This paper was written for an
audience of one: the course instructor, Prof. Henson.
Professor Henson’s marginal comments (indicated by "RGH")
and my responses ("EP") will be added to the end notes. The
paper earned a "91" from Professor Henson, which I take to
mean an A-.
"Let us introduce two antithetical terms in order
to avoid elementary confusions: To the question 'How
do you know that so-and-so is the case?' We
sometimes answer by giving 'criteria' and sometimes
by giving "symptoms.' If medical science call angina
an inflammation caused by a particular bacillus, and
we ask in a particular case 'why do you say this man
has got angina?' then the answer "I have found the
bacillus so-and-so in his blood' gives us the
criterion, or what we may call the defining
criterion of angina. If on the other hand the answer
was, 'His throat is inflamed,' this might give us a
symptom of angina. I call 'symptom' a phenomenon of
which experience has taught us that it coincided, in
some way or other, with the phenomenon which is our
defining criterion. Then to say "A man has angina if
this bacillus is found in him' is a tautology or it
is a loose way of stating the definition of
'angina.' But to say, 'a man has angina whenever he
has an inflamed throat' is to make a hypothesis.
"In practice, if you were asked which phenomenon
is the defining criterion and which is the symptom,
you would in most cases be unable to answer this
question except by making an arbitrary decision ad
hoc. It may be practical to define a word by taking
one phenomenon as the defining criterion, but we
shall easily be persuaded to define the word by
means of what, according to our first use, was a
On the surface, the concept of "criterion" seems ordinary and
commonplace. In Wittgenstein's hands, it has been greatly refined
and elaborated, to the point at which it promises to unravel some
long-standing philosophical tangles. But, like many a philosopher
with a new, significant and powerful conceptual tool, Wittgenstein
(and especially his Boswell, Norman Malcolm) has been tempted to
over-reach. It appears that a great deal of contemporary
philosophical analysis has been devoted to a careful examination of
the scope of application of his kit of philosophical tools. In this
paper, I shall add to this list. Specifically, I will contend that
Wittgenstein and Malcolm have over-reached in attempting to
comprehend the meaning of "private" psychological states in terms of
public criteria. I will further contend that their attack on
"private languages" does not succeed and thus does not compel us to
accept their criteriological theory of the belief in other minds.
What are the important features of Wittgenstein's conception of a
Two thoughtful examinations of Wittgenstein's concept of
"criterion" have been offered by Albritten2 and Wellman3.
Both allow that the quotation that heads this paper is the closest
that Wittgenstein came to giving us an explicit definition of
"criterion." Wellman then suggests a "definition which seems to be
implicit in his use of the term throughout his writings." The
definition: "A criterion is a purely linguistic ground for judging
that it is or is not correct to apply a given expression to some
object." (Wellman, 1961, 162) He elaborates:
A criterion may be said to be a criterion for a
linguistic expression fitting its object. By an object, I do
not necessarily mean a physical object, but whatever the
expression refers to, is applied to, or is about. By
fitting, I mean is true of or applied to, but in a very
special way. An expression fits an object when it is
linguistically, rather than factually, correct to apply it
to that object." (156)
And how do we distinguish "criteria" from "symptoms?"
Criteria are observable features which are directly
connected to an expression by its meaning; symptoms are
features which are indirectly connected to the expression by
being associated with the criteria in our experience, To
justify one's use of a description by giving criteria is to
appeal to a convention; to justify one's use of a
description by giving symptoms is to appeal to an empirical
The concepts of "criterion" and "symptom" suggest the elementary
distinction between "defining" and "accidental" characteristics -- a
distinction that is rarely bypassed in an introductory course in
Philosophy. These might terms might offer us an avenue into our
points of contention with Wittgenstein and Malcolm.
Consider the following sentence: (a) "All crows are black,"
then allow me to present to you a bird that cries "caw caw", has the
anatomical parts and shape of that which we call a "crow," etc. --
but it is orange. How does this presentation effect statement
(a)? There are two possible responses: (1) A. is false, for here is
a non-black crow; or (2) This thing is not a crow, for after all it
is not black, and "all crows are black."
In the former case we can imagine a specimen that lacks the
characteristic (even if, in fact, we fail to encounter it), and yet
we are willing to apply the term in question, even so. The
characteristic does not define the concept -- it is, what
Wittgenstein could call a 'symptom." In the second case, by "crow"
we mean "a thing that has characteristics A, b, c, n, and is
black." Lacking any of these, it is not a crow. Thus we could
identify blackness as a "defining criterion" of the concept "crow."4
All this is quite clear enough, simple and elementary: in Philo. 1
jargon, once more, statement (a) interpreted as case (1) is
"synthetic", and as case (2), is "analytic."
