacent marks, comparing them with his
~master-pattern'. And so on. Of course, he is not merely following his
qnclinations',56 but rather following the rule. And it is his behaviour,
including his corrective behaviour, which shows both that he is
following the rule, and what he counts as following the rule.
It might be asked how an unseen observer of such solitary rulefollowing
could distinguish Crusoe's making a mistake from his following
a more complex pattern, and his following a more complex pattern
from his non-normative behaviour. How could one justify the claim
that the solitary man is either following or breaking a rule at all? The
answer is that if one must, ex hypothesi, remain unseen, it will be very
difficult to understand him. If the rules are simple, we might guess
aright. If they are complex, we might not. Reflect that if we observed
the self-addressed speech of a shipwrecked monolingual Tibetan, our
chances of coming to understand him are remote. But he surely could
talk to himself, keep a diary, give orders to himself, play Tibetan
434 G. P. BAKER AND P. M. S. HACKER
solitaire. Once the restriction of unseen observation is lifted,
however, matters change altogether. Gestures ('the natural language of
mankinď as Augustine says), common human nature, and interaction
with the castaway provide the necessary leverage. Of course, to
understand him we must grasp his rules. Whether we are succeeding in
doing so is something we shall see from the extent to which our attempts
to follow his rules are in agreement with his behaviour. But whether he
is following a rule is independent of whether anyone else is actually
doing so too.
Second, Kripke rightly concedes that Crusoe may follow rules, but
insists that in saying of him that he does, we 'are taking him into our
community and applying our criteria for rule-following to him'. This
seems confused. This 'taking him into our community' will do little to
alleviate Crusoe's solitude. What is it supposed to mean? Does it mean
that in saying that he is following a rule we are applying our criteria for
rule-following to him? Well - are there other criteria? This, presumably,
is what 'rule following' means. When we say of the cat that it is
hunting the mouse, we are applying our criteria of hunting to it. Do we
thereby take the cat into our community? This, it might be replied, is
beside the point. Our 'taking Crusoe into our community' consists in
the fact that we judge that he is following a rule only if he satisfies the
assertion conditions of following a rule, and these, Kripke contends,
stipulate that A can be said to mean W by' W', to follow such-and-such
a rule, if he applies the rule (uses 'W') as other members of the
community do, if his responses and inclinations agree with theirs. Since
Crusoe is not a member of any community, in judging him to be
following a rule, we must be 'taking him into our community', judging
his responses to agree with ours.
Must we? Must Crusoe's rules be the same as ours? Must his colour
vocabulary be isomorphic with ours? Could he not invent new rules,
play new games? To be sure, in order to grasp them, we must
understand what counts, in Crusoe's practice, as following the rules.
And that must be evident in Crusoe's activities. But that is not the same
as checking to see whether his responses agree with ours, let alone a
matter of 'taking him into our community'. And our judgment that he is
following his rules is quite independent of any judgment about how
most members of the English Speaking Peoples would react. Indeed,
given Kripke's rule-scepticism, how are we supposed to know how our
community would react, given that the rule is novel, or is being applied
ON MISUNDERSTANDING WITTGENSTEIN 435
to a novel circumstance?
Interestingly, Wittgenstein did explicitly discuss Robinson Crusoe in
his notebooks. In MS 124, p. 213 we find an early version of Investigations
§243(a), a discussion of the imaginary monologists,
whose language is translatable by the explorer. Couldn't we imagine people
who speak only to themselves? In that case, Wittgenstein responds,
each person could have his own language. There could be men who
know only language games that one plays by oneself, viz., ordering
oneself, telling oneself, asking and answering oneself, etc. How they
learnt their language is here irrelevant, he adds. An explorer who
observed the behaviour of such monologists could translate their
languages. On p. 221 Wittgenstein remarks that the private language
that he has described above is one which Robinson could have spoken
to himself on his island. If anyone had observed him, he could have
learnt this language. For the meanings of the words of this (contingently)
private language are shown in Robinson's behaviour, s7
There is no hint here that in attributing rule-following to Crusoe, in
judging him to mean such and such by what he says, we are 'taking him
into our community'. There is no suggestion that our concept of
rule-following or of meaning is limited to our rules, or to what we mean
by signs. There is no claim that his responses must agree with ours (he
may have invented a new branch of mathematics; he may employ a
different colour geometry from ours; or he may apply names of notes
immediately, given that he has perfect pitch). The claim does not
involve insistence on community-aid for solitary rule-followers, but on
regularities of action of sufficient complexity to yield normativity. The
criteria for whether Crusoe is following a rule do indeed lie in his
behaviour, but not in his behaviour agreeing with independent hypothetical
or counterfactual behaviour of ours.
It is noteworthy that immediately following the remark about Crusoe
on p. 221 of MS 124, Wittgenstein introduces the real private language
argument, i.e., a draft of Investigations §243(b). There is no hint that
he considers such a language to have been ruled out by his antecedent
reflections on rule-following in general or on Crusoe's solitary rulefollowing
in particular.
