“ Not a Something ”

Wittgenstein’s remark in section 304 of the Investigations that a sensation “is not a something, but not a nothing either” has often been connected with his critique of the “picture of an inner process”, and there is a temptation to read “something” as meaning “something private”. I argue that his remark should be taken more at face value, and that we can understand its purport via a consideration of the notion of consisting in. I explore this multi-faceted notion and its connection with (an extended version of) the Context Principle, beginning with the case of certain “propositional attitudes” and moving on to sensations. Wittgenstein was right to think it a philosophical prejudice to say that X’s being in pain, say, must consist in, be constituted by, something.


Introduction
Wittgenstein's remarks on sensations and sensation language have typically been discussed in connection with his critique of the notion of privacy, a critique generally thought to reach its apogee in the so-called "private language argument".Those commentators who agree that Wittgenstein's position does not fall into (anything that could be called) behaviourism will often be found saying something like the following: Wittgenstein does not deny the reality of sensations like pain, a reality that transcends behaviour and behavioural dispositions; he only denies that these psychological states or events are known only to the subject, that one learns a concept like pain by some act of inner ostension, that one can only "infer" that others are in pain, etc.These commentators might well regard the first part of his statement that a sensation "is not a something, but not a nothing either" as meaning that a sensation is not a private something.For of course pain is real enough -it's something -just not anything "private".
To forestall the charge of tilting at straw men, I will give some examples of this tendency.
Ernst Konrad Specht writes that "Wittgenstein is only opposed to an erroneous determination of the naming relation between a sensation word and a sensation, but he does not deny that such a relation exists".And he goes on: "Wittgenstein's discussion is also hostile to an ontological misinterpretation of sensation; more precisely to an ontology which interprets sensation on the pattern of an object which is supposed to be internal and private, in contrast to objects that are external and public" (Specht 1969: 94).
Bill Child also speaks of a semantic relation or connection holding between two somethings, the sensation word and the sensation, explaining Wittgenstein's view by saying that a child "learns to apply the word 'pain' to herself in circumstances where she feels pain.That effects a connection between her feelings of pain and her use of the word 'pain'" (Child 2011: 167).
John McDowell feels sufficiently uneasy about the wording of PI 304 to write that "Wittgenstein could, and perhaps should, have said something more like this.The sensation (the pain, say) is a perfectly good something -an object, if you like, of conceptinvolving awareness."And he goes on to suggest that "the conceptual content of the episode of consciousness" can "be parsed in terms of the classification of a something ('the sensation itself': the pain) as the kind of something it is" (McDowell 1989:  289-90).
There are certainly grounds for thinking that part of Wittgenstein's meaning, when he says "not a something, but not a nothing either", relates to the picture of the private inner object.After all, his next remark is: "The conclusion was only that a nothing would serve just as well as a something about which nothing could be said", a remark which harks back to his discussion of the "beetle in the box" (PI 293).But of course the conclusion of that discussion was that we should not "construe the grammar of the expression of sensation on the model of 'object and designation'".Wittgenstein's main aim is evidently to "reject the grammar which tries to force itself on us here" (PI 304, my  italics).
In this essay I want to argue that "a sensation is not a something" should not be interpreted as meaning "a sensation is not a private something".And the purport of Wittgenstein's remark can only be understood when we have done as he advises: that is, when we have made "a radical break with the idea that language always functions in one way, always serves the same purpose: to convey thoughts -which may be about houses, pains, good and evil, or anything else you please".

Context and Constitution
Frege's famous Context Principle urges us never to ask for the meaning of a word in isolation, but only in the context of a proposition.This idea of Frege's surely had a crucial influence upon Wittgenstein, early and late.In his later work, Wittgenstein can be seen as extending the principle beyond the domain of words, to apply -as we might say -to phenomena, such as the phenomena of pointing, or of hope, desire, expectation, or belief.Again and again he approaches a question about some phenomenon by locating it in its wider surroundings, or by inviting us to imagine or specify such wider surroundings for ourselves; for by doing so, we gain a better understanding of the significance of the phenomenon.And here the word 'significance' should be taken broadly, so as to connect both with questions about the identity of a phenomenon and with ones about its importance for us.
Wittgenstein's remark that "an 'inner process' stands in need of outward criteria" (PI 580) is an instance of the sort of thing I am talking about; for he does not only, or even primarily, mean "behavioural criteria", as the immediately following remark illustrates: "An expectation is imbedded in a situation, from which it arises" (PI 581).But of course we can include behaviours as part of the surroundings of an "inner process".