Well, it is one thing to conjure up a textbook case. How does
this distinction apply in use -- in the language-games we
play in ordinary life? (And this, after all, is the arena in which
Wittgenstein urges us to pursue our Philosophical Investigations).
In some cases, the distinction is easy enough to recognize, but it
is most clearly found in those language games that involve
deliberate attempts at clarity, e.g. the sciences and the law. In
such "games" explicit stipulations are to be found with precise
listings of defining criteria. E.g. If a plane figure is bounded by
three straight lines, then its a triangle because, in our geometry,
that's what we mean by a "triangle."
However, once we leave the realm of exact stipulation and
consider ordinary language, we face considerable difficulties.
Consider, for example, the simple lemon.5 What
characteristics must be present for us to consider this object a
lemon? We can collect a list, easy enough, but we will soon find
that none of these alleged "criteria" are indispensible. You could
collect a basket of "lemons" and, perhaps, find yourself hard-put to
identify one characteristic common to all the lemons. (It will be a
strange collection, with some odd looking 'lemons" (?) -- but such a
situation is conceivable, even if the specimens are not to be found
now in California, Florida, or anywhere else). The common term
"lemon" will refer, not to a set of individuals with a common
"essence," but to a collection of individuals displaying a "family
resemblance" (one of Wittgenstein's most suggestive and useful
concepts) A lemon is a lemon, not if 'it has, specifically, traits a
& b & c . . . & n, but if it has a "quorum"6
of such traits. None in particular, but enough! How many are
enough? Well, if they are "strong criteria , then maybe a few will
suffice. If they are "weak criteria", then it will take a bunch of
"Wait a minute," one might protest, "let's pin this down" How
many of which criteria, in what combination." Alas, we cannot "pin
this down." That's just the way the ordinary-language-game is
played: inexact, sloppy and changing its rules in mid-play! The best
that we might do is collect a few "normal" (?!) language-users and
take a poll. Come to think of it, this sort of thing is commonly
done -- the result is called a "dictionary." But when such a survey
is conducted, we just don't expect too much precision when we get
down to the fine points.
I hope that the reader will forgive this survey of rather
commonplace "practical logic." It has, however, led us to some
substantial problems. Prominent among them is the following: What
now of that clear distinction, with which we began, between
''symptoms" and "criteria"? Wittgenstein himself admits that, on
close examination, the line is fuzzy. "In practice, if you were
asked which phenomenon is the defining criterion and which is the
symptom, you would, in most cases, be unable to answer this
question," for "in general we don't use language according to strict
rules." (Blue and Brown Books 25)
Now this doesn't destroy the usefulness of the concepts of
"criterion" and "symptom", nor does it deny that there is a
difference. But it does indicate that the boundary between can be
wide and indistinct – just as we know the difference between blue
and green without being able to denote a clear line of demarcation.
(Just which color is aqua?) The significance of this point can
scarcely be understated. Let's pursue it.
We have, on one hand, the insistence that "the satisfaction of
the criteria of y establishes the existence of y beyond question; it
repeats the kind of case in which we were taught to say Y".
(Malcolm, 1954, 87) "The satisfaction of the criterion of a thing's
being the case entails that it is the case, which no symptom. . .
can do." And yet, "it doesn’t follow that no criterion can seem to
be satisfied when it isn't, or that no symptom can be perfectly
reliable." (Albritton, 849).
But how do we distinguish between a correct "seeming to be
satisfied" from an incorrect "seeming..."? How do we resolve
Wittgenstein's admitted inability to distinguish between phenomena
that are "defining criteria" and those that are "symptoms" in a
language without "exact rules" (which means any "natural
language"*)? We find ourselves in search of an exact, unequivocal
criterion of "criteria." In a natural language, we shall search for
it in vain.7 And if our search fails, (as it must) then
our hope for a "criterion" that will "entail" (Albritton), or that
will "establish beyond question" (Malcolm) is futile.8
The point might be illustrated by the following: The Pope is said
to be infallible when he speaks ex cathedra on matters of
faith and morals. The trouble is that there is no infallible test
(for the Pope or his flock) to determine whether or not any given
pronouncement is ex cathedra. This is equivalent to saying
that the Pope is not infallible.9 So too, with no
"meta-criteria," there is no clear identification of a criterion.10
With no such identification, there is no entailment. A "criterion,"
then, appears to be a strong indicator, and a "symptom" a weak
With a backward look at the Tractatus, Wittgenstein
observed that "the more narrowly we examine actual language, the
sharper becomes the conflict between it and the requirement. (For
the crystalline purity of logic was, of course, not a result of
investigation: it was a requirement)." [Investigations, 107]
Following another metaphor of Wittgenstein, Malcolm remarks that
"what we need to do is to remove from our noses the logical glasses
through which we look at reality. We must study the phenomenon of
language as it is, without preconceived ideas." [1954, 88} Good
advice, both! We must not let the logical pictures "hold us
captive." [Wittgenstein, Investigations §115]
What does all this have to do with our recognition and
identification of subjective states?