8. FURTHER DIAGNOSIS
The discussion thus far suggests that Kripke has misinterpreted Witt-
436 G. P. BAKER AND P. M. S. HACKER
genstein's argument 'in the Large and in the Smalľ. Three substantial
misunderstandings or distortions ramify throughout his essay. These are
likely to mislead readers and to lead to futile debates about Wittgenstein's
intentions. The following observations may help to avoid
this.
(i) Kripke sets up his sceptical problem (the 'real private language
argumenť) in terms of what a speaker means by an expression, of
whether he is now using an expression in accord with what he previously
meant by it. This is a highly misleading way of broaching the
core problem Wittgenstein is concerned with, namely, what is involved
in a speaker understanding an expression, knowing what it means, using
it in accord with a correct explanation of its meaning. Kripke's strategy
is misleading because it runs together the internally related, but distinct,
notions of what an expression means and understanding an expression.
Focussing on the issue of conformity of current use with a pattern of
past use conflates the question of the persistence of understanding (an
ability) with the question of the correctness of the present use (conformity
with a norm). Kripke vainly attempts to extricate himself from
the consequent muddle by invoking community aid. Wittgenstein's
strategy is entirely different. He is careful to keep these questions
distinct. He correlates understanding both with the use of an expression
and with explaining its meaning, and he stresses that these correlations
must not be conflated. Since meaning is a correlate of understanding, the
meaning of a word is also linked both with its use and with explanations
of its meaning. The meaning of a word is what is explained by an
explanation of its meaning. And it is also the manner in which the word
is used in speech. Understanding the meaning of a word is akin to an
ability; it is the mastery of a technique of using a symbol according to
rules. The criteria of understanding lie in behaviour, in the use of the
expression in accord with (in what counts as accord with) its explanation,
the rules for its use, and in the giving of correct explanations of its
meaning (which may be by example, paraphrase, contextual paraphrase,
ostension, Merkmal-definition, etc.). Not only agreement in
judgments, as Kripke suggests, but also agreement in definitions is
essential to meaning and understanding - and the notions must be
described with sufficient subtlety that logic is not thereby abolished. It is
unclear whether Kripke satisfies this obvious requirement. On p. 297 he
insists, following Wittgenstein, that truth is not to be equated with what
most people hold to be true. He insists that Wittgenstein has no theory
ON MISUNDERSTANDING WITTGENSTEIN 437
of truth-conditions, necessary and sufficient conditions for the correctness
of one response rather than another to a new addition problem
(and, presumably, to a new colour predication). Wittgenstein's assertability
conditions story, according to Kripke, does not say that the
correct answer to an addition problem is the one everyone gives, but
only the platitude that if everyone agrees upon a certain answer, then
no one will feel justified in calling that answer wrong. But is the answer
right? What does it mean, according to this story, for an answer to be
right as opposed to wrong? Unless an answer is forthcoming, Kripke's
Wittgenstein, unlike the author of the Philosophical Investigations, will
have abolished logic!
(ii) The role of agreement is certainly paramount in Wittgenstein's
argument. My use of an expression must agree, accord, with my correct
explanation of what it means. If I explain 'reď by pointing at a sample,
saying 'This is reď, then when I judge an object A to be red, A must be
this (pointing at the sample) colour. Moreover, ceteris paribus, my
current use of an expression must agree with my previous use. I must lay
down the yardstick alongside reality in the same way, i.e., in what is
called 'the same way' from occasion to occasion. And reality must be
sufficiently stable so that the yardstick typically gives the same result
when the same object is measured on successive occasions. Otherwise
measurement in particular and the application of concepts to reality in
general become pointless. Finally, the 'language-games' I engage in
with others can be played only if we agree in explanations (definitions)
and also, by and large, agree in applications (judgments).
Kripke appreciates the centrality of the notion of agreement for
Wittgenstein, but distorts its function. We noted that he allots no
significant role to agreement in definitions (explanations) and has
nothing to say on the relationship between agreement in judgments and
agreement in definitions. However, it also seems to be the case that
Kripke's Wittgenstein conceives of agreement as constitutive of
my meaning W by ' W'. Someone is judged by the community to have
mastered such-and-such a concept 'if his particular responses agree
with those of the community in enough cases'. 58 Someone in a community
is said to follow a rule 'as long as he agrees in his responses with
the.., responses produced by members of that community'.-s9 On this
account agreement with a community is part of the assertion-conditions
of 'meaning W by "W''' and hence part of its meaning. Is this
Wittgenstein's view?
438 G. P. BAKER AND P. M. S. HACKER
For Wittgenstein, agreement is a framework condition for the
existence of language-games, but is not constitutive of any game.
Hence it is not part of the criteria for whether A understands 'plus' or
'reď. These are, rather, that A explains correctly what the expressions
mean, and typically uses them correctly. That A understands what 'reď
means is shown by his giving a correct explanation (ostensive definition)
of 'reď, as well as by his saying of my red rose 'That is reď. That others
would also characterize my rose as red is not part of the criteria for
A's understanding, knowing the meaning of 'reď. But the framework
for these concept-exercising activities is general agreement. 6°
Similarly, Wittgenstein contends that proofs in mathematics stand in
need of ratification, and in the absence of a consensus in ratifications,
mathematicians would not come to any understanding, and the concept
of calculation would have no application.6t But this is not to say that it
is agreement in ratification that makes such and such operations calculation.
