From the ALWS archives: A selection of papers from the International Wittgenstein Symposia in Kirchberg am Wechsel

• Home
• About
• Login
• Search
• Current
• Archives

Table of contents

• References

Wittgenstein’s so-called ‘private language argument,’ that is, those sections of his Philosophical Investigations following upon §242 and featuring, prominently, a critical discussion of the notion and possibility of a language involving reference to one’s immediate and private sensations, has seemed elusive, if not impenetrable, to many. Attempts to assess the validity of the argument have traditionally been hindered by vast differences of opinion regarding its overall content, differences supervening upon a myriad of smaller exegetical puzzles which have themselves, in turn, evaded satisfactory interpretive consensus.

As opposed to rehearsing these well-worn pathways of controversy, in what follows I shall attempt to give a new and simplified reading of Wittgenstein’s argument, one which I hope will undercut many of these concerns, and shed light on many contentious points, by bringing to the fore a key aspect of the considerations developed in § 242 ff that has, historically, been relatively underappreciated. More specifically, I shall present Wittgenstein as giving an argument in two basic stages, the second of which pertains to the possibility of a private sensation language, and which addresses concerns which emerge naturally out of the first stage, which pertains to identity statements. The first stage involves the development of a claim made initially in the Tractatus (c.f. 1961 5.5301-5.5302, p xviii), but to which Wittgenstein continued to adhere, though in modified form, in the Investigations, and which he introduces in various contexts throughout his later philosophy. (e.g., 1958, § 216, and 2005, p 304e) The claim, reiterated in this context at § 253, is simply that Leibniz’s principles of identity are flawed; identity statements, in so far as they are not nonsense, must state, not that a thing is identical to itself simpliciter, but rather, and contra Leibniz, either of two things that they are identical to one another (e.g., sortal identity), or of one thing that it is qualitatively discernable from itself in some respect (e.g., the morning star and the evening star are the same celestial body). This claim, I shall argue, is intended to provide support for a characterization of pains etc., as qualitatively discernable identicals, states of a living human body which ‘show up’ in qualitatively distinct ways depending, for instance, upon whether one is in the state in question, or merely observes it. One important merit of this characterization is that it renders much more plausible Wittgenstein’s various claims to the effect that it is possible for others to have direct (i.e., non-inferential) knowledge about, inter alia, our sensations and mental states.

The argument against a ‘private sensation language’ can then be seen, I shall argue, as invoking previously developed considerations regarding the concept of following a rule in an effort to address a potential response available, in light of the attack undertaken on the basis of the consideration of identity statements, to a defender of the ‘in principle’ epistemological privacy of sensations. The response is to attempt to identify the phenomena referred to by our sensation words with phenomenal items which, existing independently of the “natural expressions” (§ 256) or “outward signs” (§ 257) of sensation, it is impossible for others to know. The argument against a private sensation language is invoked to block this move by demonstrating, not that we cannot meaningfully refer to our own ‘in principle’ epistemologically private sensations,¹ nor that we can refer to such items but only insofar as our sensation words are employed within the context of a ‘public framework,’² but rather that, because we, and others, can in fact refer to our sensations, the phenomena they refer to cannot, as the interlocutor had hoped, be identified with phenomenal items of this sort.

