This action might not be possible to undo. Are you sure you want to continue?
, 1970), pp. 192-205 Published by: University of Illinois Press on behalf of North American Philosophical Publications Stable URL: http://www.jstor.org/stable/20009349 Accessed: 06/08/2009 02:19
Your use of the JSTOR archive indicates your acceptance of JSTOR's Terms and Conditions of Use, available at http://www.jstor.org/page/info/about/policies/terms.jsp. JSTOR's Terms and Conditions of Use provides, in part, that unless you have obtained prior permission, you may not download an entire issue of a journal or multiple copies of articles, and you may use content in the JSTOR archive only for your personal, non-commercial use. Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained at http://www.jstor.org/action/showPublisher?publisherCode=illinois. Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or printed page of such transmission. JSTOR is a not-for-profit organization founded in 1995 to build trusted digital archives for scholarship. We work with the scholarly community to preserve their work and the materials they rely upon, and to build a common research platform that promotes the discovery and use of these resources. For more information about JSTOR, please contact email@example.com.
University of Illinois Press is collaborating with JSTOR to digitize, preserve and extend access to American Philosophical Quarterly.
Philosophical Quarterly 7, Number 3, July 1970
PRIVILEGED II. WITTGENSTEIN, AND INCOMMUNICABILITY
central insights have thus, I believe, Wittgenstein's been obscured. Accordingly, this article will consist
this paper, IN
for the following
(A) None of the arguments about the possibility of a private language or about the privacy of sensations and thoughts which Wittgenstein in the Philosophical Investigations advances provide good reason for doubting and "pain" (a) that words like "toothache" are the names (in a nontrivial sense) of sensations which people sometimes ex?
By reviewing make possible
the literature on the topic, I hope a fresh look at some ofWittgenstein's
a tooth? (b) that when I assert truly "I have ache" or "I am in pain," I am describ? or ing the state of my consciousness, I assert of another person that when (c) "He has a toothache" or "He is in pain" the same I claim that he is experiencing sort of sensation that I do when I have a toothache or am in pain (B) None of these arguments give good reasons as senseless the claim that for rejecting
"sensations are private"
of clauses (a), (b), and (c) in (A) The wording above are taken from George Pitcher,1 who argues that these three clauses express the view which to have over? believed his arguments Wittgenstein thrown. (Pitcher does not believe that these argu? ments do overthrow this view, but many ofWittgen? stein's readers have thought they did.) John Cook,2 in contesting Pitcher's account of what Wittgenstein thought he had shown, argues that Wittgenstein's
aim was rather to show the senselessness of the
in (B) and (C). In Sect. (II) I assertions quoted which for (A) by rebutting arguments argue toWittgenstein. In Sect. (Ill) I Pitcher attributes argue for (B) and (C) by rebutting those attributed to settle the to him by Cook. (I shall not attempt ad? of whether Wittgenstein actually question vanced either set of arguments.)
II. Pitcher's Interpretation
(C) None of these arguments give good reasons for rejecting as senseless the claim that "I know that I am in pain because I feel it." I have drafted these theses with an eye to recent the about views of Wittgenstein's discussions privacy of sensations, and in the belief that certain or his committed confusions by Wittgenstein in the between "privacy" interpreters?notably sense of "susceptibility to privileged access" and in led sense of "incommunicability"?have the unneces? to attribute commentators sympathetic views to him, and hostile critics sarily paradoxical these paradoxes. to attack him by attacking
1 The Philosophy Cliffs, New of Wittgenstein (Englewood 2 on Privacy," Review, Vol. "Wittgenstein Philosophical ed. by George A Collection of Critical Essays, Investigations: 3 p. 285. Pitcher,
Pitcher calls the view expressed by the three clauses cited under (A) above "View V" and in particular, claims that they?and, (b)?entail
that "I can never know for certain whether . . .
is in pain or not?for
pain."3 He regards
reductio, but does not argue for the entailment. is the conclusion the question of whether Bating seem that the entailment would it would absurd,
only hold if
I know entailed I feel Jones's pain.
is in pain
Jersey, 1964), p. 285. in Wittgenstein: Reprinted 74 (1965), pp. 281-314. Pitcher 1966), pp. 286-323. (New York,
But one would one held
this latter entailment
or (P3) Things are private to a given person either propositions about authority
"There has or is an . . ." or "There quality" cannot is an be a certain
( i ) If Jones is certain that he is in pain, and if I am certain that he is in pain, we must both
have the same grounds for certainty
if his like
. . .which overridden.
to this view. and there seems no plausibility Pitcher claims that V entails that "I Again, cannot conceive that another person feels the same I feel a pain."4 He that I do when sensation this by saying that supports
There I could sensation are in fact no I do determine specifiable that another that conditions under feels to be the which same to able person have I would
(P4) Things are private to a given person if he he can either not has something which share with another person, or which he can share only partially (in the sense that he can have the same kind of thing as another identical person, but not the numerically thing). Now View V does not commit us to (2) even if (2) is interpreted in the weak sense of (P4). (When so interpreted, we should notice, (3) does not follow.) As it stands, View V says nothing about epistemo logical questions, and thus does not commit us to the privacy of sensations in any other sense. How? ever, Pitcher is, reasonably enough, assuming that one who holds V will also hold I know about my pains by feeling them and else can feel my pains, nor can Nobody
: to do
feel his pain (see PI, sect. 253) ; and that is impossible. as "sharp," to me He describe his pain can, of course, and so on, but this is no help what? "dull," "severe," to of telling what ever; for I have no way corresponds these being in this adjectives used analogically. the I do" truth could case. are here adjectives of is no way there "He feels the of the assertion The . . . Since possibly be determined, the
how specifying same sensation assertion
In this passage, and elsewhere, that view V entails (2) "Pain"
Pitcher of a kind
is assuming of private
is the name
and that this in turn entails (3) I only learn what which in turn entails (4) I can only apply
is from my own case tomy own we
it is not
of the meaning
to recall some and it will therefore be helpful drawn most distinctions, by Ayer, explicitly of this term.6 Ayer between possible meanings notes that we might say (Pi) Things are private to a given person if their
existence, or one or more of their qualities,
These propositions would commit the holder of V to interpreting (2) in the sense of P2 as well as that of P4. But even this would not allow the inference to (3). For we might well say that I can learn what in all sorts of pain is (and/or what "pain" means) and could learn this even if I complicated ways, to have had a pain. Even if we never happened that we could not learn what pain was granted received unless we had linguistic appropriate in pain?thus with being training concurrently granting (3) though not the entailment of (3) by (2) ?one would still not be in a position to go from (3) to (4) without the help of some such premiss as (5) If I learn the meaning of a word by ostention a (i.e., if ostensive definition is necessary part of the of the process of learning the meaning then I can meaningfully apply that word), to me in the word only to objects ostensible the original cases (the same way in which ones used in training me) were ostensible.
could be known by him but not conceivably by anybody else, or (P2) Things are private to a given person if there is at least one way in which he can detect either their existence, or one or more of their qualities, but others cannot.