Far from it, 'the agreement of ratifications is the precondition
of our language-game, it is not affirmed in iť. The subject of agreement
in judgments and definitions is treated by Wittgenstein with great
subtlety. We shall not try here to unravel its complexities. But it is clear
that he does not conceive of agreement in judgments as a constitutive
element of a language-game.
Kripke does not take sufficiently seriously Wittgenstein's insistence
that 'Following according to the rule is FUNDAMENTAL to our
language-game. It characterizes what we call description'.62 He apparently
wants to go behind rule-following to agreement. But there is
nothing behind:
It is no use, for example, going back to the concept of agreement, because it is no more
certain that one proceeding is in agreement with another, than that it has happened in
accordance with a rule. Admittedly, going according to a rule is also founded on
agreement. 63
It is not surprising that if one tries to go beyond rule-following into the
framework that makes it possible, one will, in the process, lose the very
concept of normativity one is trying to clarify; and with it too the
distinction between correct and incorrect.
What a teacher teaches a child is not to have the same inclinations
others have, but to follow a rule correctly. The learner must learn to use
the rule, the explanation of meaning, as a standard of correct use. He
must learn to see such and such results as criteria for following the rule
ON MISUNDERSTANDING WITTGENSTEIN 439
correctly. And all this takes place within the background context of a
practice of using the rule in this way. If the learner is to master shared
concepts he must learn what counts as following the rule. And
what the correct following of a rule consists in cannot be described more closely than by
describing the learning of 'proceeding according to the rule'. And this description is an
everyday one, like that of cooking and sewing, for example.64
What counts as correct is not the response we are inclined to give. The
learner is not entitled to 'judge a new response to be "correct" simply
because it is the response he is inclined to give' (and others agree). It is
correct if it accords with the rule. But we can only speak of accord with
a rule in the context of a regular use of a rule as a measure of
correctness.
(iii) It has become fashionable in the last decade, under the influence
of Michael Dummett, to view the evolution of Wittgenstein's philosophies
as a transformation of a realist into an anti-realist, a truthconditional
theory into an assertability-conditions theory. This has
been an unfortunate influence, forcing Wittgenstein into a Procrustean
bed, rather than looking carefully to see what he says. Kripke, like
others,65 falls victim to this distorted way of looking at Wittgenstein.
Distortions occur at two levels: (a) the representation of Wittgenstein's
truth-conditional theory in the Tractatus; (b) the picture of Wittgenstein's
later views on meaning. We shall limit ourselves to some
schematic observations.66
Kripke characterizes the Tractatus with the following thumb-nail
sketch:
The simplest, most basic idea of the Tractatus can hardly be dismissed: a declarative
sentence gets its meaning by virtue of its truth-conditions, by virtue of its correspondence
to facts that must obtain if it is true. For example, "the cat is on the mat" is understood by
those speakers who realize that it is true if and only if a certain cat is on a certain mat; it is
false otherwise. The presence of the cat on the mat is a fact or condition-in-the-world that
would make the sentence true (express a truth) if it obtained.67
Even allowing for the vagaries of sketches, this conflates three distinct
doctrines of the Tractatus. First, there is a pictorial theory of the atomic
proposition. The sense of such a proposition is a function of the
meanings of its constituent unanalysable names, and it consists in its
depiction of an atomic state of affairs (which may, or may not, obtain).
Second, there is a truth-conditional account of the sense of molecular
propositions. Third, there is a correspondence theory of truth. Kripke
440 G. P. BAKER AND P. M. S. HACKER
treats the Tractatus truth-conditional theory as if it were part of the
pictorial theory of the atomic proposition. But according to the
Tractatus it literally makes no sense to talk of the truth-conditions of an
atomic proposition. The truth-conditions of a proposition are the
conditions under which ' Ť occurs in the final column of its truth-table.
But there is no such thing as a truth-table for the atomic proposition 'p'.
Similarly, Kripke apparently reads Tarski's T-sentences into the Tractatus
(viz., 'F(a)' is true if and only if F(a)). But this is not part of the
Tractatus theory of meaning. If T-sentences such as '"The cat is on the
mat" is true if and only if a certain cat is on a certain mať are what spell
out truth-conditions, then the truth-conditional theory parts company
with the picture theory of the proposition, the thesis of isomorphism, the
bipolarity of the proposition and the distinctive (ineffable) Tractatus
form of the correspondence theory of truth.
Does this matter? Is it not, after all, just so much history? It does
indeed. For by thus misrepresenting history we facilitate the fit of the
distorting spectacles that allow us to delude ourselves into viewing the
evolution of twentienth-century philosophy of language (philosophical
logic) as a progressive confrontation between truth-conditional semantics
and assertion-conditions semantics.
Kripke contends that the later Wittgenstein 'proposes a picture of
language based, not on truth-conditions, but on assertability conditions
or justification conditions. '68 It is very doubtful whether this picture of
Wittgenstein's later views on language does anything but distort the
reality it is meant to represent. It is true that for some kinds of
sentences, in particular third-person sentences concerning psychological
characteristics and sentences about abilities, we explain their meaning
in part by specifying the kinds of circumstance that justify their
assertion. But it would be absurd, as well as groundless, to foist on
Wittgenstein the view that the meaning of every sentence is given thus.