Really, not too much!
"Well, then, why bring it up in the first place?"
We bring it up simply because Wittgenstein and Malcolm believe
that the concept of "criterion" has a great deal to do with our
understanding of psychological states. We aim to show that the
concept of criterion is inappropriately applied to the philosophical
problems of "privacy" and "other minds." Let us then return to our
If we have not only a "full quorum" but a "packed house" of
criteria (of a lemon), then, by George!, we have a lemon! It
is that which has these criteria. With psychological states, we have
a fundamentally different situation – moreover, a fatally different
situation for the Wittgenstein-Malcolm criteriological theory. For
with private states, no amount of objective, behavioral or
expressive (take your pick) "criteria" will add up to a pain, or a
pleasure, or whatever. If the users of a language-game are in
agreement about what constitutes a quorum of defining criteria of a
public thing or event, then the presence of this quorum will mean
that this is a such-and-such. To quote Gov. Reagan: "If it quacks
like a duck, swims like a duck, and looks like a duck, then it’s a
duck." Nothing else will dissuade them for it will not overwhelm the
way they choose to use the word (or, if you prefer, "play the
language game"). Bu no amount of behavioral display or linguistic
expression will add up to the same thing as a person’s
feelings. There can be no "entailment" here. A strong indication?
Yes. Beyond practical doubt? Most likely.
But this quite misses the point. These locutions are in the
language of evidence.-- of symptomatology. But "criteria," we
are urged, have to do with identifications. The presence of
"enough" of the criteria must, of themselves, finally settle the
question of what this thing is. There can be no other, unaccounted
for sine qua non. But not only is a subjective state a
"something-more-than" behavior, speech and objective circumstance –
it is a fundamentally-something-more. It is a something-more that no
amount of "outside" manifestations can entail.12 Yet
without the subjective component, all these "outside" criteria are
simply criteria of nothing -- a shell without an egg. These
alleged "criteria" can only indicate, albeit they can
indicate up to the point of "practical certainty." But all this can
be said of "symptoms." And we are left with the question: How do we
come to believe that these behavioral indicators are symptoms of
"other minds", or for that matter, of the "same" state of our own
It would, seem that if we wish to hold on to a "criteriological"
theory that will yield us knowledge "beyond question" of subjective
states, then we will have to suppose that these alleged subjective
states are these behavior traits, linguistic responses and
environmental circumstances. We will have to deny that there can be
a super-stoic who endures a pain without displaying symptoms. We
must discount the neglected child or the habitual liar who can feign
enough pain criteria and yet have no pain. We must, in short, be
old-fashioned Watsonian behaviorists and deny consciousness and
sensations. And this is what neither Malcolm or Wittgenstein wish to
do. Both allow that there is a subjective "something else.13
"Well then, if the pain criteria can never be the pain,
then the criteria must not be all that we mean by "his pain." That
is, they are not "criteria" at all, but variably adequate
"symptoms."14 This surely is an unstartling and
commonplace conclusion to come to. And so we are left where we
started – trying to account for our knowledge of another's pain --
trying, that is, to explain why we take the behavioral
"super-symptoms" (read "alleged criteria") to be
indicators of "his pain."
Malcolm tries to rescue the situation by suggesting that we don't
need logical certainty after all. He quotes Wittgenstein's
remark that we "can be as certain of someone else's sensation as of
any fact.:"15 [Investigations, 224] "Just try – in a real
case – to doubt someone else's fear or pain." [Investigations §303]
"Perhaps," says Malcolm, "we can imagine a doubt; but we do not take
it seriously . . There is a concept of certainty in these
language-games only because we stop short of what is conceivable."
