§1. The Transcendental Characteristic of the Approach

«What has to be accepted, the given, is... forms of life» (226eFoot note 6_2). The term «form of life» signifies the modes or structure of systematically understood social behavior, of human activity.Foot note 6_3 Form of life, according to Wittgenstein, is the ultimate reality upon which all the possible modes of intentionality should be grounded. Language-game, in Wittgenstein's use of the term, is «the whole, consisting of language and the actions into which it is woven» (7). «The term `language-game' is meant to bring into prominence the fact that the speaking of the language is part of an activity, or of a form of life» (23).

It should be emphasized from the outset that the language-game approach is of transcendental character. Wittgenstein never intends to turn linguistic analyses into an empirical study like pragmatics, although the latter itself is a worthwhile topic for philosophers. He is never interested in, e.g., what is speaker's intention in uttering a word/sentence. What is of primary philosophical significance for the language-game approach, is the transcendentality that underlies variegated uses of linguistic signs. Here the term «transcendentality» does not signify any a priori principle or any innate linguistic capacityFoot note 6_4. What should be reckoned as transcendental, the only thing that is given and has to be accepted, is the logical structure of human social activity, i.e. form of life. There are numerous ways in which one word/sentence can be employed and numerous speakers' meanings or intentions that can be attached onto a word/sentence. What is interesting for the language-game approach is how uses of words are embedded in form of lifeFoot note 6_5, and it is a problem concerning the relation between a usage and the logical structure of human activity, not the usage itself, which is of empirical nature. I call the former inquiry «transcendental» in the sense that it concerns, given a form of life, the preconditions for possible applications of linguistic signs and possible operations of consciousnessFoot note 6_6.

What signifies the transcendental characteristic of the approach, in Wittgenstein's own terminology, is «depth grammar». At an early date of his later period (in June 1931) he entitled one of his book «Philosophical Grammar»Foot note 6_7. According to him, the language-game approach «is directed not towards [linguistic] phenomena, but ... towards the `possibility' of phenomena ... Our investigation is therefore a grammatical one» (90). «These are, of course, not empirical problems; they are solved, rather, by looking into the workings of our language» (109), note, not by probing into the workings of speaker/hearers, not by digging into past, present or future intentions of utterers. This transcendental approach is concerned with the depth grammar (as opposed to surface grammar, i.e. the way words are used in the construction of sentences) (664). Nevertheless, it is not a Kantian or Tractarian one; it is not «a final analysis of our forms of language» (91).Foot note 6_8 The depth grammar, i.e., what is transcendental for individual employment of words, is itself contingent upon particular form of lifeFoot note 6_9.

§2. The Problem of Individuation

Let us now follow Quine and say, «there is no entity without identity.» Sensation words, like any other kinds of words, come into language through communal praxis and they do not have any private reference. If there are any objects of words that can be private, then the question of identity criteria inevitably arises. Wittgenstein addressed this question in 253.

However Wittgenstein's emphasis on the correctness of identification in 258 overshadowed the 253 argument. This emphasis is misplaced.

Let us imagine the following case. I want to keep a diary about the recurrence of a certain sensation. To this end I associate it with the sign `S' and write this sign in a calendar for every day on which I have the sensation. -- I will remark first of all that a definition of the sign [`S'] cannot be formulated. -- But still I can give myself a kind of ostensive definition. -- How? Can I point to the sensation? Not in the ordinary sense. But I speak, or write the sign down, and at the same time I concentrate my attention on the sensation -- and so, as it were, point to it inwardly. -- But what is this ceremony for? For that is all it seems to be! A definition surely serves to establish the meaning of a sign. -- Well, that is done precisely by the concentrating of my attention; for in this way I impress on myself the connection between the sign and the sensation. -- But «I impress on myself» can only mean this process brings it about that I remember the connection right in the future. But in the present case I have no criterion of correctness. One would like to say whatever is going to seem right to me is right. And that only means that here we can't talk about «right». (258)

Here for the correctness argument we have, (A) «I remember the connection right in the future», (B) «There is no criterion of correctness», (C) «Whatever is going to seem right to me is right», and (D) «We can't talk about `right'».