Kripke, to be sure, acknowledges that non-declaratives do not fit into
this picture, and avowals of sensation do not either. But the 'exceptions'
are not just these. It is not merely avowals of aches and pains that do not
fit this mould, but saying that I intend to do so and so, remember this or
that, want such and such. It is not merely psychological predications in
the first-person that mar the alleged pattern, but hosts of ordinary
utterances, such as 'The rose is reď, 'The table is rounď, 'It is warm
today', 'My name is N.N.', 'It is time to go', in short, most sentences.
Wittgenstein does not claim, with respect to sentences in general, that
we explain their meaning by giving their assertion conditions. The
ON MISUNDERSTANDING WITTGENSTEIN 441
injunction to look at how sentences are used is not an implicit claim
that all sentences have assertion conditions. What explanations would
we give that would justify asserting such sentences as those cited, and
also constitute explanations of their meaning? (But there is no difficulty
in explaining what 'My name is N.N.' or 'It is time to go' or 'The rose is
reď mean!)
Forcing Wittgenstein into the invented position of constructivism,
intuitionist-semantics, assertion-conditions theories, is altogether misguided.
It is a mistake stemming from a hankering after sweeping
generalizations, global confrontations of semantic theories, large scale
theory-building. But Wittgenstein builds no such theories. He does not
contend that a language is a monolithic structure run through with
truth-conditions or assertion-conditions that give meanings to sentences
and words. It is not a calculus of rules, either in the form of
classical logic or in the form of intuitionist logic. It is a motley of
language games, an endlessly variegated form of human activity,
interwoven with our lives at every level.
9. A CONCLUDING SKETCH
We have tried to show that Wittgenstein's argument as it struck Kripke
is very far removed from Wittgenstein's argument. We have denied that
Wittgenstein is ,concerned with a sceptical problem, and denied that he
gives a Humean solution to the problems he was concerned with. To
give a proper account of Wittgenstein's discussion of rule-following and
its relation to the private language argument would be a large task,
which we shall confront in another forum.69 But perhaps the following
sketch may be helpful to a reader who, weary of the flow of denials,
hankers for some positive suggestions about Wittgenstein's discussion
of rule-following prior to §243.
The conception of meaning that Wittgenstein delineated in the
Tractatus involved a commitment to various metaphysical doctrines. In
particular, fully analysed names were conceived as standing for simple
entities in reality which were their meanings. These 'objects' were
metaphysical simples, the indestructible substance of reality. The
combinatorial possibilities of simple names in a language must mirror
the metaphysical combinatorial possibilities of objects in reality. Wittgenstein
later referred to this conception as the 'Bedeutungsk6rper'
(meaning-body) picture. When he returned to philosophy in 1929, the
first element of his old way of thinking that he jettisoned was precisely
442 G. P. BAKER AND P. M. S. HACKER
the Bedeutungski~rper picture. What he argued initially was that instead
of conceiving of language as a system of (analysable) signs, connected
by means of lines of projection with metaphysical simples that provide
the signs with meanings, we should conceive of language simply as a
calculus of rules. The meaning of an expression is not a sempiternal
simple object,7°but the rules for its use, the totality of which fix its place
in a calculus of meaning rules.
Under pressure, this picture too gave way. It gave way, not because a
language is not rule-governed, nor because speaking is not a normative
activity, but because it involved a mystification of rules no less
distorting than the original metaphysics of objects that the normative,
calculus conception was meant to replace. Rules for the use of
expressions are not bits of normative machinery. They are not Platonic
entities whirring away in Fregean third realms, nor are they psychological
entities determining in a causal manner how we are constrained
to think. Two deep and ramifying confusions must be extir-
pated.
It is very tempting to conceive of a sign as standing in a projective
relation to what it represents. This conception may be variously
realized. Frege conceived the senses associated with words as determining
a reference. The Tractatus conceived of the proposition as a
logical picture representing a state of affairs. The representing picture,
the proposition, was held to include the pictorial relationship, 'the
correlations of the picture's elements with things' (Tractatus 2.1514). In
a like manner, if one conceives of the meaning of an expression as
constituted by the rules for its use, one may think that the rule must, in
some sense, contain a 'picture' or 'representation' of what complies
with it. For understanding an expression must constitute knowing,
grasping, the rules that constitute its meaning. Those rules stipulate
how the expression is to be used. So by grasping the rules one must grasp
how to use, how to apply the expression. But that would only be possible,
it seems, if the rule determines independently of us what accords
with it. Otherwise how could I, by grasping the rule, know what to do with
the expression the meaning of which is given by the rule. (Precisely
analogous is the thought that an intention, expectation, or order must
contain a 'picture' of its fulfilment.)
One of Wittgenstein's central concerns is to combat this conception.