The phrase 'inner process' has scare-quotes around it in PI 580, and much has already been said by Wittgenstein before this remark concerning the problematic idea of the "inner".He has also looked critically at the idea of a "process", e.g. in the following passage: How does the philosophical problem about mental processes and states and about behaviourism arise?-The first step is the one that altogether escapes notice.We talk of processes and states and leave their nature undecided.Sometime perhaps we shall know more about them -we think.But that is just what commits us to a particular way of looking at the matter.For we have a definite concept of what it means to learn to know a process better.(The decisive movement in the conjuring trick has been made, and it was the very one that we thought quite innocent.)-And now the analogy which was to make us understand our thoughts falls to pieces.So we have to deny the yet uncomprehended process in the yet unexplored medium.And now it looks as if we had denied mental processes.And naturally we don't want to deny them.(PI 308)   "We talk of processes and states and leave their nature undecided."In speaking of remembering, or of intending, or of being in pain, as processes or states we may well already have started down a wrong path, not because a statement like "Pain is a mental state" is just false, but because the picture that typically accompanies such a statement is liable to lead us into confusion.
A symptom of this sort of confusion is succumbing to false (i.e.bad) analogies.There are a number of sources for the picture of a sensation as a bona fide "something".In what follows I want to bring out how one of these sources may be a certain false analogy relating to that very context principle the Wittgensteinian version of which is the beginning of much wisdom.This particular source, in fact, is associated more with an anti-Cartesian stance than with a Cartesian one -to put the matter simply.But regardless of issues to do with philosophical error and its sources, it is illuminating, I think, to come at Wittgenstein's thought ("not a something, but not a nothing") from this direction -from the direction of the context principle.
The word 'rat' only has meaning as it occurs in certain contexts, such as "There's a rat in the garden".In the sentence 'Socrates loved Plato' the syllable 'rat' occurs, but not the word 'rat'.-This is a perfectly natural way of talking.If we ask, "What is it that in this case has meaning only in the context of a sentence?"we can reply either "a word" or "a syllable".But for the word 'rat' to occur in a sentence, it is typically necessary that a certain syllable do so, by which is meant a certain sound or shape.The criteria of identity for such sounds and shapes are quite liberal, and are themselves dependent to some extent on context; differing regional accents will yield very different sonic versions of 'rat'.Moreover, we might allow a coded equivalent of 'rat' to count as the word 'rat': a squiggle of a certain shape, say, or the sound bingle-bongle. 1 This is why I said that it is only typically necessary for a certain syllable to occur for the word 'rat' to occur.But there must in a given instance be some description D, different from the description "the word 'rat'", such that D is brute relative to "the word 'rat'", and such that we can say that the occurrence of the word 'rat' consists in the occurrence of something of which D is true (e.g.consists in the occurrence of the squiggle).One could say: a word can't just occur; its occurrence must consist in the occurrence of something else.And "something else" points not to another thing, but to another description of the same thing. 2Similar points can be made in connection with many processes, events, actions, etc.Only in a certain context does my pushing a bit of wood count as my checkmating you.Moreover, I can't just checkmate you: there must be something I do in virtue of which I checkmate you -which is to say, there must be some description of what I do which is brute relative to the description "checkmating'.
We might be led to think that wherever some version of the context principle applies, so does a correlative principle about there being some "more brute" description of the phenomenon in question.That this is not so has been pointed out by Martin Gustafsson (forthcoming 2017) in a paper which discusses the comparison made by Wittgenstein between words and chess pieces.A king is only a king in the context of a chess game and of the rules of chess; but a king need have no embodiment, for, as Gustafsson says, "there is such a thing as playing chess without any physical pieces or physical board; this is equivalent to both players 'playing blindfolded'".What enables a piece to count as a king as opposed to a pawn in a game of blindfold chess, is the knowledge both players have of the starting position.The king's moves are then effected simply by a player's calling out a move.There is no description of the white king which is more brute than "white king".To say this is not (by the way) to rule out the thought that the activity of blindfold chess is parasitic, perhaps even conceptually so, on the normal "embodied" game of chess.
