1. What is the Problem?

§243 of Philosophical Investigations talks of a language such that ‘the individual words of this language are to refer to what can only be known to the person speaking; to his immediate private sensations. So another person cannot understand the language’.

Why does Wittgenstein assume that the fact that a word refers to something that can only be known by the speaker means that no one else can understand what is said? Quite contrary to this, we have no difficulty at all in accepting that the taste-sensation or scent-sensation that someone else is experiencing is private to them, but that it is the sort of experience we might ourselves have in similar circumstances: we could taste what that person is tasting, or smell what that person is smelling. A person might say ‘try some of this, it's lovely’, and we try it and indeed it is lovely. We have no difficulty in understanding what has been said, though we also understand that what has been referred to is something essentially private to that subject. An inference to the best explanation assures us that we are both experiencing the same sort of sensation, though each sensation is private to the subject. What is the problem?

The puzzle is only deepened by Wittgenstein's suggestion in this paragraph that all the words in the putative language refer to what can only be known by the speaker. For this is not the claim which is the main target of his criticism. That target is the view that our talk about sensations involves reference to what is essentially private to the speaker. Wittgenstein is perfectly happy to assume that such speakers, the ‘private linguists’, can intelligibly refer to objects and events in the public world: boxes (§293), manometers (§270), the usual causes and effects of pain (§271), for example. There is no attempt to argue that such references are themselves to what is essentially private to the speaker. This must put paid to the suggestion that Wittgenstein's real target is phenomenalism, the doctrine that, since physical objects are logical constructs out of sense-data, it must be true that all the words in the putative language refer to what can only be known by the speaker. But if the private linguist can refer to the public causes and effects of pain ‘as we all do’, as Wittgenstein says (§271), then it cannot be the case that another person cannot understand the language.

§258 prompts another question. The last part of it runs:

But this passage simply gives rise to the same question about the status of criteria: if I have no criterion of correctness, how do I distinguish between what seems to be a criterion of correctness and what actually is such a criterion? What should be clear is that a vicious infinite regress threatens. This is the essence of Ayer's criticism in his paper ‘Can there be a Private Language’.Footnote ¹ Any check, or criterion, we might refer to must be grasped correctly by the subject, and if we are not to allow an infinite regress we have to acknowledge that ultimately our grasp of the import of any such check on our use of a term, any term in our language, rests on a not further checkable private grasp or understanding of what that check means. Thus, if I ask someone else whether I have understood the import of the check correctly I have to understand what he or she is saying; this process must come to a stop, or we have an infinite regress. But, once again, the main gist of Wittgenstein's attack on ‘the private linguist’ appears not to turn on this point either, as we shall now see.

2. The Private Object

In Part 2, §11, page 207,Footnote ² Wittgenstein says:

This passage would have some force if it were the case that the subject's experience comprises only private sensations. If that were the case, then it would make sense to suggest that a sensation may change but that you don't notice the change, for the only way you have of tracking such changes is your own uncheckable memory, which may well deceive you.

But this is not the nature of the subject's experience. Wittgenstein has clearly acknowledged that such a subject, the ‘private linguist’ who claims to be able to refer to his or her own private sensations, also has awareness of objects in the public, physical world. And it is this admission which undermines the sense of Wittgenstein's claim.

First, it implies the rejection of the principle ‘like cause, like effect’. Each time I knock my hand against the edge of my desk, it seems to me that I'm having the same sort of rather painful sensation. Wittgenstein suggests that I could be wrong. But that means that even though I have the same sort of cause on every occasion it may be that I have a different effect. And that involves the rejection of a principle for which there is solid inductive justification, and its replacement by a conception of a free-wheeling or free-floating relation between the world of private sensations and the world of public objects.

This passage also involves the rejection of the claim that my memory is pretty reliable, a claim for which there is, again, good inductive justification, and which can also be directly tested. I remember the number of the bus I have to take to get to the library, that I locked the door when I left the house, I remember how to get home, the date of my wife's birthday, the PIN numbers of my credit and debit cards and so on. I say, ‘that scent is coming from the sweet peas’, and I find that indeed it is. On a walk, I might say, ‘there must be some wild garlic around; I can smell it’, and sure enough there it is by the side of the path. I taste a whisky, and say, ‘surely that's Lagavulin’; and, indeed, that's what it says on the bottle.