We must sharply distinguish the lines of projection from the technique
or method of projection. The method of projection is not part of the
ON MISUNDERSTANDING WITTGENSTEIN 443
symbol. Of course, one can describe the method of projection (the
application of a symbol). But such a description is itself just another
symbol (an interpretation). If the method of application is a bridge
between a symbol and what it symbolizes, it is at any rate not built until
the application is made. It is not the interpretation that builds the bridge
between the sign and what is signified or meant, only the practice does
that. 71 It is not rules, of their own accord, that determine meanings, it is
the way in which we, in our activities, use rules, that does so. It is not
rules that breathe life into signs, but our using the signs in accord with
rules, in what we call 'accorď. And that is not fixed (magically) by,the
rule all on its own, by its containing a picture of what accords with~it and
what does not. It is fixed by our practices of using the rule (the
explanation of meaning) to constitute a norm of correctness, our practices
of teaching and explaining, of criticising and correcting, of
justifying our applications of an expression by reference to the rule.
A second great confusion is closely connected with the first. We are
inclined to think that it is the mind that infuses symbols with their
meaning. And 'when we realize that there is indeed a gap between an
explanation of the meaning of an expression and the use of the
expression, it is altogether natural to think that it is the mind that
bridges that gap, that effects the connection between a rule and its
application. This can be variously conceived. Frege thought of a sense
as an abstract entity which determines a referent, or presents a referent
in a certain way. Grasping a sense, he thought of as an altogether
mysterious72 mental act of coming into contact with such entities.
Indeed, it would not be unlike imagining a sign together with its lines of
projection (compare Investigations §141) save that a sense, being an
abstract object, is not imaginable. In the Tractatus meaning is conferred
on signs by the will, the 'metaphysical self' that thinks the method or
projection in thinking of the sense of the proposition (Tractatus 3.11).
The wayward antagonist in Investigations §§184-243 conceives of the
mind as 'drawing the projection lines' from the symbol, or from the rule
which explains its use, to its application. This is held to be done by
mental acts or processes of understanding, or by acts of meaning (e.g.,
addition by 'plus') or by intuitive insight, or by feelings of intimation
which arise from contemplating the rule.
Pricking the bubble of these philosophical flights of fancy is one of
the tasks Wittgenstein undertakes in the Investigations' discussion of
rule-following. One by one he examines these false pictures of under-
444 G. P. BAKER AND P. M. S. HACKER
standing and of rules, and shows that they constitute a mythology of
symbolism. He does not deny that we can and often do 'grasp the whole
use of a word at a stroke', nor that when we order someone to expand the
series '+2' we mean him to go on '20004, 20006'. We can and do explain
how a given rule is to be followed, and we can and do learn how to
follow rules correctly (independently of intuitions, intimations and
causal necessitations). But these mundane phenomena must be seen
aright, not allowed to evolve into mysteries about mental acts of
understanding and logical machinery of rules. It is acting according to a
rule, a practice of normative behaviour, that lies at the bottom of our
language-games. Language, far from being a reflection of thought, is a
form of behaviour. It is no coincidence that Wittgenstein often quoted
the line from Goethe: Im Anfang war die Tat.
Kripke contends that Wittgenstein 'has invented a new form of
scepticism', that 'it is important to see that his achievement in posing
this problem stands on its own, independently of the value of his own
solution of it and the resultant argument against private language'. 73
This observation not only misrepresents the character of Wittgenstein's
preoccupations and achievements, but also fails to appreciate the
diminishing significance of scepticism in philosophy. Philosophical
scepticism played a significant role in seventeenth-century culture,
stimulated in part by the deep need for a criterion of truth within
religion (given the schism within Christianity), for a criterion of truth
for scientific theories (given the number of competing scientific
theories, e.g., of the solar system, all equally reliable within acceptable
margins of error), and for a criterion of truth between science and
religion. But perhaps the most important impetus was the manifest
conflict between the new scientific picture of reality (as consisting of
material objects possessing only geometric and mechanical properties,
and powers to affect our sensibility in such and such colourful ways) and
our ordinary conception of the world. No one could gainsay the colossal
achievements of the new science, but it opened a gulf between
appearance and reality that required explanation and justification. For
if the world as it is in itself is so different from the world as it appears to
be, how can we be certain that we can ever know anything about it as it
really is? In this cultural context it was altogether natural that philosophy
should become preoccupied with justifying the ways of God (and
His world) in the face of sceptical doubts.
Those days are long past. Scepticism, in the twentieth century, is no
ON MISUNDERSTANDING WITTGENSTEIN 445
longer a serious issue in our culture (save perhaps in the domains of
ethics and aesthetics). One may use scepticism as a colourful device to
present a genuine problem. But this manoeuvre is no more than
heuristic. Achievement in philosophy today could not consist in inventing
new forms of scepticism. The deepest cultural preoccupations of
this century turn on issues concerning language and communication.
These ramify through literature and art (from von Hoffmannsthal and
Joyce to Orwell, Borges, Beckett or Pinter as well as such artists as
Picasso, de Chirico and Magritte, Steinberg and Escher), through the
'humane studies', psychology, linguistics and sociology. It is not surprising
that philosophy has followed suit, not like the owl of Minerva,
but in fruitful symbiotic relationship with the rest of our culture.