Consonant with his general approach to identity statements, Wittgenstein insists, in § 253, that the criteria for identity in the case of ascriptions of sensation (or of sensation location), are such that, in so far as it makes sense to say that one pain is identical to another, e.g., “‘my’ pain is the same as (or in the same location as) ‘his’ pain,” it must involve reporting a contingent identity. It might, for example, intelligibly be thought of as asserting identity of sortal (and so be analogous to e.g., “this chair is not the one you saw here yesterday, but is exactly the same as it” (§ 253)), or, on the other hand, the self-identity of some particular object but as determined by two or more sets of qualitatively discernable descriptive or perceptual information (e.g., “The place where Fred has a pain is in the same as the place where Bob has a pain” (where Fred and Bob are Siamese twins) (ibid.,) By contrast, in the sense in which it seems impossible to deny or doubt that a particular pain was mine, it is also impossible to affirm or be certain about it. Propositions in connection with which it makes sense to think that the opposite of what they apparently assert might obtain, such as ‘this table has the same length as that one over there,’ (§ 251) play a fundamentally different role within language than those, such as ‘Every rod has length,’ for which this is not the case. (ibid.,) In particular, while the former are empirical propositions which are capable of being true or false, the later are grammatical propositions which simply state or are derived trivially from linguistic rules. (§ 253) Since attempts to deny the latter sorts of propositions are senseless, further, emphatically stressing the point that a particular pain is mine and so cannot be someone else’s, i.e., ‘Surely no one else can have THIS pain,’ are thus not empirical assertions, (§251) but rather amount, much as does the proposition ‘Only you can know whether you had that intention,’ (§ 247) to a potentially misleading explication or explanation of a grammatical rule, in this case, that first-personal ascriptions of sensation are first-personal. Much as in the case of ‘a thing is identical with itself,’ (§ 216) the seemingly legitimate employment of a criterion for identity in the case of ‘another person cannot have my (or ‘THIS’) pain,’ is, Wittgenstein argues, chimerical. This is because, though perhaps reflecting certain contingent empirical features of our nature and circumstances, ‘another person cannot have my (or ‘THIS’) pain,’ is no more an empirical proposition than is, for example, ‘One plays patience (Solitaire) by oneself.’ (§ 248) To deny the latter proposition, clearly, would not amount to repudiating a deep metaphysical necessity, but would rather simply demonstrate a lack of understanding vis-à-vis the contingent empirical fact as to what the word ‘patience’ (Solitaire) means. Similarly, in the sense in which it seems impossible to do so, the attempt to deny the proposition ‘another person cannot have my (or ‘THIS’) pain’ would not amount to denying an important philosophical truth, but would rather simply show a misunderstanding of the grammar which, as a contingent and empirical matter of fact, governs ascriptions of sensation formulated in the first-person. Taken, by contrast, in one of the two senses in which it can intelligibly be thought of as employed to make an empirical assertion which repudiates the possibility of a certain sort of identity relation, the claim that ‘another person cannot have my (or ‘THIS’) pain,’ is simply false, with the case of Siamese Twins, and that of corresponding injuries, bearing decisively against it.

But because, in so far as pains etc., are identical they must be qualitatively (or, derivatively ‘quantifiably’) discernable (i.e., any legitimate assertion of a relation of identity must be a contingent proposition such as one of those of the two sorts outlined above), it is possible, for example, for someone to know that another is in pain without needing to have the others pain (§ 246). Pain ‘shows up’ differently depending upon whether one has the pain or observes it. In particular, while the observer cannot ordinarily have the pain of the one whom he observes (unless, e.g., he has a corresponding physical injury), it is nonetheless perfectly possible for that observer to know another’s pain non-inferentially, on the basis, specifically, of the fact that the pain in question is often quite readily discernable in certain characteristic and overt ‘natural expressions’ as well as ‘outward signs’ (such as grimaces, groans, and exclamations). The sense in which one’s sensations (and perceptions) are ‘private’ is thus not that others cannot know them, but rather simply that others ordinarily do not have them. That sensations (or sensation locations), are typically ‘had’ by only one person and so ‘show up’ in an unique way to the person whose states or sensations they are, is, furthermore, indicative neither of a deep metaphysical division, nor a consequent ‘in principle’ epistemological privacy, but is rather simply a contingent, empirical fact about our bodies which in certain cases does not obtain (e.g., Siamese Twins (§ 253)). Because the usual, if contingent, circumstances are such that it is true, however, that one and only one person has a particular pain or is in a particular mental state, etc., the grammar of first-personal ascriptions of sensation and mental conduct are such that we will say things like ‘only you can know whether you had that sensation (or intention),’ where ‘knowing’ is not here, as elsewhere, something opposed to uncertainty, but merely indicates that the expression of uncertainty is not ordinarily a genuine move in the language-game of first-personal ascriptions of sensation or intention. (§§ 247-248)

Analogously, if one is to doubt a proposition in the sense in which doubt is genuinely opposed to knowledge, it must be possible to know it. There can thus be no logical ‘half-way house’ consisting of bodies which may or may not have consciousness or sensations (i.e., may or may not be ‘philosophical zombies’), but of which it is impossible to know whether they do (or are), and so no ‘problem of other minds’ in the sense that because the minds of others are epistemologically private and hidden, we cannot, even in principle, know whether or not those minds exist. If we can meaningfully make an assertion about something, e.g., that our own or some other body does or does not have consciousness, or is in pain, etc., that assertion must rather, Wittgenstein insists, employ words with epistemologically public criteria of correct application. (§ 258, 265) Indeed, the argument against a ‘private sensation language,’ taken up in § 256ff, makes precisely this point in an effort to subvert an attempt, undertaken by the interlocutor, to identify the various phenomena referred to by our sensation words with ‘in principle’ epistemologically private phenomenal items which are (purportedly) independent of the ‘natural expressions,’ or ‘outwards signs,’ of sensation (i.e., are the sorts of thing which a ‘philosophical zombie,’ by definition, is supposed to lack).