What this does not involve, however, is any suggestion that my recognition of the private experience is always infallible. I may be quite sure that the drink I am tasting is that of a certain brand of whisky, only to find that it is a special bottling of that brand which has a slightly different flavour from the standard brand. I can confirm my mistake by tasting one immediately after the other. So, although it is true that every sensation must be as it seems, this does not mean that my classification of sensations, relying as it does on my memory of how past sensations have felt, must be infallible, since my memory, though good, is not infallible. I make this point because Crispin Wright does seem to think that belief in the privacy of sensations, the ‘private object’, commits one to the claim that our sensation-judgements are infallible.Footnote ³

But above all this, there is the extraordinary claim that we should ‘get rid of’ the private object. In our ordinary understanding, the range of sensations which any subject experiences are, indeed, private objects. Yet we are enjoined to get rid of them. How, then, are we to understand our reference to such experiences? Someone tries a new drink, and he assures me that it's nothing quite like anything he's tasted before, but that it is very enjoyable. I have no idea what that experience is; that experience is private to him. Does that mean that it doesn't exist, that it should be got rid of? This is a point to which we shall return.

§270 runs:

And §271 runs:

Now in spite of the suggestion in §270 that the problem may be that, although the sensation remains the same, I regularly misidentify it, there is no doubt that §271 entertains the idea that the same cause, i.e., ‘the usual symptoms and presuppositions’ may result in totally different effects – ‘totally different things’, as Wittgenstein says. And this is in accord with the claim of Part 2, §11. That is to say, the principle ‘like cause, like effect’ is rejected. But this is fatal to Wittgenstein's argument in §270. We would normally take it that the uncomfortable sensation is the result of what the manometer shows us to be the case, a rise in blood pressure. That is to say, the rise in blood pressure has two effects: the uncomfortable sensation and the unusual manometer-reading. But if, as I think is clear, the argument implies the rejection of the principle, ‘like cause, like effect’, then the manometer-reading can show us nothing; our assumption that it is the effect of a rise in blood pressure can no longer be relied on. It might, for all we know, be the result of the fall of a tree in China. If the principle, ‘like cause, like effect’ is rejected, then we can say nothing about the cause of either the sensation or the manometer-reading.

§271 brings further problems. Not only does it involve the rejection of the principle, ‘like cause, like effect’, and the inductively grounded claim that one's memory is pretty reliable, but it describes a possibility which is totally bizarre and utterly incredible. If it is the case that what I call ‘pain’ may in fact be a different sensation on each occasion that I use the word, and that I simply don't notice this, then not only is it suggested that the same sort of cause (‘the usual presuppositions’) might cause a whole variety of different sensations, thus clearly implying the rejection of the principle, ‘like cause, like effect’, but the fact that I fail to notice this at all involves the following amazing situation. The suggestion is that I may be woefully inept in the way of identifying my sensations, but this ineptitude is never revealed. Let us imagine a particular case. I suffer a usual cause of pain, a cut, blow, burn, or something of the sort, and feel a pain as the result, something that shows itself in a typical behavioural expression of pain. On this occasion, I get it right. On the next occasion, however, although there is again a typical cause of what we would normally call pain, and a typical behavioural manifestation of that feeling, I misidentify the sensation. It is not pain but, say, the taste of rhubarb – or whatever you like. Why, then, was my mistake not revealed? Because the world has changed around in such a way that what used to be the typical causes and effects of pain are now the causes and effects of the taste of rhubarb. This hyper-miraculous coincidence is repeated next time. This time, although once again we have a usual cause of what we call pain, and a typical behavioural expression of ‘pain’, what I actually experience is, say, the scent of a rose. Again the natural world conspires to hide my abysmal incompetence in the way of identifying my sensations. And so on. Now this passage is just about the most ludicrous piece of philosophy I have ever read, but it has been enthusiastically endorsed by pretty well every exponent of Wittgenstein's ideas.

3. Speaking of Sensations

Well, how do we speak of sensations in Wittgenstein's view? In §293 Wittgenstein concludes: ‘[I]f we construe the grammar of the expression of sensation on the model of “object and name” the object drops out of consideration as irrelevant’. I innocently assume that when I talk of, say, the scent of a rose the phrase ‘the scent of that rose’ picks out, or names, a certain experience or sensation. But if it is not doing that, what is it doing?