[Malcolm, 1954, 89] This means, I presume, "short of logical
entailment." Still further:
The man who doubts the other's pain may be neurotic, may
'lack a sense of reality,' but his reasoning is perfectly
sound. If his doubts are true then the injured man is not in
pain. His reaction is abnormal but not illogical. The
certainty that the injured man is in pain (the normal
reaction) ignores the endless doubts that could be proposed
and investigated. [Malcolm, 1954, 90]
Let's drop a flag on this play! Is Malcolm employing a double
standard here? Does he demand of his rivals a logical purity that he
does not require of himself or of Wittgenstein? (A common enough
failing amongst philosophers). Let's try it out. Where, above, he
quotes Wittgenstein as saying "Just try – in a real case – to doubt
someone else's fear or pain," let us substitute ". . . to doubt that
that human figure has a mind." Now let's follow through with
Wittgenstein’s challenge.16 Cannot the analogy argument
for other minds, or even still other theories, be justified in just
such a manner? Could not these rivals get by with just such a
less-than-total standard of certainty? But to apply the test of
logical doubt is not to pretend that we seriously entertain such a
doubt, but only that we can conceivably do so. Who in fact is a
solipsist? Yet Malcolm uses a reduction to solipsism to refute the
analogical argument. [Malcolm, 1958,156-7] Would not the same
standard of total logical rigor refute Malcolm's own criteriological
But doesn't the "expression theory" cover up these difficulties?
After all, if I am in pain, then my report "it hurts" can not
possibly be incorrect. I can't even be said to "know" I am in pain,
for it is inconceivable for me to be in pain and not to be aware of
it. And so, with no counter-concept, there is no concept. The pain
and its expression are therefore incorrigible, and "that the natural
pain-behavior and the utterance 'it hurts' are each incorrigible is
what makes it possible for each of them to be a criterion of pain."
Let us grant the entire list of premises above (although, in
fact, I do not believe we must). The criteriological theory remains
unfounded. The utterance "it hurts" may be incorrigible. But only to
me -- the "first person." But I am in no need of behavioral criteria
to tell me that I have a pain (as Ziff well argues). The situation
is wholly different to the second person. For reasons already
argued, the expression "it hurts" is not incorrigible to the
audience. The hurting hurts only the first person. It is not
thus immediately available to the second. The first person gets the
egg, the second merely the shell. The expression of the first person
is a contingent symptom, quite conceivably faulty, as may be
the audience's perception or judgment of it. It is, accordingly, no
criterion at all!
In summary, then, it would seem that the Wittgenstein-Malcolm
concept of "criterion' proposes a strange hybrid indeed! Criteria
can "logically establish", but do not "entail" (at least in the case
of subjective states). Expressions of sensation are "incorrigible"
but in a contingent sort of way. We can know that x is in pain
'beyond question" provided that a certain unspecifiable number of
conditions are not in force. It would be most helpful to have an
explanatory method that has the advantages of logical force without
its shortcomings – but to accomplish this, one had better come up
with a new logic, and explain it in detail. I cannot see that
Wittgenstein or Malcolm offer us any such thing. Moreover, I
seriously doubt that such a logic, devised to do the work they
intend, is at all possible – or if possible, of any use.
But if we invalidate 'criteria', what is left to account for
Well, perhaps we will just have to defend "private language" from
the assaults of Wittgenstein and Malcolm. If we succeed, we then
might be able to make sense of (a) "I have a pain;" (b) "Its the
same sort of pain that I had last week;" (c) "You seem to have the
same sort of pain."
Perhaps the most telling defense of "private languages" can be
derived by challenging the critics, Wittgenstein and Malcolm, to
account for some perfectly commonplace events. If they are correct,
and we cannot in fact make sense of "private rules", then how can we
account for the following poignant little scene;
Mother; "Johnny, its time for you to go to the Dentist."
Johnny; "I don' wanna. Last time he took that big drill, and he
hurt me, somethin' awful!"
Now it isn't the drill and probes that causes Johnny to protest.
It is what these instruments did to the "private Johnny." It is not
his holding of his cheek or his yelling, but that which caused these
expressions which provoke his present anguish. He might conceivably
have completely forgotten all about the drills and probes and
yelling, etc. But what he does remember, and quite well indeed, is
that "it hurt." If a philosophical Johnny is asked to explain what
he means by "it hurt" per se – not what caused it, or what
expressed it – he may find himself at "the point where one would
like just to emit an inarticulate sound." [Wittgenstein,
Investigations §261]He may be at a "basic fact." But to say that he
doesn't understand it, or that he cannot remember or imagine
it, or even that it has no meaning to him – all this is clearly
false. He very well understands what "it hurt" means, and He'll try
almost anything to avoid going through the "same experience" again.