* Ibid., p. 288. 5 Ibid., p. 288. * Cf. The Concept of a Person (New York, 1963), p. 79. In stating these possible meanings, rather than simply of" an object, them by speaking of "the existence of, or some quality
statement I have touched up Ayer's of" the object. of "the existence
in View V seems to support such a But nothing should one prima facie premiss. Why implausible who holds V not answer Wittgenstein by saying that the criteria for the truth of "He feels the same and sensation I do" are the standard behavioral environmental ones? Pitcher implicitly grants the possibility of this reply when he says that "The fore? [various arguments, including going considerations the one now under discussion] do not show that
cpain' is not the name of a private sensation. They
is not a that particular (a mistake which of the language?of, result of ignorance of terms like "pain," e.g., the meanings
"sensation," to "stabbing," etc.)
of pain show only that some outward manifestations are required for the teaching and learning of the use of the word epain'."7 Pitcher says the admission of of this latter point constitutes a "modification" V, but it is not clear that it is. He so describes it, I believe, because he implicitly assumes that one who to upholds V is also committed
name of things of a certain If *N' is the (common) save some then least kind, kinds) (for at nothing utterances awareness of of such things, plus correlated which know the language contains *N* by those who to learn what 'N' means. *N', are required
But, defenders of V may reply, why does not the case of mental particulars simply show that the inference from (I) to (II) is fallacious? Why should be cases of corrigibility ? all cases of identification When we have to choose between an analysis of a concept like "identification of a particular" (e.g., one which would license the inference from (I) to wisdom such as (II)) and a piece of conventional View V, what is our criterion ? It is not too much to have say that certain interpreters of Wittgenstein as follows: "We can make mistakes when argued we identify, but not when we express feelings. There
are reports no mistaken are expressions pain-reports. of Therefore feeling, rather pain than
of particulars." So put, the argu? identifications ment is to say the least of it, an obvious petitio. To a does have Here, however, argument good against. get a good argument we need to argue that no Iwant simply to note that it is logically independent a particular" which does analysis of "identifying of V. First, however, for the sake of completeness, not permit the inference from (I) to (II) will be to let us look at a final argument which, according adequate. But I do not think that any interpreter of the "un? advances Pitcher, Wittgenstein against has offered such an argument. Rather, Wittgenstein form of V. modified" This assumption is one which I think Wittgenstein
This argument rests on the premiss that "The they have assumed that any adequate analysis of
expression of doubt has no place in the language like "I am game" (which we play with utterances
in pain saying tion appears now").8 if V that Pitcher were before presents Wittgenstein a correct "when then a person's mind, he as sensa? must
a particular" will "identifying of "observing the particular
"noting whether the particular
include (or its
the notions effects),"
identify that item (as, say, a pain rather than itch or an ache or a twinge). In that case, the possibility in his arises at once that he might make a mistake identification. He might always misidentify it, and hence it must always be possible for him to wonder he has done so or not."9 But the quoted whether premiss shows that he cannot so wonder. Pitcher is infers from here suggesting that Wittgenstein (I) I can identify to (II) It is in principle
criteria," Cartesian of host
etc. Since they note, rightly, that the tradition in philosophy has built up a out of the perplexities philosophical
that we possess an inner eye which
a certain particular possible for me
inspects (and, so to speak an "inner mind" which forms judgments about) mental they particulars, assume to overthrow that the fastest way this the notion that when we tradition is to abandon report, e.g., pains or thoughts we are identifying It is indeed a fast way, but it is cer? particulars. tainly not the only way. More will be heard of the fact that we cannot
have doubts about whether we are in pain when we
however, arguments to
Pitcher's the account
and practices, or failure
barring to master
Pt. I, sect. 288.
of V. On
p. 292. Pitcher, 8 Philosophical Wittgenstein, 9 p. 290. Pitcher,
the names of public objects and qualities denote their objects cannot be even remotely like the ways in which and this is ?pain' denotes a sensation," seen by noting that I can do practically none of the things [with pains] I
with or colors or shapes, physical objects so the modes and observable publicly things, in which behavior alone the connection between do with and the thing it names something public are not available in the case of "pain."10 of can i.e., of the is
Since "pain," under this definition, obviously is an observation commit term, the truth of (6) would " those who hold that 'pain' is the name of private to something sensation" like the "private diary"
picture of how we learn the names of mental
defender of V is then envisaged by Pitcher as replying that the process of learning by ostensive definition which we undergo while coming to use "tree" or "red" correctly may, in fact, be paralleled the one which the by a similar process?namely, celebrated diarist" conducts when he "private decides to call a given sensation 6E9. taken Having this line, he is now a patsy for the familiar objection that "the concept of correctness and incorrectness
does not apply" to 'E9.11
ask: is this appeal to a need merely one who private diary the only rejoinder which holds V can make to the argument which Pitcher to Wittgenstein? The attributes in difficulty this question is a result of the vagueness answering Here
of the phrase "the way in which names denote. ..."
But, given Wittgenstein's particulars. critique of the whole notion of "ostensive definition," and the doubt this critique casts on (6), cannot we simulta? neously repudiate (6) and maintain V ?Once again we need to ask whether we cannot simply appro? theories about language and priate Wittgenstein's the learning of language without thereby giving up V. As in the case of choosing an analysis of "identi? there may be a perfectly good fying a particular," of "denoting" which covers both "tree" analysis and "pain." At least we may say that unless it is shown that such an analysis cannot be given, and that any adequate analysis will entail (6), the does not show that argument "private diary" "pain" may not be the name of a private sensation, where "private" has the senses of (P2), (P3), or In my examination so far, it of the arguments seem that I have perversely evaded the may point. It seems clear that, despite the gaps in the argu? ments which I have tried to point out, there is
something to these arguments, and that our treat?