Wittgenstein's central concerns, TM in both his philosophies, were with
the nature of language, its function and structure (cf. Investigations
§92), and the myriad philosophical illusions propensity to which is the
unavoidable condition of every language user. It is here, and not in the
invention of new forms of scepticism, that his achievements lie. His
reflections on rule-following not only undermine a conception of
language rampant in philosophy, theoretical linguistics, and psychology,
but also yield a novel and more profound conception of logical
and mathematical necessity than any yet achieved by philosophers. His
private language argument, the real private language argument, not
only undermines a tradition of philosophical thought running from
Descartes to the present day, but yields novel and more profound conceptions
of self-consciousness, of the relation of mind to body, and of
the will than any available hitherto. What exactly his conception was
needs exposition, which we have not offered save en passant. Whether
his conception was right needs argument, which we have not given,
save per accidens. What we have done is to show that it does not lie in
the arid area of sceptical questions, let alone of Humean sceptical
solutions.
NOTES
S. Kripke, 'Wittgenstein on Rules and Private Language', in I. Block (ed.), Perspectives
on the Philosophy of Wittgenstein, Blackwell, Oxford, 1982. Since we wrote this paper,
Kripke's essay has been slightly modified, extended by a postscript on other minds, and
published independently. All references here are to the original essay.
2 Ibid., p. 241.
3 Ibid., p. 242.
446 G. P. BAKER AND P. M. S. HACKER
4 Ibid., p. 296.
5 Ludwig Wittgenstein, Notebooks 1914-16, Blackwell, Oxford, 1969, p. 44.
a Ludwig Wittgenstein, On Certainty, Blackwell, Oxford, 1969, p. 65.
7 Kripke, 'Rules and Private Language', p. 268.
8 David Hume, A Treatise of Human Nature I-iv-2.
9 David Hume, 'Abstract of Treatise of Human Nature'. Cf. Enquiry Concerning Human
Understanding I-v-i, I-xii-2.
lo "Nature has not left this [belief] to his [the sceptic's] choice, and has doubtless
esteem'd it an affair of too great importance to be trusted to our uncertain reasonings and
speculations" Treatise I-iv-2.
11 See A. Ambrose (ed.), Ludwig Wittgenstein, Lectures Cambridge 1932-35, Blackwell,
Oxford, 1979, p. 97; also C. Diamond (ed.), Ludwig Wittgenstein, Lectures on the
Foundations of Mathematics, Cambridge, 1939, Harvester Press, Sussex, 1976, p. 103.
12 Ludwig Wittgenstein, Philosophical Investigations, Blackwell, Oxford, 1953, §128.
Hereafter referred to as PI.
x3 Kripke, 'Rules and Private Language', p. 273; 'quus' is a function defined by:
x~y = x + y if x, y < 57, otherwise xGy = 5. This is Kripke's operative example of the
paradox involved in rule-following.
14 Wittgenstein, PI, §187.
15 Kripke, 'Rules and Private Language', p. 239.
16 For the detailed account of this complex bibliographical history, see G. H. von Wright,
'The Origin and Composition of the Philosophical Investigations' in his Wittgenstein,
Blackwell, Oxford, 1982, pp. lllff.
17 The missing remarks, inserted only in the final draft, are: §§247-52,257,262-69, 292,
294, 297, 299-301.
18 This explanation derives from MS 180(a), 68 f.
19 Or indeed to effect the transition from seeing that it is this to seeing that it is red! And
what is the this which this is?
20 In MS 180(a), 72, the passage began: 'I was (earlier) in the difficulty that a rule could
not determine any action, since anything can be brought into accord with the rule.' There
PI §198 occurs on page 1.
zl This leads to a vicious infinite regress, as Wittgenstein repeatedly noted. No
interpretation would be complete without a rider, i.e. there would be a need for
explanations of the terms used in formulating the interpretation, a need for further
explanations of the terms used in giving these explanations, etc.
22 It is, however, noteworthy that in MS 180(a), 72 an alternative drafting to 'And so
there would be neither accord nor conflict here' was (translated) 'And so "accord" and
"conflict" here lose their sense altogether', which does not imply a paradox leading to the
conclusion that language is impossible, but implies a misunderstanding of what counts as
acting in accord with a rule and what as acting contrary to one. And what goes for MS
180(a) goes for the final version too!
23 For detailed analysis of P! §§143-184, see G. P. Baker and P. M. S. Hacker,
Wittgenstein: Understanding and Meaning, Blackwell, Oxford, 1980, pp. 621ff.
24 Given that Wittgenstein's discussion of reading intentionally runs parallel to his
analysis of meaning and understanding, it should seem puzzling to Kripke that Witt-
ON MISUNDERSTANDING WITTGENSTEIN 447
genstein did not introduce a 'new paradox of reading' viz., that since reading is not just
mouthing words while looking at writing, nor is it any mental accompaniment, therefore
there is no 'facť that constitutes my reading. Therefore, paradoxically, dear reader,
reading is logically impossible!
25 Although not consistently with the meaning of the expression (e.g., 'plus') nor with
what we understand by it. But that is just what is in question - what does determine
meaning?