Suppose I name my sensation S_t1 (occurs at time t₁) «S», and I then apply the name «S» to my sensation S_t2 (occurs at time t₂), then to S_t3, S_t4...S_tn (occur at t₃, t₄...t_n), all of which are similar to each other. In the process of applying «S» to S_t2, S_t3, S_t4...S_tn, is there any connection between «S» and S_t1 that I have to remember rightly? Wittgenstein would say it is the naming relation between «S» and S_t1. At any rate (A) means that I memorize S_t1 and recall it in the future accurately.

In order for understanding (B), we have to know what this criterion of correctness could be. In public language the conventional rules, including identification criteria and paradigms, are criteria of correctness. In private language, if I apply «S» to S_t2, S_t3, S_t4... S_tn, there is no publicly accessible paradigm by which we can check each of these applications is made right or wrong, in terms of how similar S_tn is with S_t1. (B) and (C) virtually assert the same thing and they indicate that 258 assumes that S_t1 is analogously a private paradigm. The fact that we cannot apply this private paradigm or private criterion of correctness to later occurrences of sensation amounts to that we simply don't have it as a paradigm.

The correctness argument fails on two accounts. Firstly, it relies on contingency of human memory capacity. Secondly and more importantly, it misrepresents the logical structure of recognition act.

If Robinson Crusoe (assuming he was left alone, having not yet learned to speak, nurtured by some animal) possesses a very extraordinary capacity in memorizing sensation, then he should be fully capable of inventing a private language. Whether or not other people have access to his private paradigm would make no difference: he doesn't need public criteria of correctness. One may argue that «it is not possible to obey a rule `privately'» (202). Yes, but in the current context this is the very thesis that one should establish and therefore it cannot serve as a premise.

For the second point, let us come back to one of Wittgenstein's 1930's texts, where he performed a meditation on recognition.

[1] How do I know that the color of this paper, which I call `white', is the same as the one I saw here yesterday? By recognizing it again; and recognizing it again is my only source of knowledge here. In that case, `That it is the same' means that I recognize it again.

[2] Then of course you also can't ask whether it really is the same and whether I might not perhaps be mistaken; (whether it is the same and doesn't just seem to be.)

[3] Of course, it would also be possible to say that the color is the same because chemical investigations do not disclose any change.

[4] Recognition is what is primary and identity what is secondary.Foot note 6_10

[4] indicates the nature of this discussion: it belongs to the transcendental discourse: seeing from the empirical angle, identity is primary and recognition is secondary. From the transcendental perspective, it is community's recognition that decides an object's identity criteria, which in turn, makes individual's (empirical) recognition possible.

[3] indicates that the community's agreement on choice of paradigm, which establishes identity criteria, is primary. For example, we can choose, or not choose, chemical investigation as the final decision procedure of identification.

[1] concerns situations in which a sample, an instance of paradigm, is not available, e.g. the white paper placed here yesterday has been destroyed, and we only have memory to rely upon. Or, a sample still exists but according to our memory the color has changed (56). Under these circumstances, as [2] makes explicit, it does not make sense, to talk about right and wrong or talk about whether or not two color-appearances are really the same. And they are not even needed: identification hangs on practical purposes. [2] is a straightforward denial of the correctness argument.

Whether or not a publicly accessible paradigm is available is largely irrelevant, and even if it is available, S_t2, S_t3, S_t4...S_tn still have to be identified through senses.Foot note 6_11 This undermines the correctness argument. Note that the correctness argument rests upon the assumption that there is no publicly accessible paradigm by which we can check each of applications of «S» is made right or wrong, in terms of how similar each S_tn is with S_t1. Here the problem is rather the range of similarity from S_t1 to S_tn and this is a problem of individuation. Whether S_t1, S_t2, S_t3, S_t4... S_tn are private or public makes no difference.

The (transcendental) question is, how a rule, or an identification criterion, is decided. Central to the private-language case is the problem concerning the social nature of individuation. We don't know how a private criterion of identity is generated.

«Another person can't have my pains.» -- Which are my pains? What counts as a criterion of identity here? ...I have seen a person in a discussion on this subject strike himself on the breast and say: «But surely another person can't have THIS pain!» -- The answer to this is that one does not define a criterion of identity by emphatic stressing of the word «this». Rather, what the emphasis does is to suggest the case in which we are conversant with such a criterion of identity, but have to be reminded of it. (253)

The decision on identification criteria is an essential part of a linguistic community's form of life. What counts as an individual, the whole body of the rabbit, or each part of it, or a temporal slice of it, depends upon our collective operational relation with the object, depends upon the object's status in our way of life. In the case of sensation, e.g., as long as doctors treat your pain and my pain in the same or even similar way, two pains should count as the same (kind). What counts a kind, a type, an identity criterion, is hinged upon our way of life.