How much like the way in which "tree" denotes trees does the relation between "pain" and pains have to be? What parameters of similarity and dif? ference are appropriate here ?Does the fact that we can do "practically none of the things" with trees and "tree" that we do with pains and "pain" count
Wittgenstein showing " is out of place here? May of 'object and name' it not be that the model can be kept, as long as we are same on our model not guard is used to assume for "tree" that, and because for "pain," the as that what calls "the model
ment so far has not got at it. I think this is so, and I think that many threads can be knotted together on the central premiss of the next by focusing (and inter? last) argument which Pitcher presents?his of Wittgenstein's "beetle-in-the-box" pretation argument. This premiss is :
are private, that sensations Everyone acknowledges that no one can experience another sensations, person's so that the felt quality of each person's sensa? special tions is known to him alone and to no other.12
that holds for the former holds for the everything latter? By putting the claim that "we can keep a private diary" in the mouth of the holder of V, one
seems to assume that
Pitcher does in this sentence is to move from privacy in the sense of (P4) to privacy in the sense of (Pi)?from (7) Sensations cannot be shared to cannot communicate (8) We (the "special felt qualities")
to know what
certain qualities of our sensations
people make without having gone through a conscious process of inference) except by
ostention. 10 Ibid., pp. 293-294. 11 Ibid., p. 297. 12 Ibid., p. 297.
in pain, of your I do not know, cannot for
you are character
example, it is exactly like what Imight feel ifmy hand
were thing wounded altogether as yours different. is now, or whether it is some?
But, restricting ourselves toV and its consequences, than is no more convincing this line of argument the sceptical suggestion that you and I, when we inspect a patch of red or a beetle, are having wildly or seeing wildly different different experiences things, even though we say all the same things. If we both say the same things about our sensations, the patch, or the beetle, then it is not clear what it of means to say that there are certain qualities
sensations, patches, or beetles which are "incom?
I do not know whether it would be appro? to apply the name of a given sort of priate
sensation?e.g., "pain"?unless I know
whether I am talking about something which has certain incommunicable "special
to squash scepticism We may wish municable." our knowledge of beetles by noting that a about is not reflected in a possible dif? difference which
ference in what we say is, in some sense, not a real
that the plausibility of Wittgenstein's as they have been presented by Pitcher, arguments, arise from his tacitly treating (8), (9), and (10) as an intrinsic part of view V. This treatment is by no means disingenuous, for it is the case that philoso? who held V had frequently held (8), (9), and phers important both (10) as well. But it is nevertheless to distinguish between what have philosophers tacitly accepted and what common sense would say, I think
and least to recognize a few steps that farther these away various from theses common are at sense
But one may happily grant this point difference. and yet insist that the fact that pains cannot be to show that the relation shared does nothing between "pain" and pains differs from the relation "red" and red patches or "beetle" and between about If one refuses to take scepticism beetles. that the ground patches or beetles seriously?on that itmight be, so to speak, the sceptic's suggestion sheer chance that we all agree in what we say about one should not sense?then them does not make take scepticism about pains seriously either, and thus one should not grant (8). Nor does it help to say that it is only the false view of pains as private sensations which licenses the scepticism of (8). For one who holds view V may be committed only to in the sense of (P4) (and sensations as private and (P2) as well), not to their privacy possibly (P3) in the sense of (Pi), and thus not to (8).
Given this point, we can account for our sense
than the three assertions which express Pitcher's "official" version of V. I want now to finish the job of showing that, in
Pitcher's account of Wittgenstein's arguments,
everything turns on the truth of the additional premisses (8), (9), and (10). So I now come back to the argument which Pitcher constructs on the basis of his premiss about "special felt qualities." Pitcher passage quotes the celebrated "beetle-in-the-box" and says
If "pain" is is perfectly clear. with pain analogy a a somewhat to denote (including nothing) supposed can observe in his own which each person case, only if the and sole "cancels then the somewhat out"; is to denote is of the word function it, the word "pain" The at once deprived of any use.13
that there is something to the various arguments which Pitcher attributes toWittgenstein by noting more if that if (8) is accepted?or, generally, in the sense of (P2), (P3), or (P4) is taken privacy a in the sense of (Pi)?then to entail privacy which we have of dubitable number premisses a isolated above?e.g., (1), (3), (5), and (6)?look deal more plausible than they do in isolation. good we adopt two more They look still more plausible if
are various objections which might be made clause "in? about the parenthetical here?notably to and about the attribution cluding a nothing" holders of V of the claim that this is the "sole function" of "pain." But let us bate these issues and is being said when we say consider simply what 'cancels out'." The original that "the somewhat after describing people's passage in Wittgenstein, access to the contents of their boxes, says privileged
But suppose the word "beede" so, it would had not a use be used in these as the language??If people's in in the box has no place name of a thing. The thing at all; not even as a something', for the language-game one can "divide even be empty.?No, the box might through" whatever the by it is. thing in the box; it cancels out,
a given sort of sensation is only (9) I know what incom? certain about I know because are municable "special felt qualities" which of certain sensations I myself characteristic 1_
18 Ibid., p. 298.
WITTGENSTEIN, That expression designation" irrelevant.14 is to of say, if we sensation the object
construe on out
of the grammar of "object and as
says that Wittgenstein
denying that a particular the word "pain" the person that
thesis names about or language, this
so "neutron" in anything and cannot denote like the to directly that words which refer way experienced entities does tracks) denote. What (e.g., cloud-chamber a part are pointer in neutron games play language in mathematical results calcula? unexpected readings, ... tions.