26 Our translation. The German reads: "Dadurch zeigen wir n~mlich, dass es eine
Auffassung einer Regel gibt, die nicht eine Deutung ist;" there seems no license for the
translation "there is a way of grasping (Effassen) a rule". And "yon Fall zu Fall der
Anwendung' seems unhappily rendered by "in actual cases".
27 Cf. Ludwig Wittgensteirl, Remarks on the Foundations of Mathematics (3rd. ed.)
BlackweU, Oxford, 1978, p. 410.
28 Ibid., p. 335.
29 It is important to note that Wittgenstein countenances the logical possibility of
creatures being born with the ability to speak a language (cf. Ludwig Wittgenstein,
Philosophical Grammar, Blackwell, Oxford, 1974, p. 188; and The Blue and Brown Books,
Blackwell, Oxford., 1964, p. 12). How one learnt or acquired a language is irrelevant to an
account of what one has learnt. So if one could be born speaking German, would the
absence of speakers of Etruscan be a logical barrier to being born speaking Etruscan? Since
Robinson Crusoe could talk to himself, keep a diary, follow rules, would he cease to be able
to do so if, unbeknownst to him, the rest of mankind were destroyed by a plague? Obviously
not. Is his continuing to be able to do so dependent on the history of his acquisition of his
linguistic skills? That seems inconsistent with the principle that 'Teaching as the
hypothetical history of our subsequent actions.., drops out of our considerations' (Blue
Book, p. 14).
In MS 165 Wittgenstein imagines a solitary cave-man who uses a picture-language on
the walls of his cave. Such a language, he says, would be readily intelligible. Later he
imagines a solitary caveman who speaks only to himself, gives himself orders, etc.
Provided he uses simple signs which we could interpret, we could come to understand
him. A few pages further on Wittgenstein concludes that to describe the language of a
people is to describe a regularity of their behaviour, and to describe a language which
someone speaks only to himself is to describe a regularity of his behaviour, and not
something which can happen only once (cf. MS 129, 89).
Of course, we could not understand another's language unless we could grasp the rules
of his language, foIlow them as he does, agree with him in the manner of applying them.
3o Kripke, 'Rules and Private Language', p. 277.
32 Kripke, 'Rules and Private Language', p. 248.
32 Ludwig Wittgenstein, 'Notes for Lectures on "Private Experience" and "Sense Data" ',
Philosophical Review 77, 1968, 300: "The solipsist flutters and flutters in the flyglass,
strikes against the walls, flutters further. How can he be brought to rest?" Note that in MS
165 Wittgenstein states explicitly that the discussion of a private language concerns the
problems of idealism and solipsism.
33 Wittgenstein, 'N,otes for Lectures on "Private Experience" and "Sense Data" ', p. 314.
'The private experience is to serve as a paradigm, and at the same time admittedly it can't be
a paradigm.
448 G. P. BAKER AND P. M. S. HACKER
The "private experience" is a degenerate construction of our grammar (comparable in a
sense to tautology and contradiction). And this grammatical monster now fools us; when
we wish to do away with it, it seems as though we denied the existence of an experience,
say, toothache.'
34 Americans prefer the terminology of the computer age - so: given that we are all
'wired-up in the same way'!
35 Kripke, 'Rules and Private Language', p. 286. In effect Kripke's interpretation of §202
assigns a meaning to the word 'privately' which is at odds with Wittgenstein's use of the
term. Kripke holds that someone would follow a rule 'privately' if his actions are
considered in isolation from the behaviour of his community in respect of responses to this
rule. But on Wittgenstein's account it is essential that the putative rule allegedly being
followed 'privately' is one to the expression of which only I have access, i.e., it is
expressed by a private ostensive definition. Hence it is, for a quite different reason from
Kripke's, impossible for another to ascertain whether or not my applications of this 'rule'
are correct. Kripke's objection to private ostensive definitions must be that the attempt to
apply any such rule must leave the agent stranded, as it were ex officio, on his own desert
island. But this leaves open the possibility that all agents are stranded on the same island
(that public language is a congruence of private languages built separately on private
ostensive definitions).
36 Kripke, 'Rules and Private Language', p. 245.
37 The plausibility of this bizarre picture seems to turn on the now popular conception of
language as quasi-contractual, as if the distinction between correct and incorrect use of
an expression turned on the prior adoption of an explicit convention. This seems as
misguided as the now derided conviction that a social contract would provide the only
possible foundation for political liberties, rights, and duties.
38 Kripke, 'Rules and Private Language', p. 259.
39 Kripke, 'Rules and Private Language', p. 263.
4o It could also be an oblique way of explaining that the expression of uncertainty here is
senseless, as in 'Either I mean red by "red" or something else, I'm not sure'. Cf. PI
§§246-7.
4t There is, in certain contexts of 'mean' or 'intenď, a problem about self-deception. The
present context is not such a one.
42 For detailed discussion, See Baker and Hacker, Wittgenstein: Understanding and
Meaning, p. 595 ff.