In fact, the only sensation Wittgenstein says anything about is pain. Part of §244 says:

This is totally inadequate even as an account of our talk about pain. I go to the doctor and tell him about a pain I've been having in my lower back. There is at this stage no behavioural expression of the pain: I simply report its occurrence. The doctor then feels around my back and asks me where it hurts and what sort of pain it is. On another occasion, I try to focus my attention on the pain in an attempt to control it, and to diminish it. I attempt to focus on what is essentially private to me: the pain.

If Wittgenstein's suggestion is a totally inadequate account of our talk about pain, it is completely hopeless if it is meant to provide some sort of account of our talk about sensations in general. Our behavioural responses to the great variety of sensations we have amount to those which suggest pleasure, those which suggest displeasure or perhaps disgust, and those which suggest indifference Wittgenstein surely did not assume that there was a distinctive behavioural expression for each and every sensation that we experience. How, then, do we manage to talk about sensations? Clearly not by reference to the various causes of our sensations, since, first, we may be entirely ignorant of the cause of a new sensation and, second, we often identify a sensation – a particular scent or taste, say – and only later establish what its cause is. Just think of the weekly articles in the papers recommending various wines, and the descriptions of the various flavours we might enjoy. And I read recently that one new perfume has a suggestion of liquorice and another a hint of mint…

Our ordinary view is that we do indeed refer to sensations, and that what we refer to is private to the subject. Wittgenstein's attempt to undermine the view, as we have looked at it so far, has totally failed. Other comments by him on the idea of privacy only add to the confusion. Thus, in §246 he declares that the assertion ‘only I know if I'm in pain’ is either false or nonsense. But, on the contrary, its sense is perfectly clear. That is why the doctor has to ask me if I feel a pain at this or that point as he feels around my lower back. And §248 declares that ‘sensations are private’ is like ‘one plays patience with oneself’, which is simply nonsense.

In §253 Wittgenstein questions the sense of the assertion that another person can't have my pains, and says, ‘In so far as it makes sense to say that my pain is the same as his, it is also possible for us both to have the same pain’. This betokens a pretty obvious failure to distinguish between qualitative and numerical identity. Someone else might have a pain qualitatively exactly similar to the pain I experience, but not that particular token of pain. No one else can have this chair while I'm sitting on it, or this watch while I'm wearing it, though someone else may be sitting on a qualitatively identical chair, or wearing a qualitatively identical watch. However, in cases such as these ownership is transferable. I may give up my chair to someone else, or give my watch to someone else. But in the case of mental particulars, ownership is non-transferable. I cannot transfer my ownership of the pain I am having, or my thought ‘it's a cloudy day’, to someone else, though others may have similar pains and similar thoughts. These distinctions I take to be absolutely obvious, though Wittgenstein rides rough-shod through them all. Underlying this failure may be a startling failure to acknowledge the reality of other selves, having the same sorts of experiences as one does oneself. Thus in §302 he says:

In fact, it's something we all do without the slightest difficulty. I can sympathise with someone who has experienced the pain of kidney stones, because I have had it myself. I can understand an amateur singer's nerves before a public performance because I used to do that sort of thing myself. And of course we all understand what it is for anyone to enjoy the pleasures of a warm, sunny day. Wittgenstein's point suggests a pathological condition rather than any sort of philosophical insight. Would Wittgenstein understand what it is to feel sympathy for, and to empathise with, the victims of the tragic events in Manchester and London this summer of 2017?