Only «if I assume the abrogation of the normal language-game with the expression of a sensation, I need a criterion of identity for the sensation; and then the possibility of error also exists» (288). Otherwise «to use a word without a justification does not mean to use it without right» (289). Given the premise that the argument is of transcendental nature, section 288 provides an exact answer to Kripke's question (regardless of what kind of interpretation he is imposing on Wittgenstein):

How can I possibly have any difficulty identifying my own sensation? And if there were a difficulty, how could `public' criteria help me?

Surely I can identify these [sensations] after I have felt them, and any participation in a community is irrelevant!

It seems to me that we have sensations or sensation qualia that we can perfectly well identify but that have no `natural' external manifestations.Foot note 6_12

If I have no difficulty in identifying my sensation, it is due to the fact that at the empirical level sensation has already been individuated by our form of life! I agree with Kripke in that the view that an inner process always has outward criteria is «empirically false»,Foot note 6_13 e.g. my feeling of pain may well be more reliable as in indicator of my illness than a CT scanning result is. But public criteria belong to the transcendental makeup of sensation language. It is in the transcendental sense that Wittgenstein says «That is not agreement in opinions but in form of life» (241), and «What people accept as a justification -- is shown by how they think and live [i.e. what their form of life is like]» (325). Essential to the debate, as I shall show in the following section, is the difference in metaphysical position.

§3. The Metaphysical Status of Sensation

In 1959 Ayer claimed that having tried to construct a language all of whose words refer to private things he believed that in any language which allows reference to individuals there must be criteria of identity which make it possible for different speakers to refer to the same individual.Foot note 6_14 By Wittgenstein's standard this amounts to abandonment of the private language view, but Ayer didn't withdraw from the battle and in fact in his lifetime he never did.Foot note 6_15 In 1973 Ayer declared that in the empiricist construction program (1) the observer was not permitted to conceive of the data with which she works as private to her and (2) the observer was not identified either with her or with any other person.Foot note 6_16 If it is so, what is left with respect to privacy for what he termed as «a reformed Robinson Crusoe approach», i.e. a construction program developed by a single observer? Ayer says that this Robinson Crusoe approach is supposed «to do justice to the fact that any knowledge of the world which anyone acquires is bound to be based upon his own experiences [i.e. his own sensation and memory].»Foot note 6_17

Ayer's central thesis is this: since the ultimate ground for language use is the individual's judgment upon her own sensation and memory only, the distinction between public and private objects is unfounded in the first place. This is Ayer's final position in his lifelong campaign against Wittgenstein's private language argumentFoot note 6_18. The thesis is so pivotal to the metaphysical foundation of those empiricist programs that it merits a careful exploration.

The crucial fact which it seems to me that Wittgenstein persistently overlooks is that anyone's significant use of language must depend sooner or later on his performing what I call an act of primary recognition. In Wittgenstein's example [265], it is supposed not to be sufficient for someone to check his memory of the time at which the train is due to leave by visualizing a page of the timetable. He has to check the memory in its turn by actually looking up the page... But unless he can trust his eyesight at this point, unless he can recognize the figures printed in the table, he will be no better off. If he distrusts his eyesight, as well as his memory, he can consult other people, but then he must understand their responses. He needs to identify correctly the signs that they make. The point I am stressing is not the trivial one that the series of checks cannot continue indefinitely in practice, even if there is no limit to them in theory, but rather that unless it is brought to a close at some stage the whole series counts for nothing. Everything hangs in the air unless there is at least one item that is straightforwardly identified.

If this is correct, Wittgenstein is wrong in taking that the corroboration of one memory by another is an inferior substitute for some other method of verification. There is no other method. Whatever I have to identify, whether it is an object, an event, an image or a sign, I have only my memory and my current sensation to rely on. The difference lies only in the degree to which the memories are cross-checked.Foot note 6_19

In this primary recognition argument Ayer commits two mistakes and both of them are essential to the empiricist vision of language and knowledge. The first is that he assumes that memory is capable of working independently and internally without any external assistance and corroboration, and the second is that he writes as if all the multifarious employment of memory and senses enjoy equal credibility and significance.