Such cases could easily be multiplied. It seems quite clear that
we do in fact recognize and compare our "criteria-less" sense
recollections, and do attribute some psychological states to others.
Wittgenstein and Malcolm admit as much too. But, in abolishing
"private languages" they undercut the grounds for this behavior. To
them, it makes no sense to say that we do these things for, recall,
it is not the alleged criteria that we are remembering and comparing
– it is those private somethings that the "outside" indicators;
We could continue this frontal attack upon the critics of private
languages. I suggest a different approach. I prefer to shrug it off
with a casual 'so what?" The accomplishment of Wittgenstein and
Malcolm [RGH which] is, from a logical point of view, subtle,
clever, and in a way, rather elegant. But really, what does it all
What if we have no "criteria" for determining that "this pain" is
the same as "last week's pain." What difference does this make?
Suppose that we did? Suppose that there were some "marks",
say a flash of colored lights, to accompany our pains. 'Jut how
would we know that the "same" color accompanied the "same" pain.
Even if we did, how would we determine that the same color-pain
appears now as appeared before? Well, with imagined colors, we can
go "outside" and (say) check with a color chart. So we might note
that this is the same "color of pain" as that of last week, which I
also checked with the chart. But what criteria are to tell us that
our faculty for matching pain-colors is the same now as it was then?
To put it generally, if we seek to compare two things, we then must find criteria of comparison. If we
determine that the "mark " of identity on item A is the same as the
"mark" of identity on item B, then A and B are the "same" in that
respect. But how are we to determine that the marks that identified
the 'sameness" were in fact the same in each case? Well, lets look
for some super-marks that will apply to both marks in the "same way."
Of course, we have an infinite regress. We simply have to begin
somewhere. There must be a criterionless criterion.
In a sense, we do in fact have a "mark", albeit faulty, to
indicate, e.g., that today's pain is "the same as" that of last
week. It is the "it-seems-to-be-the-same" feeling. And this just
might be an ultimate. But any decision of "sameness", subjective or
objective (e.g. comparing instrument dials), must eventually,
through a hierarchy of "marks", be brought to this ultimate mark of
Now this is not to suggest that the "it-seems-to-be-the-same"
criterion is the court of last resort Were this the case, we never
would have discovered that the deja vu experience is illusory
-- the feeling that "I've been here before," or "I've met you before
somewhere" when a quick autobiographical canvas reminds us that we
most certainly have not. But this only shows that we often have a
variety of means by which to verify, or correct, our recollections.
This is true in that most cited of cases, the sensation of pains.
With pains we do often have the additional support of "first
person analogies" Consider for example: "This thumb pain seems like
the same pain as the one I had last week. Come to think of it, it should, since I hit it with a hammer then, just as I did right
now!" Of course we can ask, "how are you sure that you hit your
thumb then in just the same way as you did right now?" Well, we can
call upon further evidence. Perhaps someone else (!) was present at
both mishaps. But how does he know it was the same sort of
accident? Or we could say that the hammer must have hit me the same
way because it hurts me the way it did before. "But that's circular
You can't tie two horses to each other and expect to hold them --
you've got to find a post." Well, maybe. But if you tie enough
horses together, they'll be a rather stable herd.
This suggests another answer to the attack upon "private
languages." Philosophical analysts have the understandible habit of
contemplating bits, pierces and parts -- in and of themselves, and
as they interact. Is it then not possible that Wittgenstein and
Malcolm have overlooked the fact that psychological states come in
"clusters"? This fact, and some tried and true principles of
associational psychology, might well lead us out of the
Consider the argument that Malcolm gives against recognizing "the
One will be inclined to say here that one can simply
remember this sensation and by remembering it will know that
one is making a consistent application of its name. But will
it also be possible to have a false memory impression? On
the private-language hypothesis, what would show that your
memory impression is false -- or true? (1954, 77)
In response, let's return to Johnny:
Mother: "Johnny, you are going to see the dentist today."
Johnny: "Oh boy, lucky me!"
Johnny is soon to find our that his memory impression of his last
visit is false. But let's leave this unfortunate child and turn to
our own situation. I propose the following experiment: think of the
pain of a pin prick. Got it? OK! Now take this pin and push it
against your finger. Do your sensation and your memory match up? If
so, they are the same sort of sensation! But, you reply (bleeding
finger in mouth), "how do I know.that they are the same --
what validates this impression of sameness?" Maybe the
impression of the sameness of sensation is all that you have.