§304 introduces a new, and very puzzling, claim. Answering the charge that he seems to be asserting that the sensation itself is nothing, he responds that the sensation ‘is not a something, but not a nothing either!’ This strange utterance is hardly illuminated by the claim that what we need to do is to ‘make a radical break with the idea that language always … serves to convey thoughts – which may be about houses, pains, good and evil, or anything else you please’. This leaves us completely in the dark as to how we do in fact talk about sensations, an issue on which Wittgenstein's one suggestion about the way we talk about pain is clearly a non-starter. The only other possibility that occurs to me is that suggested by Norman Malcolm's treatment of the concept of dreaming. Malcolm's claim was that he did not purport to say what dreams actually were, but only to say what the criterion is which justifies our saying of someone that he had a dream. That criterion is simply that the subject is inclined to tell a sort of story in the morning, prefaced by the words ‘I dreamt that…’ But if that is the suggestion, its application to talk about sensations would compel us to say that there is some sort of behavioural criterion for each and every sensation we can distinguish; and this, as we have already seen, is clearly not the case. The approach might even suggest that great swathes of our mental life – judgements, intentions and emotions – have this curious status, being ‘not a something, but not a nothing either’. In brief, it is quite impossible to make sense of Wittgenstein's suggestion here.

§350 runs;

This passage, again very often referred to by commentators on Wittgenstein, and enthusiastically endorsed by David Pears,Footnote ⁴ is in fact extraordinarily confused. To begin with, the analogy with different time-zones simply misfires. There is not the slightest difficulty in understanding that, say, it is now 2 o'clock in London, 3 o'clock in Berlin, and 4 o'clock in Moscow. These are all the same moment in time, even though they belong to different time-zones. Next, we all understand what it is to judge that, say, we experience the same sensation on different occasions: it's the taste of almonds again, or raspberries, the scent of roses, or hyacinths. We hear the same note on different occasions, the A above middle C, or the same tonal quality, that of the oboe. We can judge of two pieces of cloth whether they are of the same colour. In short, there is no problem at all in our claim that we judge two or more sensations to have the same quality. We do it all the time.

But if this is not what we do, how do we make such judgements, on Wittgenstein's view? The idea that there is a distinct behavioural expression for each and every sort of sensation we experience is clearly a nonsense. We have been here before, and once again nothing coherent has emerged.

It is worth-while contrasting what is, in my view, Wittgenstein's wholly confused treatment of the notion of privacy, or the ‘private object’, with the essence of what has come to be known as the Knowledge Argument, originally stated in Frank Jackson's paper, ‘Epiphenomenal Qualia’.Footnote ⁵ Jackson's original statement of the argument, turning on the plight of Mary and her terribly restricted experience of colour, is, I think, open to question, but the core of the argument can be restated in a way which avoids any reference to Mary's predicament. Someone deaf from birth can not only be given a complete description of sound waves and what happens when they enter the human ear, but he can also be shown what happens. He therefore has complete knowledge of the physical aspects of hearing, not a mere descriptive knowledge. But he does not have the faintest idea of what it is like to hear. Equally, someone who lacks the sense of taste can be shown everything that happens in the human body when food or drink is consumed, but cannot have the slightest idea of what it is to taste something, and what it is for some tastes to be enjoyable and others disgusting. And those unfortunate people who cannot experience pain, and have to be watched like a hawk in case they damage themselves, can again be shown what happens in the human body when someone is in pain, but this will be to no avail. The sensations of hearing, of taste, and of pain are therefore intrinsically private. They are not part of what can be accessed publicly. Each token of such sensations is necessarily private to the subject, and its ownership cannot be transferred from subject to subject. This approach to the issue of privacy seems to me absolutely clear, but Wittgenstein's approach amounts to confusion upon confusion.Footnote ⁶

4. Privacy and Intention

This is the title of the final section of Crispin Wright's paper, already referred to. Wright correctly characterises the claim that ‘the intentional state is, typically, simply salient to consciousness, and directly recognised by the subject for what it is’, and says that this position is a major target of Wittgenstein in the Investigations. Wright goes on to assert that ‘knowing of one's own beliefs, desires and intentions is not really a matter of “access to” –being in cognitive touch with – a state of affairs at all’. How, then, should such knowledge be construed? Wright's proposal is as follows:

A number of questions are provoked by this passage. The first is that there seems to be any number of examples of the subject's having just the sort of privileged access to his beliefs, desires and intentions that Wright denies him. The door squeaks, and I say to myself , ‘I must put some WD40 on that hinge’. No one else knows about this. I stop, and ask myself, ‘What was I going to do next? Oh, yes…’, and then do it. I try to operate a machine, and get nowhere. ‘What are you trying to do?’, someone asks, and when I tell him he shows me that I was going about it in the wrong way. I walk round the local pond and think ‘the swans are looking beautiful, and there are four signets’. I think through, and enjoy, a favourite passage of music, silently. In attempting to go through the procedure of a relaxation exercise, I find that I'm constantly having a whole variety of thoughts instead of concentrating on my breathing. And so on. In all these cases, I want to say that I have access to something, my own experiences, beliefs, desires and intentions, an access which is privileged to me. Certainly, in none of these cases has there been any sort of expressed avowal. In the light of these obvious features of our experience, it is difficult to understand what Wright is actually denying in rejecting the notion of privileged access.