As in the case of the rule-following considerations introduced prior to § 242, the argument against a ‘private sensation language’ involves introducing a certain sort of scepticism concerning the movement from present (or past) to future employments of a particular sign, in this case, a putatively ‘private’ ostensive definition. Here, however, the scepticism is neither constitutive,³ nor epistemic, but rather heuristic. Wittgenstein’s point in § 258 is that in remembering, in the future, the connection which I have established in the past between sign and sensation, I will have no genuine criterion for the correctness of the memory, since, given the privacy of the definition, whatever seems to me to be correct will be in fact correct. While going on to concede, of course, that “the hypothesis that I make a mistake is mere show,” (§ 270) i.e., it is not genuinely possible to make a mistake in the case of (putatively) employing a sign the use and definition of which it is impossible for others to understand (this, in fact, is precisely the problem with such a sign), Wittgenstein brings this point out heuristically by noting that, given that memory is ordinarily fallible, it is intuitively problematic to rely exclusively, as would seem to be required in the case of a private ostensive definition, upon our personal and introspectively available memories in order to guarantee, or provide a criterion for, the ‘sameness’ of usage in each case. This simply reflects, however, the more fundamental point that there being a correct usage of a particular sign involves its being governed by standards conformity to which can readily beindependently determined or confirmed. The manner in which a sign with epistemologically public criteria for correct application evades the difficulties characteristic of a ‘private’ ostensive definition is, further, not that it facilitates communal agreement in regards to the use of the sign (understood in this way a ‘public’ definition will inevitably seem no more or less tenuous than a ‘private’ one),⁴ nor that it allows for the introduction of public evidence which facilitates or assists in the correct identification of our epistemologically private sensations,⁵ but rather that, because its definition can be understood by others, it is possible for there to be genuine disagreement about whether it has been used correctly; it is thus possible in such a case, and as it is not in the case of a language which it is impossible for others to understand, for someone to misapply a term referring to e.g. a sensation.

But if a private language of the sort envisioned is impossible, then claims to the effect that ‘I am in pain’ or ‘She is in pain,’ cannot, Wittgenstein insists, refer to a phenomenon which is knowable only to the person that has it. It is therefore not available to the defender of the ‘in principle’ epistemological privacy of sensations to identify the various phenomena referred to by our sensation words with phenomenal items which, according to her view, others cannot know. Simply put, if it is impossible for others to know it, then it cannot be what I or anyone else refers to by a sensation word. This is not to say, however, as has sometimes been supposed, that it is impossible to refer meaningfully to our own ‘in principle’ epistemologically private, and inner, sensations. Indeed, given the validity of the private language argument, so conceived, this statement about its import would be meaningless (i.e., it would amount to the self-defeating assertion that ‘I have pains (or sensations) to which I cannot refer.’) Nor is it the point to show that we can refer to our ‘in principle’ epistemologically private sensations provided the words employed to do so are situated within a ‘public framework.’ Understood in this way, the argument against a private sensation language would render mysterious Wittgenstein’s claims to the effect that sensations etc., are non-inferentially knowable. The point of the argument is rather to show that our sensations, perceptions, and so on, and in fact because we can meaningfully refer to them, are simply not epistemologically private ‘in principle’.⁶ In so far as they are ‘inner,’ furthermore, it is not in the sense of their being encapsulated in an epistemologically private, and metaphysically distinctive, ‘inner’ space, but rather in the sense of their being ‘in a living human body,’ the mental and other states of which are qualitatively discernable identicals and so non-inferentially knowable from a third person point of view. Sensations, perceptions, consciousness, unconsciousness, and so on, are thus states of a living human body, that is, a person, and though, as a contingent matter of fact, they ‘show up’ differently to the person who is in those states than to a person who observes them (much as a table, as a contingent matter of fact, looks different from different angles), they are nevertheless perfectly discernable in that person’s behavior (and, derivatively, their physiology) within certain circumstances.

Refbacks

• There are currently no refbacks.