But this is no less than a scientist in his laboratory has when he
matches readings of his dials, or a musician, when he tunes to A, or
two witnesses, when they report "the same" event. Surely this should
meet 1alcolm's less-than-complete test of certainty!
Now of course we don't go about testing our sense recollections
in such a way, each time we have them. But when we test them, they
seem to work out. Moreover, as noted, our "seemings of sameness"
seem to be accompanied by clusters of other familiar experiences --
that is, we are constantly involved in tacit "testings," of which we
take no especial notice, except when something doesn't pan out (as
with the deja vu experience).
Now the criteriologist will reply, "that's just what I have been
saying all along! The pin against the skin is a criterion of
the pain!" I hope that I have, by now, shown that it is no more than
a test -- a "symptom" (here the word "symptom' seems queer indeed).
I can have the pain without the pin, or I can be pricked without
feeling the pain (with a post-hypnotic suggestion or an anesthesia).
It could feel different. In fact, if it did -- if,
say, it caused a dull ache or a tickle -- I would recognize that
this sensation just isn't the same thing, which indicates that
something else is operative here! The pin offers just a partial
test. But enough of such tests will suffice. Our experience comes in
clusters. When these clusters verify and strengthen our impressions
of "sameness", life goes on as usual. When something is out of order
-- not the same -- the strangeness of it jolts us and brings it to
our attention. The fact that this happens as rarely as it does,
gives strong validation to our assumption that the same sort of
stimulus results in the same sort of sensation response.
.What then of these other biped figures about us? How can we tell
that their "minds" have the same sort of experiences as our do?
Well, unlike Wittgenstein's beetles, these things act back. A doctor
examining a patient will be told the time of occurrence of a pain,
its location, intensity, "sharpness," etc. This we can imagine in
ourselves. Indeed, we have a common descriptive language of pains.
In his attempt to reduce the analogy argument to solipsism,
Malcolm argues that "pain' cannot mean the same for the first and
second person. "When I say 'I am in pain,' by 'pain' I mean a
certain inward state. When I say 'He is in pain,' by 'pain' I mean
behavior." [1958, 157] If experience is cut into analytical bits,
then we may well find ourselves in such a trap. But if we hold to
the clusters which our experience presents to us, then Malcolm's
observation would read "When I say 'I am in pain,' by 'pain' I mean
a certain inward state which prompts my behavior. When I say
'He is in pain,' by 'pain' I mean that which prompts his
behavior.'20 The step to the analogy argument for
other minds is then quite simple. "If the behaviors are alike [and
"alike" would be Malcolm's prime target here], then since I know the
feelings that accompany my behavior, such things may be
presumed to accompany his behavior." Or, to put it in tabular
form, it is quite true that I feel my pain and not yours, and so:
From Cut =
Wish to avoid =
Wish to avoid
As we add to this list, and repeat this and similar experiences,
the analogy will grow stronger, which in fact it does. This is
called "maturation." But isn't this just the old analogy argument?
Perhaps it is this -- and more. Perhaps we are dealing here with
analogies that are not merely 'arguments," but are fundamental
conditions of human thought and growth!
It is not my purpose here to defend the analogy argument.
However, I hope to demonstrate that analogy, or perhaps some other
hypotheses, are opened to us once we are willing to accept that our
"private rules' -- our recollections and identifications of private
states -- come out of a complex experience-in-process. If we
analytically dissect experience, we may find ourselves lost in
confusion and paradox. Wittgenstein and Malcolm have taken the clock
apart and have failed to find the "tick." This is understandable. I
trust that we will be forgiven if we continue to believe that it
makes sense to speak of and to recognize the ticking of our working
What, after all, does Wittgenstein have to tell us anyway.
A great deal!
Lest I conclude with an overly harsh note of criticism, I would
like to add my voice to those who have acknowledged Wittgenstein to
be one of the most significant of twentieth century philosophers. I
opened this paper with the hypothesis that Wittgenstein and Malcolm
had "over-reached" in applying their concept of "criterion" to
private states. This is, I believe, a vulnerable target among a fund
of ideas that includes some insights of lasting significance.
Indeed, I get the impression that much of the vulnerability of the
criteriological theory comes from a corruption, by Malcolm, of the
sound doctrine of Wittgenstein. In his zeal, Malcolm may have rushed
in where Wittgenstein trod with great care. So it is that,
concerning psychological states, Wittgenstein seems the more
interested in expression theory, and Malcolm the more interested in
I had little to say about the expression theory of sensation
statements, precisely because I find it persuasive and productive.