Of course, when in a game of blindfold chess I call out "king to bishop four", it is my doing so that constitutes the event of my moving my king.So there is a description of the move that is brute relative to "Roger's moving the white king to bishop four", namely: "Roger's saying the words 'king to bishop four'".Note how the grammar of the word "description" is here being determined.(There is not only one thing which may be called description.)In this case we are exhibiting what can count as a "further description of a chess move".
possible defeaters (e.g."they are not acting in a film"), since the list of possible defeaters is open-ended.
So far we might say: a king in chess can just be a king -but a chess king's moving must also be something else, something more brute.Does this betoken some deep difference between objects and events?-something for the metaphysicians to get their teeth into?Well, inheriting a house only counts as such in a context of laws, legal documents, property, and so on, so to that extent the context principle applies; but if in turning eighteen a girl inherits the house left to her in a will, shall we say there must be a description of this event more brute than "inheriting the house"?Do I give a further description of that event by saying, "She's just turned eighteen"?Or indeed by saying, "She's just turned eighteen and there's a will which says such-and-such"?3If we do speak thus -and why forbid it?-we are evidently adopting a certain, perhaps rather novel, grammar for the expression 'description of an event'. 4or that matter we could if we wish decide to say that the white king in a game of blindfold chess does have a more brute description, namely: "the set of utterances by players A or B of the word 'king', spoken as part of a move-utterance, and bearing suchand-such relations to other move-utterances made by A or B".
This last proposal might be objected to on the grounds that you cannot replace 'the white king' by 'the set of utterances, etc.' in the context of a proposition.You can hardly move from "The white king has taken a black pawn" to "The set of utterances, etc. has taken a black pawn".A set of utterances can't take a pawn! 5 This objection relies in effect on Leibniz's Law (i.e."When a = b, everything that is true of a is true of b, and vice versa").But our notions of being constituted by, and of different descriptions applying to one thing, are too various to be constrained by Leibniz's Law.A credit card can be redescribed as a piece of plastic, but you can't move from "I've cancelled my credit card" to "I've cancelled a piece of plastic" -and Napoleon's becoming emperor can be redescribed as his putting a crown on his own head, but you can't move in the opposite direction from "Napoleon put a crown on his head with both hands" to "Napoleon became emperor with both hands".
The point of these remarks is not to make out that all determinations of the grammar of 'describe', in connection with formulating "descriptions of the same thing", are equally justifiable, or are all on a par, or anything of the sort.It is rather to point to the variety of things that do, and can, count as cases of X's being constituted by Y.Some of these cases are paradigm cases, others more peripheral, others quite stipulative.(When I said above that in blindfold chess there is no more brute description of the white king, I meant that only stipulation could supply such a description.)My overall aim is to prepare us (soften us up) for what ought to be said -or ought not to be said -about sensations, and indeed about other "mental events".

"I was thinking"
Let us start with some of these other mental events.I want to look specifically at statements involving what are called intentional verbs -verbs like 'want', 'believe', 'think', 'decide', 'regret', 'hope', 'expect'.In connection with statements involving such verbs there is a useful, if rather rough-and-ready, distinction to hand, namely that between disposition-statements and episode-statements.Because our eventual quarry will be sensations, let us focus on the latter. 6Examples might include "After investigating the wound for a few seconds, the doctor decided to apply a tourniquet", and "When I said those words I was thinking of Margaret Thatcher".In these examples, the deciding and the thinking, we would like to say, take place in real time; some might call the first an event and the second a process.Wittgenstein argues, in ways I won't now rehearse, that such decisions and thinkings can only have the content they do -that is, can only be those decisions and thinkings -on account of the wider context, both that of the subject's biography and that of the world more generally (e.g. a world containing Margaret Thatcher, in the second example). 7To that extent the context principle (in the sense I have been using the phrase) applies to such decisions and thinkings, if Wittgenstein is right.Quite a few philosophers would be happy to agree with this, especially those who endorse "externalism" about intentional content.
