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

• Literature

A basic assumption of traditional philosophy of mind, which Wittgenstein attacks, is that we extrapolate the meaning of third-person ascriptions of mental-states from the meaning of first-person ascriptions (hereafter, the Extrapolation Thesis). In this paper I examine two fundamental strands in Wittgenstein's attack. My main conclusion is that none of them succeeds. But, as I indicate at the end, this may not be a reason for traditionalists to rejoice.

Following Kripke's lead (1982, 114-133) the first strand can be reconstructed as follows. Obviously, in order to extend talk about mental-states from 'ourselves' to 'others' we must abstract from particular features of these states that make them 'ours'. But what could these features be? The traditional answer is 'soul', 'mind' or 'self' qua an immaterial entity, which mediates between our mental-states and bodies: On this view, the key to the idea of mental-states 'had' by things other than oneself like one's neighbour or one's chair, is the idea of mental-states that resemble one's own in all respects except for 'belonging' to a different 'mind', which is somehow attached to the body of one's neighbour or to one's chair. However, a fundamental problem of this account is that the 'having' relation between physical objects and 'minds', which it assumes, is unintelligible. This is particularly clear in the case of inanimate objects: To paraphrase Wittgenstein 1953, §361, what, if not spatial location, which 'minds' lack, could relate a given 'mind' to one chair say, rather than another, or to the chair as a whole, rather than to its back or legs, or the air around it? But the 'having' relation between animate objects and 'minds' is also not spared. For on the account at issue it is exactly the same relation as that between inanimate objects and 'minds'.

It follows that a substance-dualism, which assumes some indirect relation between bodies and mental-states via the mediation of 'minds', cannot ground the dualistic idea of mental-states ownership required by the Extrapolation Thesis. Perhaps then, this idea may be grounded in a property-dualism that assumes some direct connection between physical objects and mental-states. Maybe, to be more specific, the idea of mental-states that are not one's own, rests on the idea of mental-states that resemble one's own in all respects except for belonging to a different body cum physical object?

However, in so far as mental-states, like a rash of anger, that cannot be said to have any specific bodily location are concerned, this proposal is in no better position than its precursor. For the same difficulties that beset the 'having' relation between physical objects and 'minds' also inflict the 'having' relation between physical objects and mental-states of the non-locatable sort. Indeed, the problems indicated above concerning the former relation are pointed out in §361 with respect to the latter relation.

When it comes to mental-states like pain, which can be said to have a bodily location, these problems no longer arise. But then other serious problems emerge. For, as Wittgenstein insists, the possibility that one has a sensation like pain in the body of someone else is perfectly intelligible (1953, §302; 1969, 49-51). So the idea of a pain, or any other locatable mental-state had by someone else, cannot be based on the idea of location in another body.

It follows that property-dualism also cannot ground the dualistic idea of mental-states ownership required by the Extrapolation Thesis. So this idea is vacuous, which means, the argument concludes, that the Extrapolation Thesis cannot be true.

As this argument claims, the 'having' relation between bodies and 'minds' may indeed be unintelligible. But, contrary to what it argues, the analogous relation between bodies and mental-states makes perfectly good sense. Consider for example our visual perceptions. A striking feature of these experiences is that they present the perceived objects as standing in various spatial relations to the perceiver (Evans 1982, 153-4). And this, it should be noted, implies that although visual experiences cannot be related to a specific body via a bodily location, they can be related to it via their "egocentric" spatial content: Such an experience presents a specific location as the point of origin of the spatial relations which it involves, thereby reflecting the location of the perceiver's body. Moreover, visual experiences can be related to specific bodies by the ways they affect their behaviour. Indeed, the content of these experiences is intimately linked with behaviour, since the egocentric terms - viz. above, below, right, left, in front, behind - in which it can be specified, derive their meaning in part from their complicated connections with bodily actions (Evans 1982, 154-7). Finally, visual experiences can be related to specific bodies by their relation to experiences that have a bodily location. For example, a visual perception of an object may be related to a perception of this object by touch, which is bound up with a sense of parts of the body where the tactile perception is felt. And it may also be related to a proprioceptive awareness of the body, which is a type of experience that takes a back seat in consciousness almost all of the time (Bermúdez et al. 1995, 12-15, 175). Indeed, owing to visual experiences' strong ties with behaviour and bodily-experiences, they actually carry with them a sense of the bodies whose location they reflect (Cassam 1997, 52-3).

In like manner, a non-locatable mental-state like a rash of anger, which has no "egocentric" spatial content, can be related to a specific body by (1) its relation with a locatable mental-state - e.g. a sharp pain in the knee; (2) its relation with a non-locatable mental-state with a spatial content - e.g. an auditory experience of the man who hit the knee running away; (3) its behavioural impact - e.g. a shout at that man; (4) its relation with proprioceptive awareness of e.g., pangs in the chest. Thus, our mental-states, locatable and non-locatable alike, "latch" so to speak, onto our bodies in complex ways. This means that the non-locatablity of a great many of them is no bar against body-based accounts of mental-states ownership. But it also indicates that recent accounts of our notion of the self, which give the body a prominent place (Bermúdez et al. 1995), are very likely on the right track.