After all, criticism is a much more stimulating philosophical game
than agreement, The expression theory is strongest, I believe, as an
account of the genesis of subjective language in the
individual (to which Wittgenstein gives due attention.
[Investigations, §244]). The theory becomes less persuasive as an
account of the sensation language-games of the adult. A metaphor
from the Tractatus [6.54] is suggestive here. Is it not
possible that the "expressive" use of sense-language serves as a
ladder, upon which the child ascends until he reaches the point at
which he can conceptualize and abstract -- that is, treat his states
of feeling as objects of thought. Once he reaches this point, he can
discard the ladder. (I don't wish to pretend that such words as
"ouch" are absent from adult vocabularies). Once feeling-concepts
are acquired, then it is not difficult to account for a belief in
other minds -- say, by analogy -- provided that we are willing to
settle for an inductively strong but deductively leaky inference.22
Of the lasting value of Wittgenstein's general account of
"criteria,' volumes have been, and will continue to be written. He
has left to the philosophical community a vast supply of rich, if
often unrefined, ore. Contemporary philosophy is asking his
questions and laboring in his vineyards. This is the truest tribute
that ca n be given a critical philosopher. I will not attempt to
list his lasting contributions, but will close with but one example
that grows directly from his examination of 'criteria." I can
scarcely improve upon Wellman's account:
Since the time of Socrates it has usually been assumed
that all the instances of a descriptive predicate have some
one thing in common. This assumption has been accepted both
by those ontologically oriented philosophers who conceive of
philosophy as discovering the real natures of the basic
kinds of entities, and by those analytic philosophers who
aim only at stating precisely the necessary and sufficient
conditions for the application of some linguistic
expression. To have challenged this assumption and to have
partially explored an alternative to it is surely a major
contribution to the progress of philosophy. [1962,163]
To which I can only add, amen!
NOTES AND REFERENCES
1. The Blue and Brown Books, Harper
2. Albritton, Rogers, "On Wittgenstein's Use of
the Term ‘Criterion’," Journal of Philosophy, 1959. p, 845.
3. Wellman, Carl, 'Wittgenstein's Conception of
a Criterion," Philosophical Review, 1962.
4. Do criteria refer to expressions, their
uses, characteristics, or what? Apparently any or all of these --
and more. Wittgenstein gives us no clear guidance on this. (Cf.
Wellman 1961. 155).
5. As does Scriven ("The Logic of Criteria")
and Wellman (1961
6. This marvelous metaphor is from John
Hospers’ Introduction to Philosophical Analysis
7. A point that I am prepared to defend at
length -- but cannot do so in this space. Cf. my paper "Whatever
Became of Logical Positivism; or The Vienna Circle Squared." (Alas,
unpublished -- the last I heard).
8. RGH: "Doesn’t follow that there are no
clear cases of criteria."
EP: Agreed! As Witttgenstein says, we
can show things that we can’t define."
9. This may not be an accurate account of the
Roman Catholic doctrine. However, true or false, it serves to
illustrate my point
10. Again, let me acknowledge that "exact"
language-games such as science can almost provide meta-criteria by
stipulation. But even so, the terms of the stipulation must
eventually be derived from the natural language. (The "open texture
of language.") Formal languages such as logic and mathematics
constitute an additional case, but do not bear directly upon our
11. Grice and Strawson's article "In Defense of
a Dogma" (Philosophical Review, April, 1956) is roughly a defense of
the sort of dogma I am attacking here -- a sharp analytic-synthetic
distinction. However, my position is much milder than that of W v0.
Quine, whose paper "Two Dogmas of Empiricism" provoked the Grice and
Strawson rejoinder. Moreover, even if the gist of the G & S paper be
allowed, this would not, I believe, rehabilitate the criteriological
theory. But to argue this would require still another paper, and I
have papers enough, thanks!
12. "Do the propositions that describe the
criterion of his being in pain logically imply the
proposition 'He is in pain'?" W ittgenstein's answer is clearly in
the negative. Pain-behavior is a criterion of pain only in
certain circumstances." [E.g. He is not rehearsing a play, or is
not hypnotieed. And, says Malcolm, there is an "indefinite" list of
such circumstances]. "Therefore, entailment conditions cannot be
formulated; there are none."