In what did my thinking of Margaret Thatcher consist?This is not intended as a question about the nature or essence of thinking, or about the meaning of the expression 'to think of Margaret Thatcher'; we are not here in danger of adopting the strategy described by Wittgenstein at PI 316: "In order to get clear about the meaning of the word 'think' we watch ourselves while we think; what we observe will be what the word means!"One can after all note that in a given context the white king was constituted by a bit of balsa wood without being tempted to regard the description "bit of balsa wood" as helping us in understanding what a white king is, what it is for something to be a white king.Nevertheless, if we look for something in which my thinking of Margaret Thatcher consisted, and if we require that it be something about which I myself could tell you straight off (as I can tell you I was thinking of Margaret Thatcher), then we will either find nothing that satisfies us, or will be faced with one or more of an indefinitely wide range of phenomena, all of which might on different occasions be etc., as it does to pains and itches.Wittgenstein's net can usefully be cast wider than he happens to cast it in PI 304. 7The sort of contextuality which is at issue is illustrated by the following: "An intention is embedded in its situation, in human customs and institutions.If the technique of the game of chess did not exist, I could not intend to play a game of chess" (PI 337).regarded as constituting my thinking that thought -e.g.imagining her face, saying her name, frowning characteristically, pronouncing certain words with a doleful emphasis, etc.All these phenomena might qualify as things that constitute my thinking of Margaret Thatcher, if we are content to use as our central criterion of "constitution" the fact of something's going on in or with the subject at the relevant time -which in this example just means the time when I was uttering the words in connection with which I said, "When I said those words I was thinking of Margaret Thatcher".
Of course not anything going on in the subject at the time will strike us as a likely candidate for constituting a bit of thinking.The person's fidgeting with his pen will not do so.If we are talking of things about which the person himself could tell us straight off, then it seems that only those things which themselves may be described in terms of the thought will be candidates; thus the doleful emphasis is a candidate because we can hear it as, and describe it as, "the spoken dolefulness so expressive of his attitude to Margaret Thatcher". 8Note that there is little mileage in the idea that any special connection exists or gets set up between the expression "thinks of Margaret Thatcher" and any or all of such assorted phenomena as having an image, frowning, writing, etc.
What of the criteria for my having thought of Margaret Thatcher?Such criteria lie elsewhere, in how I respond to questions, how I explain myself, and the like.I may of course never be called on to do these things.The temptation now arises to invoke counterfactual conditionals -as, "If I had been asked what I was up to, I should have said 'thinking'; and if I had been asked what I was thinking of, I should have said 'Margaret Thatcher'"and then to say that here we have the something in which my 8 I do not think that this in itself throws doubt on the idea that we could be dealing with descriptions of a phenomenon that are brute relative to the original description (viz."NN's thinking of Margaret Thatcher").A description may be conceptually dependent on the description than which it is "more brute".Thus a spoken sound is only a candidate for being that in which the word 'rat' consists on a certain occasion in virtue of our being able to hear it, and take it, as that word.One cannot even in principle assemble all the sounds of a certain (sonic) sort, together with the contexts of their production, and thereby determine that they count as utterances of the word 'rat'.See Teichmann 2016.thinking of Margaret Thatcher consisted on that occasion.For we seem to have lit on something that existed or subsisted at the relevant time.The counterfactual was true then, we want to say; so haven't we got a further description of the thinking, a description more brute than "NN's thinking of Margaret Thatcher", and which moreover could plausibly be taken as the something with which the description is semantically connected?
As with the case of the stipulative identification of the white king with a set of utterances, there is no reason to forbid saying that someone's thinking of Margaret Thatcher consisted in the truth of a counterfactual conditional(s).There is however good reason to discourage it, since the proposer of it is likely to be a philosopher who thinks he has hit upon something substantive and important, something going beyond "mere criteria".For present purposes it is enough to note that if my thinking that p did consist in the truth of a counterfactual conditional, we could hardly speak of this thinking as taking place -i.e. as being an episode or "mental process".
We are considering what a certain alleged process or episode consists in; and if our model or paradigm of this is something like Napoleon's becoming emperor consisting in his putting a crown upon his head, we might end up deciding that, in the case of a person's thinking of Margaret Thatcher, there is nothing in which it consists.Frowns, imaginings, and the like seem better candidates than counterfactual conditionals, but there seems no rationale for picking out certain of these phenomena rather than others, in the case where several took place -and taking the thinking to have been constituted by the lot of them looks a bit desperate.Or rather, it seems to be motivated only by the thought that the thinking must consist in something.
So perhaps we should allow the possibility that one can just think of Margaret Thatcher, in the sense in which a white king in blindfold chess can just be a white king.Such a conclusion might on the face of it be palatable to some of those externalist philosophers who want to say that intentional contents "aren't in the head".Externalists who also lean towards functionalism will however resist the idea, preferring to say that a person's thinking of Margaret Thatcher must, on a given occasion, consist in the person's being in some "intrinsic", probably neurological, state.This position is often called "token-identity theory", and it has attracted many people.What should be said about such a proposal?