43 Kripke, 'Rules and Private Language', p. 245.
44 Ibid., p. 249 f.
45 Ibid., p. 273 f.
46 Ludwig Wittgenstein, Philosophical Grammar, pp. 161 f., 199 ff., 212 ff.
47 For a more detailed analysis of Wittgenstein's resolution of the problem of the
pictoriality of thought, see P. M. S. Hacker, 'The Rise and Fall of the Picture Theory', in
I. Block (ed.), Perspectives on the Philosophy of Wittgenstein, p. 103 tI.
48 Cf. Kripke, 'Rules and Private Language', p. 281 f.
49 Ibid., p. 285 f.
50 It is curious that Kripke carefully encloses 'correcť with scare-quotes. Is it that there is
no such thing as correctly using a sign on this account? The only option Kripke has
explicitly left open is a Platonist one, viz. it is in the nature of such and such mathematical
ON MISUNDERSTANDING WITTGENSTEIN 449
object to yield such and such a value for such and such arguments (cf. Kripke,
'Wittgenstein on Rules and Private Language', p. 264 f.). But this is certainly not
Wittgenstein's view. And Kripke does not explain what he thinks Wittgenstein's view is.
51 Or at least, in the arithmetical case, that he goes through the correct procedure.
52 To this the reply might be that the requirement is not merely that he gives the same
reply I am inclined to give, but rather that given the same information, he gives the same
reply. So given that he knows that Genghis Khan was an unmarried man at the age of 22,
he would answer the question of whether he was a bachelor in the same way as I am
inclined to, given that information. This may be conceded; of course he will give the same
answer - because he knows (as I do) what 'bachelor' means. It is not our agreement that is
the ground for the judgement that he means bachelor by 'bachelor', but rather the
explanation of its meaning which he gives and the applications of the word that he makes.
The agreement is a framework condition for the language-game (in#a, p. 437 f.), not part
of the rules of the game.
53 Parallel to Kripke, 'Rules and Private Language', p. 245.
54 Ibid., pp. 277, 284.
55 Ibid., p. 296; a similar bizarre view is expressed by C. Peacocke in his defence of what
he calls 'The community view' of rules: 'The community view can count such a person [as
Crusoe] as a genuine rule-follower if he reacts to new examples in the same way as would
members of our community, or of some other conceivable community'. See C. Peacocke,
'Reply: Rule-following: The Nature of Wittgenstein's Arguments', in S. Holtzman and C.
Leich (eds.), Wiugenstein: to Follow a Rule, Routledge & Kegan Paul, London, 1981.
Note that the subjunctive, together with the requirement of mere conceivability of a
community, robs the qualification of any restrictive content whatever.
56 There is something very wrong about recent talk of 'following one's inclinations' when
applying rules. We do not teach children arithmetic by teaching them to follow their
inclinations. We do not even teach them to have the same inclinations as we have. We
teach them to follow arithmetical rules, we teach them that getting such and such results is
what counts as following this or that rule. When we hit bedrock we do indeed follow the
rule blindly, but that is not to say that we follow our inclinations blindly!
57 A similar discussion of the language of Robinson Crusoe 'considered in isolation'
occurs in MS 116, 117, where Wittgenstein examines a distinction between subjective and
objective understanding. Crusoe could certainly play language-games by himself, Wittgenstein
remarks. If one secretly observed his sign-using activities, and if one discerned in
them certain kinds of complex regularities, one would rightly judge him to be using a
language of his own.
58 Kripke, 'Wittgenstein on Rules and Private Language', p. 286.
59 Kripke, 'Wittgenstein on Rules and Private Language', p. 289.
60 See Wittgenstein, Remarks on the Foundations of Mathematics, p. 323: 'peaceful
agreement.., belongs to the framework out of which our language works'. Cf. PI §240.
6a Wittgenstein, Remarks on the Foundations of Mathematics, p. 365.
62 Ibid., p. 330.
63 Ibid., p. 392.
64 Ibid., p. 392.
65 Ourselves (in the past) included. Kripke's position differs substantially from Dummetťs
in respect of the issue of anti-realism, for nothing in his remarks about assertion-
450 G. P. BAKER AND P. M. S. HACKER
conditions commits him to Dummetťs anti-realist doctrines. While Dummetťs interests
lie, as it were, in the geometry of proof, Kripke's lie in the sociologyof meaning.
66 For a detailed analysis of the different conceptsof truth-conditions, from the Tractatus,
through Tarski, Carnap and onwards, see G. P. Baker and P. M. S. Hacker, Language,
Sense and Nonsense, Blackwell, Oxford and New York, 1984, chapter 3.
67 Kripke, 'Rules and Private Language', p. 273 f.
68 Ibid., p. 274.
69 See G. P. Baker and P. M. S. Hacker, An Analytic Commentary on the Philosophical
Investigations, Volume 2 (forthcoming).
70 The metaphysical object, he now thought, was an illusion. Its role was filled by
samples, conceived as belonging to the calculus itself.
71 Wittgenstein, MS 165, p. 82.
72 Gottlob Frege, Posthumous Writings, Blackwell, Oxford, 1979, p. 145.
73 Kripke, 'Wittgenstein on Rules and Private Language', p. 267.
74 This is not to say that his discussions of scepticism in On Certainty are unimportant,
merely that it would be misconceived to represent these themes as lying at the heart of the
Philosophical Investigations.
St. John's College
Oxford
England