is feels, something even like the way that words for publicly remotely name or designate observable them. In the things we play with like "tree" words and language-games trees and "red," and it is in these the name and the numerous "pain," cannot redness games (red things) play that the connection some part, between
designates in a way which
"tree," a part does play ... and behavior short, "pain" the external is used.15
the thing named is established. in But we play with the word language-games no part, sensations and so "pain" private play denote them in anything like the way that for example, denotes that kind of object. What in pain language-games behavior in which is pain ...?in the word
I take these parodies to show ifwe use the notion of to explicate playing no part in the language-game too easy to cancel "cancels out," it will be much things out.16 The trouble is that it is never very hard to describe the process of learning and using a seems to be which, prima facie, given term?one a language which used to denote particulars?in makes no reference to these particulars. All that one has to do is to use only terms which occur in criteria for applying the term in question. One will thus be able to argue that the particulars putatively denoted in the language-game "play no part" played with the putatively denoting term, and thus
Pitcher thus identifies "is canceled out" with no part in the language game." Two points "plays may be made about this passage. First, in saying that "private sensations play no part" Wittgenstein a petitio or else using "private is either committing sensations" as an abbreviation for "those special felt qualities of private sensations which are not in language." In the latter case, he communicable is perfectly justified in his claim, but then we must realize that it is not self-evident that there are such qualities, and that even if there were it might be that private sensations had plenty of other qualities which were communicable. consider the Secondly, following
I have developed the second point to emphasize the importance of the first, and to show that, on Pitcher's the "beetle-in-the-box" interpretation, reduces to the argument that what is argument incommunicable can play no part in a language-game. This argument may be expressed as follows : (a) Suppose that (?is an incommunicable quality of a particular?a quality which cannot be in language characterized Let G* be the (indefinitely (b) large) set of rules which govern the correct application of a ' 6P (putatively) referring expression iPi has (c) Suppose that we cannot tell whether been unless we know correctly applied it is being applied to something that whether
the language-games tree-in-itself plays in anything denote which no like we play so and part, the way that entities (e.g.,
in the numerous
"tree," cannot "tree" which
green sense-data) tree language-games mental entities desires, etc.)?in of awareness
to directly experienced denote. does What is sense-data, (intentions, short, which we the
a part in play and certain other volitions, apprehended consciousness "tree." we play no part,
judgments, directly are before our
(d) Then, by (c) G* must contain some rule of the form (R) "Call it a P only if you have good reason to think that it is (?" But then G* is inexpressible in language, for, (e) in any by (a), no expression language characterizes is therefore (?, and (R) inexpressible. (f ) But the notion of there being rules for using a language which cannot be themselves in any language is incoherent. expressed
objects in situations But with
to use which
in the numerous "neutron,"
language-games the neutrons themselves
14 Investigations, Pt. I, sect. 293. Philosophical 15 Pitcher, pp. 298-299. ie The fact that such as Strawson, can be constructed to such writers and Chihara has suggested parodies is taking for granted the Absolute the sort of operationalism which Idealists, Wittgenstein inspired Berkeley, I think that he did indeed do this, confusing the kernel of truth in the verificationist theory phenomenalists. with a stronger, false, theory about meaning.
and Fodor, tha and positivistic of meaningfulness
(a) and (g) Therefore patible hypotheses. If one grants
is sound.17 But we now see
(c) are mutually
(f ), then I think that this argument
that, the argument?in
(10)?viz., (c)?presupposes cable "special felt" qualities
are of the essence of those
that the incommuni? ofmy private sensations
sensations. The reason
"I cannot determine that another person feels the same sensation I do'."19 Cook argues that Wittgen? stein did not commit himself to (8)?on a construal of (8) according to which the "special felt qualities" are objects of knowledge?nor to (9) and (10). But his argument depends on attributing certain other dubious theses toWittgenstein, notably the follow?
ing: (according to Pitcher) thinks that why Wittgenstein There is no criterion of numerical private sensations cancel out but that trees and (as opposed to neutrons do not is that the former, but not the latter, identity of sensations generic) If I am right are "beyond the reach of language." The notion that "sensations are private objects" are of (8)-(io) that the denials in suggesting is senseless then central to Wittgenstein's thought, absolutely It is senseless to say that "I know that I am in should have it is easy to see why Wittgenstein pain because I feel it" to denying so much that the attached importance relation between pains and "pain" can be construed Since we have presupposed the contradictories of all If of 'object and designation'." "on the model three of these theses in our discussion of Pitcher's were true, then we would have to grant we need now to inquire whether (8)-(io) interpretation, the existence of knowledge unmediated by language in favor of Wittgenstein provides good arguments ?the central Cartesian fallacy. For we would have any of them. to grant that words in a public language could be Although Cook's criticism of Pitcher comes at the learned and used only by calling upon knowledge to it will be convenient very end of his discussion, usual of (as yet) incommunicable qualities?the with an analysis of this criticism to connect begin of language-learning. Cartesian-Lockean picture the remarks I have already made about Pitcher's This exegesis of the final argument which Pitcher version ofWittgenstein with the criticisms I wish to to Wittgenstein attributes may be confirmed by make of Cook's. Cook says that Pitcher thinks that : is criticizing a common-sensical view noting one more remark which he makes Wittgenstein whereas he is in fact only criticizing a "philos? in pain," to someone he "I am (V), say you If, after
he can comforts with you, does what you, sympathizes has done its work? to help you, then the word "pain" of what the nature to tell him and it was not used you be that cannot because had before your consciousness, told.18 notes not that Pitcher does Cook picture," opher's the last sentence of Wittgenstein's "beetle-in quote con? if we "That is to say: the-box" viz., passage, on strue of the expression of sensation the grammar
the model here is unexceptionable,
it was not used . . .")
as "if" but here
the object drops out
He continues it is not Wittgen? construes the on the model of
consideration The stein's grammar "object stein, as paradoxical "pain," irrelevant. opposite ing that since tions for word view and
except the claim that "that cannot be nothing told" and the vague suggestion that if an utterance the does one job, it cannot do two. If we discount latter suggestion (as I think we should) then we are left with the claim that what is before my conscious? ness "cannot be told." This claim rests entirely on
is crucial, for he opposes of sensation therefore who that the
of the expression and name," Pitcher thinks, consequence example,
it is not Wittgen? to the is committed use of the word
in the sensation
(8) and (10).