In the example of a timetable, if my eyesight is not reliable I will naturally appeal to that of others. This decision presupposes my trust upon my ear and understanding. If both my eyes and ears are not dependable I should e.g. purchase auxiliary equipment, and that again presupposes the effectiveness of modern technology, the reliability of manufacturer, and so on. But in no time should I belabor the universal doubt. It is not that a series of checks has to stop somewhere by virtue of our special trust upon any particular employment of senses, it is simply that no such an obsessive scrutiny is needed to be conducted by the assumed judges of senses. Only when the primacy and sufficiency of sense/memory are assumed such a scan and a primary recognition should be called for.

First, experience is a symphony of coordinate actions rather than a solo of successive sensations. Success, failure, stableness, anomaly, smoothness and incoherence etc. are the most salient parts in the organic whole of our experience. Empiricists are tempted to say that these are nothing but groups of sensory indices, the objection is that they dominate the whole web of experience in which sense/memory inquests are only scattered and dependent episodes. Action is metaphysically primary to sensation, and sensation ultimately should be understood through action, not the other way around.

Second, certitude pertaining to multifarious uses of senses is varying and thus should not be assigned indiscriminately. In the example of the timetable, the credibility of memory is different in category, rather than in degree, with that of senses. If I am weak in senses, then the credibility of them should be different in category with that of the aid of sensing-equipment plus my neighbor or colleague's contribution plus my capability of understanding. So on and so forth, reliability can never be accredited equally to all the contemporaneously active sectors.

It is the sensualist empiricist's dogma that human experience only or ultimately consists of sensation and memory. The fact that our judgments upon sensation and memory are coordinated and corroborated with our every-minute fulfillment in action and the dynamic equilibrium of the whole system of experience is so familiar to us that these factors are easily overlooked. When Kripke declares that «Surely I can identify these [sensations] after I have felt them, and any participation in a community is irrelevant (!)»Foot note 6_20 doesn't he commit to this Ayerian weak form of private language view? It is indeed a difference between two metaphysical positions.

For sensualist empiricists social reality is nothing but a total sum of individual sense experiences. Ayer thus wrote in 1985:

The practice of the community is supposed to bestow meaning on my utterances. But what is the community except a collection of persons? And if each of those persons is supposed to take his orders about meaning solely from the others, it follows that none of them takes any orders. The whole semantic house of cards is based upon our taking in each other's washing, or would be if there were any laundries to wash. On this interpretation, Wittgenstein's argument, so far from proving that private languages are impossible, proves that they are indispensable.Foot note 6_21

Ayer here appeals to the empirical notion of community. Indeed, a university, a nation, etc., consists of persons. But this is irrelevant to our question because the notion of social praxis is of transcendental one and should not be reduced to that of individual's practice. The empirical causal chain of learning doesn't count because the meaning of word (1) is inculcated into an individual's mind through her everyday-and-lifelong participation in social life and (2) has to be cashed out in every-minute social transaction.Foot note 6_22

Sense data do contribute a substratum to the linguistic meaning, and how these data are processed and integrated into the connotation of words should be properly accounted for.Foot note 6_23 But the question here is that if one chooses sensory input alone as the metaphysical basis for meaning, then certain commitment to privacy seems to be inevitable, and then skepticism as well seems to be unavoidable.Foot note 6_24

• Ayer, Alfred Jules
  • 1954: Can there be a Private Language, in Wittgenstein, University of Notre Dame Press 1966.
  • 1959: Privacy, in The Concept of A Person, St. Martin's Press 1963.
  • 1973: The Central Questions of Philosophy, Penguin Books 1976.
  • 1982: Philosophy in the Twentieth Century, Unwin Paperbacks 1984.
  • 1985: Wittgenstein, Random House 1985.
  • 1988: A Defense of Empiricism, in A. J. Ayer Memorial Essays, the Press Syndicate of the University of Cambridge 1991.
• Kripke, Saul
  • 1982: Wittgenstein on Rules and Private Language, Harvard UP 1982.
• Wittgenstein, Ludwig
  • 1974: Philosophical Grammar, Basil Blackwell, 1974.
  • 1975: Philosophical Remark, Basil Blackwell 1975.