Although it is still unclear to me what exactly Wright means in this context by a ‘constitutive principle’, the following is what I take to be the position which Wright seems to be denying: I taste or smell something, say, and I find it absolutely delightful, or utterly disgusting. The state of consciousness, the delightful or disgusting sensation, is one thing and the avowal, the expression of delight or disgust, is another. I may express my delight or disgust in a public avowal, but what I avow, the state of consciousness, is private to me, and something to which I therefore have privileged access. Furthermore, what I believe, is that the scent of the rose, the scent of hyacinths, the taste of raspberries, the taste of an Islay malt whisky, and so on, are all delightful, and it is these individual experiences that I intend to repeat. These are clearly not the same belief, nor is the intention with regard to each of them in turn to experience that sensation again the same intention. The object of each of these avowals of delight, and of the associated avowals of intention, is something essentially private to the subject, something to which, contra Wright, the subject has privileged access. But the notion of privileged access is fundamental to our grasp of what it is to be a conscious subject, as I shall now try to make clear.

5. Avowals and the First Person

On the London Underground, if I remember correctly, one is sometimes enjoined to ‘mind the gap’. On the train that connects one terminus with the other at Gatwick Airport one is warned, ‘doors closing’. On an early word-processor, now long superseded, to go to delete anything provoked the response, ‘are you sure?’. One used to be able to dial the letters TIM on the telephone and get a beautifully articulated statement of the exact time. My microwave oven gives the instruction ‘open door’. When I switch on my computer it says ‘Welcome’. None of these is an avowal. One is not being told to mind the gap, or warned that the doors are closing. The word processor is not full of concern about the possibility that I might do something that I'll regret later. And TIM did not genuinely intend to inform me about the time. The microwave oven does not literally command me to open the door, nor does my computer literally extend a welcome to me. All these are mechanical or computer simulations of genuine avowals; none of them is the real thing.

Wittgenstein admittedly says that only of human beings can we say that they have desires and intentions but if avowals of intention and belief do not issue from the subject's privileged access to his or her own states of consciousness it is not at all clear why only human beings can give expression to them. Subjects can give orders, but so, apparently, can microwave ovens (‘open door’). Subjects can show concern, but so, apparently, can word-processors (‘are you sure?’). Subjects can advise, but so, apparently, can underground trains (‘mind the gap’). Subjects can inform, but so, apparently, can telephones (‘at the third stroke, it will be…). And so on. What has gone wrong?

Part of what has gone wrong is that Wright and others have focussed on a feature which is common to all entities which seem to be able to issue avowals of various kinds – warnings, commands, expressions of concern, and so on – but have not asked what it could be for one apparently arbitrary token of such avowals to be one's own, to be mine. David Chalmers has argued that while an explanation of consciousness might give one a conception of ‘points of view’ in general, it is hard to see how one apparently arbitrary point of view can be mine, unless solipsism is true. This seems to point to the need to posit in our ontology a ‘primitive indexical fact’: one particular consciousness is mine or me.Footnote ⁷ This echoes Nagel's point that a complete description of the world given without recourse to indexicals or token reflexives will fail to grasp the truth that one tiny sub-segment of that reality is mine, or me. And I have myself argued for the notion of the irreducible property of mineness.

Chalmers goes on to downgrade the importance of this claim, quite wrongly, I think. However, when Chalmers talks of an explanation of consciousness which might give one a conception of ‘points of view’ in general, it should now be clear that any attempt to give an account of consciousness in general, or points of view in general, will be likely to come up with a list of capacities which entities other than conscious subjects seem able to exhibit: issuing warnings and commands, expressing concern, and supplying information, and so on. And, while conscious subjects have a perceptual point of view, so, arguably, do television cameras. An individual might be able to scan his or her own body for signs of a medical condition, but computers can scan their own operations. But we know that the apparent avowals of intention, the giving of advice, expressions of concern, and the possession of a point of view in machines are happenings or states, while those of human beings are states of consciousness, and some are actions, or manifestations of agency.