III. Cook's Interpretation
as out drops The of the passage, is quite the then, point Pitcher than show? of what Rather supposes. cannot sensations have it shows that names, that "no sensations place in are private the language to give any use of sensation allows sensa? and game" account of the words, we
the view to have makes (that
In the article referred to earlier, Cook criticizes to the view that Pitcher for attributing Wittgenstein
171 think that (f ) is true, and that Wittgenstein nor do I. 18 Pitcher, p. 299. 19 Cook, p. 313 (322). The number given 2 above). (see footnote anthology thought so too. But
thereby actual I do not
it impossible is, the "public")
think he had any convincing
in its favor,
is the page
if we reject
to give an account the view that sensations
alike?" This kind of identification question has no to
in the grammar of "tantrum," and so neither do place did have its two answers: the same one" "Yes, they the same one, two and "No, they did not have only same alike." Now this for the holds exactly point . .That of "toothache".. not grammar is, it would sense to say, as if in answer make to that question, have had the same toothache" either "They the same toothache."22 or "They did not
private, they be can't
to say that
then, seem to agree
premiss objects can't have
I wish to argue that Wittgenstein has given good reasons for holding this view. In examining
we are once again led to ask for an explanation
"private," but Cook is no help here. He thinks that to criticize Wittgenstein it is pointless for not having explained more clearly what he meant by "private language" because
the ably idea under not only suggested, the philosophical idea of Moreover, clearly explained. a private not merely it is confused in that language a mistaken notion of language supposes (or meaning) but in its very notion of the privacy of sensations.21 confused be investigation and hence turns can out to be irremedi?
can agree that such questions We "should not be asked," but we can also hold that the reason they are not asked is not that we lack a criterion for numerical identity of toothaches, but simply that it is obvious that the criterion for numerical difference the following definition : is satisfied. Consider the same if (D) Two sensations are generically and only if the persons who have them them in exactly describe the same terms, answer questions about them in exactly the same way, etc., and two generically identical sensations are numerically if and different only if they are had by different persons or by the same person at different times.
as it is? is wrong with (D)?simple-minded What Cook here seems to be saying that the very notion if one wants a criterion for the numerical of a "private object" is too confused to be explained. identity of toothaches? Cook is right in suggesting that Ayer In the light of Ayer's vigorous efforts to formulate would not accept it, since Ayer wishes to leave open of various senses of this notion, this is explanations the possibility of "co-consciousness," and thus does a strong claim. Cook backs it up by claiming that not want its impossibility built into the language a false presupposition that sensations can be only we just identical permits Ayer to formulate his game. But suppose that, for the moment, numerically to describe decide all cases of possible "co Thus, according to Cook, when Ayer explanations. cases consciousness" (e.g., cases of split personality, says that in the sense in which two people can have
the same pain or the same thought we do not have of interchanged or inter-communicating organs,
"numerical identity" in mind, he is presupposing the view that sometimes we do have numerical identity in mind, and this is false. Cook thinks that
we must answer use of "no" a familiar sensation to the question words with "is there, a criterion then of
etc.) in a way compatible with (D). Do we not then have all we need to give sense to the notion of the same, but numerically different, "generically
Cook would presumably would insist that
say that we do not and
passage Thus, :
answer comes out best in the following
has described one remark like of her children's
if a mother someone a tantrum here
(I) If a yes-or-no question does not normally occur in extra-philosophical discourse (al? all the words used in it do), then the though and all direct answers to it, are question,
tantrums, had had like" with That may
else might "exactly
in contrast, not so forth. "rather and like," different," not be asked: "Do you suppose is, it would they have had the same one and not just two exactly is used "rather
that her child . . . that" "Exactly with but "same,"
regards this principle as superior to the one usually adopted by philosophers who claim that "It is a necessary truth that no two people can have
the same toothache"?viz., a principle such as
20 Ibid., p. 312 (322-323). 21 Ibid., p. 281 (286-287). **Ibid., p. 308(312).
the answer regard to numerical identity. Further, "No, the builds are different, but they're exactly alike" is "senseless," if any grammatical English sentence is. But this example just forces us to ask: or like like mountains (a) "Are pains more since (D) would, mutatis mutandis, builds?"; (b) seem to work for builds as well as for sensations, is a necessary not our so-called criterion for numerical truth. expresses asked) identity of sensations susceptible to a reductio ad absurdum ? One may sympathize with Cook's doubts about The answer to (a), I think, is that pains are more of "necessary truth" (or (II), since the notion like mountains simply in the respect that we do truth" or "grammatical is "conceptual truth") can have sometimes say things like "But nobody indeed very murky. But I think the attempt (which this pain" and "But the one in Antarctica can't be to substitute theme of Cook's article) is a major this mountain" and don9t say things like "But from mixing notions of "confusion up arising nobody can have this build." (If it were said that and of "senselessness" distinct language-games" for but a philosopher would use the first of the nobody is worse is a cure which than the such notions quoted sentences, I would not wish to argue a point disease. One can applaud Cook's remark that "This which seems tome irrelevant. It would be relevant talk of grammar 'forbidding us' to say something to only if philosophers played on the limited?and, is nothing but the most recent jargon for calling a between pains my mind, quite innocent?analogy to halt to an analogy whose oddness has begun and mountains created by their odd linguistic dawn on one,"23 while regretting that Cook himself habits in order to infer to some further analogy.) does not stick to blocking bad analogies, rather than as far as I can There is no particular reason, see, resorting to charges of "senselessness" and invoking why we treat pains but not builds as particulars, is wrong with the (I), consider (I). To see what but then why shouldn't we? The answer to (b) is argument against the claim that there is a following that we could treat builds as particulars, in which : criterion of numerical identity for mountains case we doubtless would use a modification of (D) two mountains?one that in It might be discovered as a criterion of numerical for builds. It identity in Antarctica?were like each and one Alaska exactly just happens that we don't. same configuration, the had precisely the other. They In making these points, I have no wish to deny made seemed rock of which they were qualitatively Cook's point that builds and pains are analogous
a radius of, so on for all features within identical?and from their respective say, two air miles peaks. "Exactly not with in contrast, is used here but like" "same," so forth. "rather and "rather with different," like," That may tion not be asked "Do is, it would This be the same mountain?" you kind suppose they of identifica? of "moun? "Yes, they in Cook a way says that that neither the "my" are analogous in "my coat" with coats. is a "posses?
(II) If a yes-or-no question does not occur in discourse because it is extra-philosophical never the case that the answer to it is "yes" it is never the case that the (or because answer to it is "no"), then a suitably form of the answer which would generalized be given (if the question were invariably
in the grammar has no place question so neither do its two answers: and tain," are the same" and "No, they are not."
sive of ownership" because the question of whether it's my coat (whether I own it) is not settled by the fact that I've got it. In this (stipulated) sense, the "my's" in "my build" and "my toothache" are, to be sure, not possessives of ownership. But Cook goes from
In cases where the "my" same question''24 is not as his a possessive of ownership, a "genuine X?" is not
This argument does not show anything except that have different if we know that two mountains we don't ask whether locations, they are the same It certainly does not show that "The mountain. cannot be in two different places at same mountain the same time" is senseless. So sometimes (I) is false. On the other hand, Cook has an example which makes (I) look fairly plausible. This is the noun "build," as in "His build is exactly like his father's." Here indeed the question "Do they have the same in different ways? cannot be construed build?" once in regard to qualitative identity and again in
23 24 Ibid., Cf., p. 303 (311). ibid., p. 296 (303).