Moreover, the variety of ways in which mental-states can be related to a body enable body-based accounts of mental-states ownership to accommodate cases like Wittgenstein's example of a pain felt by one person in the body of another. For while this pain is related to the one body by way of its location it may still be related to the other body via affective, emotive and motivational states which "latch" onto it. And, if significant enough, the latter rather than the former can be considered as what matters for ownership of the pain. Indeed, owing to the aforementioned variety, the range of possible deviant relations between mental-states and bodies is much wider than Wittgenstein's example may suggest. Think, for example, of the possibility of a body related to visual experiences centered on another body, which is related to auditory experiences whose point of origin is a third body etc. (Strawson 1959, 90-2). Admittedly, if the deviant relation is too complicated, body-based talk about ownership by a particular subject would lose sense. However, proponents of the body-based approach may bite the bullet here, and not implausibly accept the consequence that when mental-states become too dispersed among bodies the notion of a single owner of them becomes vacuous (Evans 1982, 250-3).

These comments certainly require elaboration. But they suffice to show that, at least as it stands, the first strand in Wittgenstein's attack on the Extrapolation Thesis is unsound.

However, this attack has a second strand, which runs as follows. Extending our concepts of mental-states to others on the basis of our own case is tantamount to extrapolating from talk about our own mental-states to talk about the mental-states of others. But the former talk is a private communication of ourselves with ourselves, which is rendered impossible by the private language argument. So given that we can meaningfully ascribe mental-states at all, the Extrapolation Thesis must be false.

A fundamental assumption of this argument is that the private communication it deals with (hereafter, Privatelogue), must be a private language in Wittgenstein's sense. When introducing the latter notion, however, Wittgenstein provides two definitions (1953, §243). On a standard reading of the first, what makes a language private is that the meaning of its terms is constituted by a relation they have with epistemically-private mental-items - i.e. mental-items whose nature can be known only by the speaker of the language (Blackburn 1984, 92-3). On the second definition, which he considers as closely related to the first, what makes a language private is that no one apart from its speaker can understand its terms. As I will now show, the second strand is either unsound or viciously circular, depending on which of these notions of privacy it involves.

The only reason of some weight, it seems, to consider Privatelogue as private in Wittgenstein's first sense is the following. Privatelogue's terms are first and foremost terms for mental features. So if terms for outer, supposedly behavioural features contribute to their meaning, then, arguably, there must be some necessary connection between these behavioural features and the said mental ones. However, the Extrapolation Thesis goes hand in hand with a strict denial of any such necessary connection. So, from its viewpoint behavioural cum non-private features cannot contribute to the meaning of Privatelogue's terms.

The traditional denial, however, of a necessary connection between the mental and the behavioural essentially boils down to two theses. First, that the core of any mental-state consists of an epistemically-private inner item. Second, that the possibility that the same private-items involve very different behavioural patterns, or even none at all is perfectly intelligible. Obviously, as far as these theses are concerned our private-items and behaviour can be systematically correlated. But if such a correlation obtains, we can certainly speak about our mental life by way of terms whose meaning is determined both by our private-items and by our behaviour. True, anything falling under these terms would have of necessity features of both sorts. Further, and connectedly, the mental concepts they would reflect would have a behavioural dimension. Finally, we could not use them if the correlation between our private-items and our behaviour were very different. However, these implications are perfectly in keeping with the aforementioned traditional theses. Thus, proponents of the Extrapolation Thesis need not be committed to the view that the meaning of Privatelogue's terms must be constituted by a connection they have with private-items. And this means that the second strand in Wittgenstein's attack on this thesis is unsound, if its fundamental assumption involves his first notion of privacy.

However, even if the meaning of Privatelogue's terms need not be completely determined by private-items, it certainly must be partly determined by such items, which is the same as saying that this meaning can be fully known only by one subject. And isn't this privacy with respect to knowledge of meaning tantamount to Wittgenstein's second sense of privacy, or privacy with respect to understanding of meaning? A basic belief we all naturally share is that others have, in broadly similar external circumstances, broadly similar inner-states. But given this Similarity Assumption, we will be able to arrive at true beliefs about the meaning of others' expressions, which involve private-items (Craig 1997, 131). And isn't this enough to constitute understanding? True, since there is no question of knowing whether the Similarity Assumption obtains, we will not be able to know in this case that we understand. However, this does not seem necessary for understanding (Craig 1997, 131). So, unless it is shown that the Similarity Assumption cannot be true, the assumption that Privatelogue must be private with respect to understanding is unfounded. But how can this be shown if not by demonstrating that the Extrapolation Thesis must be false, thereby rendering meaningless (1) the idea of interpersonal comparison of private-items, which goes hand in hand with this thesis, as well as (2) the Similarity Assumption, which involves the latter idea. It follows that the second strand in Wittgenstein's attack on the Extrapolation Thesis is required for establishing the assumption that Privatelogue must be private with respect to understanding. And this means that this move is viciously circular, if its fundamental assumption involves Wittgenstein's second notion of privacy. Thus, the second strand is either unsound or viciously circular.

Both strands, then, in Wittgenstein's attack on the Extrapolation Thesis are unsuccessful. However, mental-states owe their identity as particulars to the identity of those to whom they belong (Strawson 1959, 97; Evans 1982, 253). So by the rebuttal of the first strand, which is committed to a body-based account of mental-states ownership, the identity conditions of such states must be bound up with bodily cum behavioural features. Similarly, one part of the rebuttal of the second strand brings in a behavioural element into our mental concepts, and so also may the other part. So both rebuttals may well be committed to the view that the meaning of third-person ascriptions of mental-states and their behavioural justification conditions are not completely independent. This is somewhat ironical since the Extrapolation Thesis goes hand in hand with a traditional assumption to the contrary, and it seems that Wittgenstein's attack on the former is intended as part of an attack on the latter (1953, §353). Thus, both rebuttals may turn out to be a Pyrrhic victory. But a fuller examination of these points will have to wait for another occasion.^*

Refbacks

• There are currently no refbacks.