Token-Identity Theory
When I looked at the idea that thinking of Margaret Thatcher might consist in such things as pronouncing words with a doleful emphasis, I noted that we were only considering phenomena about which the person could tell us straight off.Such phenomena do not of course include brain-states and the like.If when you said you were thinking of Margaret Thatcher you were in fact talking about a brain process that had taken place within you, this is not something of which you could ordinarily be aware.But after all, could it not also be said that when you told us you had moved your king you were in fact talking about moving a collection of molecules of such-and-such description, despite the fact that this is not something of which you could ordinarily be aware?
But consider: if you are teaching a child how to play chess, using a normal chess set, you will count on the child's being able to take a certain piece of wood (say) as the white king; and it makes good sense to say that the child comes to call that piece of wood "the white king".Does the child come to call that collection of molecules "the white king"?
The sense of the verb "to call" in which I am here interested is one that concerns a person's mastery of language.What can the child who has cottoned on to our teaching do?What do we want to get her to do? -and what will we take as her successfully doing it?She needs, for example, to show that she can tell the white king from the white queen; will this be a matter of her telling the difference between this set of molecules and that set of molecules?No -for if it were, we would not be able to know by straightforward observation that she had cottoned on to this aspect of the game.We can know, and indeed she can know, that she has distinguished those two pieces on a given occasion, and normally this is because we and she can know that she has distinguished those two pieces of wood or plastic or whatever.
But surely she might not know that she is talking about a piece of wood?(She might not have heard of wood.)Nor need she know that the king is of such-and-such a shape -at any rate, if such "knowledge" is to be articulable.Perhaps then she has some sort of implicit or practical knowledge, that she is talking about and moving a piece of wood and that it is of a certain specific shape?
We do not need to say any of this.She can recognise this -that is to say, recognise the white king, say "This is the white king", etc.When I said that it made good sense to say that the child comes to call that piece of wood "the white king", that was only because our knowledge that she has successfully and intentionally moved her king is typically articulable using descriptions more brute than "white king", descriptions like "that piece of wood".We can say, and agree, that the child recognises this piece of wood as the white king.But the possibility of our articulating our knowledge in this way, although typical, is not necessary.We can, in the context of teaching a language, know that a child is able to tell the difference between dim lighting and bright lighting, even though we typically cannot articulate our (or her) knowledge using expressions more brute than "dimly lit" and "brightly lit" -cannot specify things in which given cases of being dimly or brightly lit consist.
Does this then open the door to the identity-theorist's claim that when I say I was thinking of Margaret Thatcher I call a certain brain process "my thinking of Margaret Thatcher", although I recognise it only as my thinking of Margaret Thatcher?Hardly.For we do not yet have any reason to say that I am talking about anything when I utter the words "I was thinking of…" -unless of course we merely want to recapitulate the surface grammar of the sentence by saying that I am talking about thinking.In this sense of "about" we can talk about unicorns or time travel; "about" used in this way carries no metaphysical or philosophical weight.To say that I was talking about thinking will then cast no light on what a statement like "I was thinking of Margaret Thatcher" is doing.In any more robust sense of "about", the idea that I must be talking about something when I utter the words "I was thinking of…" is a philosophical prejudice, as Wittgenstein says or implies when he urges us to "make a radical break with the idea that language always functions in one way, always serves the same purpose: to convey thoughts -which may be about houses, pains, good and evil, or anything else you please".
One can indeed say that in talking of the white king a child is talking about a piece of wood, or a set of molecules, if by this one means that in that situation the king can in fact be investigated and a description arrived at -e.g.'piece of wood' -such that it is okay to say that the king is constituted by something of which that description is true -in one of the various possible senses of the expression 'constituted by'.But such an investigation may be impossible, as with Gustafsson's blindfold-chess king.In that case, we can only say things like, "When she said the white king had taken a pawn, she was talking about the white king".In the same way you can, as we saw above, say: when Smith said "I was thinking of Margaret Thatcher" Smith was talking about thinking.