X "is my identification to
"Is my X the same as his X?" answers to it, direct
to criticize this inference. If a I do want identification question" means "one that "genuine would not be asked by anybody who knows the language" then we still have to ask "Do we not ask it because the answer is so obvious, or because we and
sense of it?" In the case of builds,
"because we can't make sense of
sense of it although I can imagine someone making in which "my by creating a new language-game is treated as the name of a particular, build" rather than as the name of a collection of qualities. In the case of pains, I am inclined to say "because the answer is so obvious." All that Cook has done is to note that there would be no great difficulty in treating "my toothache" as the name of a sharable collection of qualities. He has not shown that we do so treat it or that we should so treat it. To show the former he would have to establish some sort of invidious distinction between language-games and language played (mainly) by philosophers which games played by non-philosophers?one enables us to discount the former as ipso facto "mis? taken" or "confused." To show the latter, he would to show that the philosophers' have habit of treating "my toothache" as the name of a particular has had disastrous philosophical and consequences, that the only (or the best) way of avoiding these consequences was to break the habit. I think that the latter project is closest to Cook's " actual intention, for he says that when 'no two can feel the same toothache' comes to be people called a necessary truth," then "one easily con? cannot that we cludes know about anything
another person's toothaches."25 But this is not an
are private" in saying that "sensations is wrong senseless, and thus all I can to argue for the claim labeled (B) in the first paragraph of this article. I turn now to the one labeled (C), and ask whether Cook has produced good arguments for rejecting "I know that I am in pain because I feel it" as senseless. If he can do this, he will have shown that to interpret "privacy" in the sense of (P2) iswrong headed. Cook sets about making this latter claim in an odd way. He says that Wittgenstein wants to get around "Argument A" which goes as follows: (i) No one can feel (experience, be acquainted with)
another sensations. person's means and necessary of coming proper (ii) The know what sensation another is having person to feel that person's sensation. (iii) Anyone because sense Conclusion person who he has feels a sensation it, and whatever to exist can know in any knows can that he has be known to is it to
exist by being felt cannot be known (in the same
of "known") : No one other way. another what sensations
One would think that (ii), and the second half of (iii), were so implausible that an attack on them would suffice to get around Argument A. But Cook employs an indirect method, arguing that (ii) and that (iii) presuppose
use of the verb is a genuine (iv) There an of with expression certainty sensation statements.27 present-tense "to know" first-person as
It does not follow at all unless one easy conclusion. grants some further premisses?e.g., (i), or (5) or are of which, I have argued (8)?all above, dubious. A philosopher who refuses to grant any of these further premisses, while holding out for the view that "sensations are private" and explicating "private" in the sense of (P2), (P3), or (P4), is not thereby going to get involved in any of the standard about other minds, of the puzzles knowledge external world, and the like. If he interprets "private" in the sense of (Pi), and thus assents to (8), he will indeed be in trouble. But, once again, we need to notice the distinction between "privi? leged access" ("privacy" in the combined senses of (P2) and (P3)), "unsharability" ("privacy" in the sense of (P4)), and "incommunicability" (privacy in the sense of (Pi)) and to ask whether, if these are borne in mind, we cannot get the distinctions of Wittgenstein's treatment benefits of private particulars without accepting Cook's paradoxes. I have now done all I can to show why Cook is
25 Ibid., p. 309(318). 26 Cf., ibid., pp. 283-284 27 Ibid., p. 285 (240). (289-290).
and that this presupposition is false. The odd thing about this argument is that neither nor "certain" in (ii) or in appears "certainty" and thus it is hard to see how either could (iii), presuppose (iv). Suppose we forget about "cer? for a moment and ask simply: can we find tainty"
an account of the meaning of "I know that..."
to which "I know that I am in pain" according would be sensible? Let us try a conventional analysis, according to which "I know that/>" is true ifp, I believe that p, and I have adequate evidence for my belief that p. Do I have good reasons for that I am in pain, when I believe that I believing am? Two answers suggest themselves: (A) My and (B) My reason is that I know the language, and reason is that I feel it
know which consequently which "anxieties," "pains,"
states are called which "thrills,"
Since Cook has a rather detailed discussion of (A), to which we shall shortly turn, let us concentrate to say first on (B). At first sight, (B) seems merely "I know that I am in pain because I know what seem satisfactory. But pain is," and this does not for now let us read (B) as an abbreviation am inclined to say "I am (B') My reason is that I in pain," and I know that this inclination, had by one who knows the language, is itself and indeed conclusive evidence, evidence, for the truth of what is said. the person who is asked to justify his belief Here that he is in pain is calling attention to a convention that the utterance convention of the language?the to utter, certain first-person of, or the disposition reports are taken as the best possible present-tense cannot be evidence (and, indeed, evidence which for their own truth. In other words, he overridden) between the fact that he is in pain is distinguishing and his own belief that he is in pain, and saying that the latter is evidence of the former. One could rule this answer out if one held that Beliefs can never be evidence proposition believed for the truth of the
which sounds more funny (a result which would, as the recent literature shows, lead to an irresolvable impasse), we have to look further and ask which of the two views leads to the greater quantity of funny sounding views. In other words, we have to look to of taking the further philosophical consequences
one alternative rather than the other. Wittgen
to steinians, by and large, think that a willingness to say CIknow what I am think? "It is wrong say " is the price we have to pay for not having to ing' the bulk of the funny things which epistemolo say tradition have been wont to gists in the Cartesian as I remarked the familiar above, say. But, Cartesian puzzles would only follow if one accepted, of (ii) and (iii) that the presupposition not merely itmakes sense to talk about knowing that one has a but (ii) and (iii) themselves. certain sensation, Once again, we are led to ask why Cook does not attack the premisses rather than their presupposi?