13. Says Malcolm, "Wittgenstein does not deny
that there are inner experiences any more than he denies that there
are mental occurrences." [1954, 94. See context of this quotation
for citations from Nittgeristein that appear to support this
observation of Malcolm.
14. RGH: Not strictly right: A could be a
logically necessary or sufficient condition of B without being
either identical w B or a symptom of B. I.e., these two sentences
seem to mean that "identical with" and "symptom of" are the only
relationships we could have.
EP: Very well – then
the question remains: how do we establish necessary or sufficient
(or whatever) relationships between pain and pain-criteria if we do
not, in fact, have an independent notion of what a pain is – and
therefore what "pain" means?
15. Once more, an admission that this is a
factual, rather than a formal question. [RGH: What is
"this"? The question whether someone is in pain?]
16. Better yet, try this out in the original
[Malcolm 1954, 89] (I am "plugging in" the analogy argument above).
17. RGH: Good! Though it would be well to spell
out exactly where too much rigor is required – in the theory or in
the inference to someone’s pain.
18. RGH: Don’t see this. "Hurt," "pain,"
al, are bits of the common language, which Johnny has
learned to use like the rest of us. The are not parts of a
EP: Sure – but
why are we able to compare pains recollected – "like this," but not
"like that" – unless there is a private somethings before us
to be attached to the "common language"?
19. RGH: Plausible, at least.
20. If we prefer to dismiss the causal language
as unwarranted, then we might use Hume's ploy and substitute "which
is accompanied by' for "which prompts."
21. This I will have to check out as I read
ahead in The Blue and Brown Books, and the Philosophical
22. This is but a suggestion that is in great
need of study and expansion. So we'll just put a number on this
paragraph and call it an "Investigation."
Albritton, Rogers, "On Wittgenstein's Use of the term
'Criterion,'" Journal of Philosophy, LVI (1959) 845.
Ayer, A. J. "Can There be a Private Language?"
Society Supplementary Volume, (1954), 63-78. Reprinted in Morick
Ayer, A. J. The Problem of Knowledge, Pelican Books, 1956.
See especially Chapter 5, "Myself and Others."
Casteneda, Hector-Neri, "Criteria, Analogy, and Knowledge of
Other Minds," Journal of Philosophy, LIX (1962) 533
Chappell, Vere, 'Introduction" to The Philosophy of Mind,
Englewood Cliffs, N.J.: Prentice Hall, 1962. (ed. Chappell ),
Chisolm, R. M. Theory of Knowledge, Prentice Hall, 1966.
Grice, H. P. And Strawson, P. F., "In Defense of a Dogma,"
Philosophical Review, LXV (1956), pp. 141-158.
Hartnack, J., Wittgenstein and Modern Philosophy, transl.
A. Cranston., Garden City, N.Y.: Anchor Books, 1965.
Hospers, John, An Introduction to Philosophical Analysis,
Prentice Hall, 1967.
Kenny, Anthony, "Criteria," Encyclopedia of Philosophy,
Macmillan Co. The Free Press, 1967. Vol. 2, p 257.
Malcolm, Norman, "Knowledge of Other Minds,"
LV (1958), 141. Reprinted, Chappell (above p.
Malcolm, Norman, ''Ludwig Wittgenstein," The Encyclopedia of
Philosophy, The MacMillan Co. and The Free Press, 1967. Vol. 8,
Malcolm, Norman, "Wittgenstein's Philosophical Investigations,"
Philosophical Review, LXIII (1954), p. 540. Reprinted in
Chappell (above), p.74.
Morick, Harold, "Introduction" to Wittgenstein and the Problem
of Other Minds, (H. Morick, ed.), McGraw Hill, 1967
Quine, W. vO, "Two Dogmas of Empiricism," From a Logical Point
of View, 1961..
Scriven, Michael, "The Logic of Criteria," The Journal of
Philosophy, LVI, (1959), p. 857.
Shorter, J. A., "Other Minds," The Encyclopedia of Philosophy,
The MacMillan Co. and The Free Press, 1967, Vol 6. p. 6.
Wellman, Carl, "Our Criteria for Third Person Psychological
Sentences," The Journal of Philosophy, LVIII (1961) p. 281.
Wellman, Carl, "Wittgenstein's Conceotion of a Criterion,"
Philosophical Review, LXXI, (1962), p. 433. Reprinted in Morick
Wittgenstein, Ludwig, The Blue and Brown Books, Harper
Wittgenstein, Ludwig, Philosophical Investigations, The
MacMillan Co 1953