What assures me of this is that I know in my own case that such avowals and states are states of consciousness, and that some are manifestations of agency. That is to say, I have privileged access to my own states of consciousness. And it is because I know in my own case that certain states and events in the world are not just states or happenings, but states of consciousness and sometimes manifestations of agency that I am able to infer that the best explanation of similar behaviour in other human beings is that they also are conscious beings, and that every such being has privileged access to his or her own states of consciousness. Our understanding of what it is to be a conscious subject is therefore necessarily tied to the notion of privileged access, and this notion in turn rests on what ‘I’ denotes as a primitive indexical fact or property.

Among other philosophers who have had difficulty in acknowledging this, I mention just two. Ryle, having rejected the notion of privileged access in chapter 1 of The Concept of Mind, declares in chapter 2 that ‘silent argumentation has the practical advantages of being relatively speedy, socially undisturbing and secret’ (page 47, my italics). That is to say, to state the obvious, I have privileged access to my own ‘silent argumentations’. What Ryle should have acknowledged is that I know my own ‘public argumentations’ to be genuine exercises of intelligence and not a mere computer simulation of such processes because I have privileged access to my own states of consciousness.

P.F. Strawson argues that the notion that I begin with an awareness of my own consciousness and its connection with behaviour and then argue by analogy to the existence of other selves must be a mistake, since I cannot have the concept of myself unless I have at the same time the concept of others who are not myself.Footnote ⁸ But this is a mistake. It is because I begin with an awareness of my own case, my own states of consciousness and the way in which they manifest themselves in behaviour, that I am able to recognise certain other particulars in the world to be persons, having a similar consciousness. I must begin with an awareness of what is in fact my own case, but it does not follow from this that I must begin with the concept of my own case. I may infer from the late appearance of a certain flower that the rest of the summer will be bad. That flower, perhaps, has a complicated botanical name of which I am ignorant. But clearly I can infer from the late appearance of the flower that has this Latin name that the summer will be bad without having the concept of that flower as designated in botany. The fact that I infer that q from what is in fact p does not entail that I must have the concept p. Strawson allows that on the view which he is attacking one might well notice that experiences stand in a special relation to this body. We surely need to ask here what entitled Strawson to pick out any one body as this body, and I think the only answer that makes sense here is that it is the body which is the centre of what he can go on to conceptualise as his experiences. At any rate, it is but a short step from the position that Strawson has reached to the observation that other bodies behave in a way similar to this body when certain experiences occur, and to infer from this that this indicates that there are other centres of experience, other conscious beings. This I infer from what is in fact my own case. To repeat, the fact that Strawson was able to pick out one particular body as this body should have indicated to him that he did after all have privileged access to the connection between one particular body and states of consciousness, the one he calls ‘this’, which he has yet to conceptualise as his own. A further relevant consideration is that, while the propositions ‘each body is identical with itself’ and ‘each body as the properties that it has’ are clearly necessarily true propositions, the proposition ‘one particular body is this (i.e., my) body’ is a contingent truth.

Strawson argued for what he called the primitiveness of the concept of a person, a double aspect view, and the argument fails for the reason I have just outlined. What is very strange in the light of Strawson's argument is that he declares at the end of the relevant chapter that ‘each of us can quite intelligibly conceive of his or her individual survival of bodily death. The effort of imagination is not even great’. This simply highlights the extraordinary uncertainty of aim in the argument of this chapter.

In summary, I find it astonishing and depressing that Wittgenstein's multiply confused argument should have had such an influence, and spawned such an enormous literature. At no point can I find an argument or a claim which is remotely acceptable. What reflection on this topic does bring out, however, is, firstly, how the anti-Cartesian stance, amounting to a venom, although now fading, has distorted the philosophy of mind, and, very importantly, the absolute primacy of the first-person perspective. That is something which the philosophy of mind still has to fully recognise.Footnote ⁹