but this is false, since other people certainly take my states as evidence for their beliefs about my mental states. So why own beliefs about my mental shouldn't I ? Can we justify holding I can never take the fact that I believe ? adequate evidence for the truth of p p as
A partial answer to this question will emerge if of go back to the phrase "as an expression ab Cook drags which in, seemingly certainty" extra, in formulating (iv). The reason why he does it makes so is suggested by his saying "Whereas doubt sense to speak of ignorance and knowledge, and certainty, in the case of the stone in the shoe, it does not make sense to speak this way in the case of the man in pain."28 The crucial move here is the implicit suggestion that
sense to say "I doubt it does not make (v) Where sense to say "I am certain p" it does not make p," and conversely. that that
it does it does the not make not make of sense sense to say "I am certain to say "I know that intended in (ii) and
(vi) Where that/?" p" (in
not see how. It seems to Perhaps we can, but I do me that we have a choice between letting "privi? as expressions of knowledge by leged" reports count count, or else ruling letting this sort of evidence I do them out by ruling out this sort of evidence. tomake this choice except by looking not know how of the views which to the degree of paradoxicalness If we choose the will result from each alternative. the fact that former we have to say "Sometimes beliefs are held can be adequate evidence for the truth of those beliefs," which sounds a little funny. If we choose the latter we have to say with Wittgen? stein "It iswrong to say CIknow what I am think? a little funny. Unless we ing'," which also sounds to fall back on our intuitions about are simply
**Ibid., p. 288(294).
(iii),) and conversely. Since one ismuch more ready to say "I cannot have I am in pain" than to say doubts about whether " to "It is wrong to say CIknow that I am in pain' and (vi) is to have persuasive reasons for accept (v) did not saying the latter. If Cook and Wittgenstein that move to help them, it is doubtful have this would have been quite so certain that those they who believe (iv) are "confused." But this move is at Pitcher which the same as the move bottom from attributes toWittgenstein I can identify to a certain particular
It is in principle possible for me tomake amistake in identifying that particular (a mistake which is not a result of ignorance of the language), a move which
to accept an
in which that is a circumstance I can be of the truth value of statements made by ignorant then
means of the sentence. Since there is no reason to
is asking us
which is not so-called "incorrigible knowledge" to count as knowledge. In both cases, the argument is clearly "reversible" (in Waismann's phrase). That is, it is not clear whether we should accept em? the analysis and get rid of the (potentially) barrassing fact that we possess incorrigible knowl? we should reject the analysis edge, or whether because it does not cover a certain sort of knowledge (while trying to overcome our embarrassment by other tactics). Prima facie, my certainty that I am in pain is an obvious counter-example to (v). If I
am told that I cannot render a satisfactory account
for Cook's purposes. What he obviously intends is that we should take "the speaker" (in both the sentences quoted above) as short for "someone who knows the language." If we make this emendation, see that Cook has simply ruled out then we "knowing the language" as a factor to be taken into account in determining whether someone is in "as good a position as one could want for correctly
answering a certain question or making a certain
of the meaning of "certain" that will not entail (v), and of "know" that will not entail (vi), then we are driven back to the question about whether the fact of believing p may not sometimes be adequate evidence for p. For if itmay be, then I am entitled to reject the conjunction of (v) and (vi) as false. The importance of (v) and (vi) for Cook's view will be evident from the following passage. In dis? cussing the reasons one might adduce in support of "I know it's raining"?e.g., "I'm looking out the
window"?Cook What for makes someone says it possible : to use "I know" here is not, as one as an
If he did not rule it out, then we could simply reply to him and Wittgenstein: "Certainly I need not suppose that Jones is in as good a posi? tion as one could want for determining whether he is in pain ; after all, the poor boy is only three, and statement."
he still uses ways." 'pain,' 'hurts,' etc., in some very peculiar
Once again, we find ourselves brought back to the alternatives presented earlier :either admit that the language" counts as part of "being "knowing in a good position" (i.e., as a sufficient belief and accept first-person present justifying reason)
tense pain reports as cases of knowledge, or rule
expression of certainty is that it would be intelligible
particular want for making know to suppose that the speaker in as good a position a certain correctly answering instance, statement. More in the could or "I
the language" out and refuse to allow "knowing these cases to count as knowledge. The primary reason why Cook prefers the latter alternative is that he thinks that making sensations "objects of knowledge" Specifically,
engenders philosophical it suggests the question:
mental states as to make
perplexities. What is so
a certain that..." that
least necessary blank allow circumstances of by means
to be an expression of certainty, it is at the sense of the sentence the filling to be the in some speaker ignorant of the the truth value of statements made sentence (or equivalents thereof).29
The first sentence of this passage makes clear that Cook tacitly adopts (v). His taking for granted the relevance of "as an expression of certainty" to (ii) and (iii) demonstrates his adoption of (vi). His refusal to consider the possibility that "knowledge of the language" can count as a good reason for that I am in pain can be seen by the fact knowing that his necessary condition for "I know that..." an expression as it stands, of certainty, being excludes nothing whatever. For if I do not know
what 29 some words in the sentence in question mean, Ibid., pp. 285-286 (291-292). 30Cf. The Blue and Brown Books
edge of them incorrigible ? Now if we answer this by saying "They are directly present to conscious? ness, and nothing else is," and add on a few other the resulting epistemological plausible premisses, dualism will beget metaphysical dualism (which will beget Idealism, which will beget Neutral which will beget Logical Positivism) as Monism, as Sin gave birth to Death. But suppose we surely answer the question by saying: "There is nothing that special about them, apart from a convention
first-person present-tense reports of them are taken
as the best possible evidence about them." If we that when we reach con? add, with Wittgenstein,30 ventions we reach rock-bottom, we will not be tempted to go on to ask "And why do we apply this convention to some things and not others ?" We will not fall under the spell of what Pitcher calls the
in degrees of that differences principle" or of corrigibility, to meta? correspond in the objects known. physical differences This last suggestion ismerely a sketch of a "con? ventionalist" view of non-inferential knowledge, one which I cannot argue here.31 It is inserted here to show that there may not be as much merely reason as Wittgenstein thought to be frightened of are private of the view of sensations objects to the business at hand, I knowledge. Returning shall conclude my discussion of Cook by making one further criticism of detail and then spelling out a more general criticism, which I have sketchily above. The criticism of detail concerns adumbrated sense of cto feel'." his discussion of "the perceptual of the defects of Cook's general method Here seem to me particularly obvious. He argument to argue, in the face of cases like "I feel a wants inmy knee," that "Sensation words cannot be pain in first-person the objects of verbs of perception sentences."32 Clearly, he needs to say that "feel" in So the sentence cited, is not "a verb of perception." " as the sense of 'to feel' he defines "the perceptual sense in which "feel" is used in the sentence "I know it because I feel it." He then says that in such sentences as "I feel a slight pain in my knee when I bend it" the words "I feel" may be replaced by "I have" or "there is" without altering the sense of the sentence. So far so good, and all we need certitude,
now is an argument to show that when "feel" is
between uses of a term to analogy or disanalogy If he had con? convict his opponents of confusion. out that pains, and the fined himself to pointing way we know about them, differ from stones and the way we know about them, and that something in Argument A, or some other argument with a undesirable conclusion, depended philosophically that we knew about them in upon the assumption just the same way, then all would be well. In certain
passages?passages which, to my mind, are the best
for does just this. Thus, parts of his article?he of A he notes that "the plausibility example,
depends on its seeming to be analogous to some?
thing like this : to ascertain whether my neighbor's crocuses are in bloom, as opposed to merely taking his word for it, I must see his crocuses.33 Later he as follow s : points up the relevant disanalogy It [Argument A] makes
first-and circumstance third-person
out the difference
on a matter
to see my unable neighbor's (like being us realize has made whereas Wittgenstein crocuses) in the language-game itself. resides that the difference rest on some not The difference does circumstance, and such feel therefore another's Argument a circumstance with sensations" A, which the words is inherendy purports to name to unable "being confused.84
I think that in the first sentence of this passage Cook an important characterizes part of accurately
Wittgenstein's of the notion contribution of "privacy." In to our the understanding second sentence,
is sense such substitution in the perceptual impossible. But we do not get this. All we get is the sense fact that in some cases of the use of perceptual
"feel"?e.g., "I feel a stone in my shoe"?such
he confuses the fact that we may be however, to accept the false propositions inclined (ii) and on the basis of a false analogy between knowing (iii)
about-pains and knowing-about-crocuses, or "confused." with the
is impossible. This does nothing what? substitution ever to show that satisfaction of the initial definition of the non of the criterion satisfaction entails is" for "I feel." One of "There substitutability disanalogy does not make a difference of sense, and even if it did, itwould not necessarily make the very special difference of sense which Cook needs for his argument. The fact that when we feel stones in our there for checking whether shoes we have methods is one, but that when we feel pains in our knees we to do not, hardly shows that it is a "confusion" answer the question "How do you know that there I feel it." is a pain in your knee?" by "Because as in his comparing Here, "pain" with "build"
31 It is a view which at length has been elaborated vol. 4, pp. 204-212. by Sellars,
that (ii) and
(iii) are not simply false, but are
There is a
to accept a difference between people being willing senseless quasi-proposition) false proposition p, (or another because (or proposition they accept or senseless quasi-proposition) q, and p entailing are the phenomena q. Sometimes presupposing but it has been the burden of my co-extensive, that, in the case at hand, they are not. argument between first- and The claim that the difference is built about sensations statements third-person can be accepted by some? into the language-game one who accepts all the premisses of Argument A,
and which I try to summarize in "Intuition," The Encyclopedia of
1966), Philosophy (New York, 32 Cook, p. 289 (295-296). 38 Ibid., p. 290 (296). **Ibid., p. 291 (297).
Both, in other words, focus on what Strawson calls to privacy," rather than Wittgenstein's "hostility since the implausibility of (ii) and (iii) is such as to on what to im? Strawson calls his "hostility us dismiss the argument at once. Granted let that it to break the spell which took Wittgenstein these mediacy."35 In the discussion so far, I have admitted that it is we not to recite need continue exerted, premisses to construct analyses of "know," and of possible the powerful incantations (such as "It is incorrect or "name" "denote," "refer," (or "identify"), to say 'I know what I am thinking' ") which or which will entail the two theses just mentioned, used for this purpose. (It took Hume's Wittgenstein to construct distinctions between alternatively views to wake 18th-century philosophy paradoxical different "senses" of these terms which will produce from its dogmatic but we are fortunate slumbers, the desired result. But I have argued that Cook and sense of gratitude that no misguided to led Kant Pitcher present no good reasons for saying that we reiterate Hume's slogans). should adopt such analyses or make such distinc? tions. I have argued further that both men pick up IV. Conclusions the problem by the wrong end, that what needs or distinctions between is senses, analysis, My discussion of Pitcher and Cook has suggested suffice to do "privacy," and that Ayer's distinctions the following view of Wittgenstein. Wittgenstein this job. I have suggested that it was Wittgenstein's to cut Cartesian scepticism off at the roots, wanted failure to make the distinctions which Ayer makes, and hoped to do so by arguing that the Cartesian and to recognize that an object can be "private" in and of incorrigible picture both of non-inferential one of these senses without being "private" in all In particular, he noted knowledge was incoherent. the others, which accounts to for his "hostility that the Cartesian tradition assimilated the privacy." terms like language-game played with sensations In another article, forthcoming, I hope to show to that played with physical-object terms "pain" that the proper thrust of those portions of the like "tree," thus produced and various false has grouped together Investigations which Malcolm to analogies. These analogies inclined philosophers as "arguments of a private against the possibility make various false statements (or to invent a new, is against the notion that we can have language" and then, specifically philosophical language-game, knowledge of something distinct from the knowledge drawn from this artificial by combining premisses are true of it, and against that certain propositions with from more drawn language-games premisses the notion that we can have knowledge of the truth "natural" to produce arguments language-games, are not formulated of propositions which in which committed fallacies of ambiguity). Pitcher to argue In other I want that words, language. the differences between in the way emphasizes what is novel and exciting in these portions of the which we name trees and pains, and Cook the dif? is the attack on the Cartesian notion ferences between the way in which we know about Investigations of pre-linguistic awareness?the notion that there trees and pains, but both agree that is a species of awareness which antedates and If sensations are private objects, they can't have our coming underlies to be able to justify the names in which (in the same sense of "name" utterance of sentences. But this is a separate topic, trees have names) which I have only had space to hint at in the and article. present
ness of any relevant statements. Nor need he try,
criticisms For, if the foregoing cannot get from this difference
are sound, Cook to the "senseless?
If sensations are private objects, we can't know about them (in the same sense of "know" in which we know about trees).
Cf. P. F. Strawson,
This action might not be possible to undo. Are you sure you want to continue?
We've moved you to where you read on your other device.
Get the full title to continue listening from where you left off, or restart the preview.