Even where there is such a thing as an investigation of X, and that investigation does yield an answer to "What constitutes X?", so that in talking about X one can be said to talk about Y, this will not in itself enable us to posit any sort of association in thought, or semantic connection, between the idea or expression "X" (on the one hand) and Y (on the other).In particular this is because there will be no guarantee that the criteria of identity through time of Xs and of Ys will be in harmony.The white king to be found at the end of the game is the same piece as the white king to be found at the game's start, but will not be the same set of molecules.9Hence no semantic or psychological connection can have been set up between "white king" and a given set of molecules.
However, this last point is more of an objection to the views of some of the Wittgenstein commentators I mentioned at the start of this paper than it is to the views of identity-theorists.An identitytheorist will probably not talk of words or concepts at all, except by way of the assertion that a psychological statement must be about a something, i.e. a psychological state, in some robust sense of "about".This state will be what constitutes NN's thinking of Margaret Thatcher, for example.In order to home in on this state he adds the assumption that whatever the state is must be something that was "going on" or subsisting at the relevant time.What is the relevant time?The most obvious answer will be: "It is the time indicated by the surface grammar of the psychological statement".So if I say "When I uttered those words I was thinking of Margaret Thatcher", then the brain process in question can be taken -at least as a first shot -as simultaneous with that earlier utterance.
Why a brain process?Indeed, why any process going on in this human being?Surface grammar does not help us in this case, since (as Wittgenstein pointed out10 ) the pronoun 'I' does not pick out a particular human being -at any rate not in a statement like "I was thinking of Margaret Thatcher".Of course it is true that it was a human being who was thinking of Margaret Thatcher, since I am a human being and I truthfully uttered those words.Why not then take the process allegedly constituting my thinking of Margaret Thatcher to be a process involving a whole human being?To be sure, we have found that various kinds of damage to the brain hamper or destroy such things as the ability to say "I was thinking…", and have made other related findings, but since something that happens in or to a human brain ipso facto happens in or to a human being, this can be no objection to claiming that it is a process of the whole human being which constitutes my thinking of Margaret Thatcher.Such a claim would appear to be as respectable as the identity-theorist's claim from the perspective of neurological enquiry, the perspective which is evidently guiding our thought here; for it does not preclude a scientist's going to the relevant bit of a human being, her brain.
Be all this as it may, the identity-theorist in fact prefers to identify the thinking (or whatever) with some neurological state or process.Insofar as this homing-in is motivated by the assumption that a psychological statement must be about something it is philosophically questionable.And there are further worries, concerning the assumption that the surface grammar of a sentence can guide us to a proper determination of when something happened.In various places Wittgenstein brings out how, and in what sense, a person's subsequent assertions "make the connection" between an earlier statement and some intentional object, something which if true would suggest that nothing going on at the time of the earlier statement determines that connection.This indeed would not rule out my saying something true when I assert "I was thinking of Margaret Thatcher when I said those words", for the notion of truth need not be tied to the idea of something's making the assertion true. 11oming in on brain processes that went on during the time I was saying the words in connection with which I later said "I was thinking of Margaret Thatcher then" may simply be looking in the wrong place altogether -which is not to imply that there is any right place.Analogously, if a court declares a contract made three years ago to be null and void, you would be looking in the wrong place if you investigated what went on at the time when the contract was made in the hope of finding its nullity.The court may base its later decision on facts about what went on then (though it need not), but it is the decision which renders the contract null.12

Investigation
A key notion in all this is that of investigation.I imagined above someone's investigating a white king and finding, or confirming, that it is a piece of wood of a certain shape, or that it is a set of molecules.It was because such an investigation can yield such answers that it is correct to say that the white king consists of, is constituted by, a piece of wood (or set of molecules), at any rate at the time of the investigation.What counts as investigation in this case is relatively clear.These facts together show that this particular claim about X's being constituted by Y is a non-stipulative one.By contrast, no investigation of the white king in blindfold chess could ever show that it was constituted by the set of utterances by players A and B of the word 'king' made in the context of such-and-such move-utterances; this constitution-claim is a stipulation.
There is also a fairly clear sense of "investigation" applicable to events and processes.Having been invited to attend the ceremony of Napoleon's becoming emperor, you could (in principle) investigate or observe the event with a view to saying, or confirming, what it consisted in -the investigation yielding such an answer as: he put a crown on his head (in certain surroundings).There will evidently be some indeterminacy as to how much of the phenomenon gets investigated, and correlatively as to how compendious our answer is to "In what did it consist?"These would be matters for stipulation.We can say either that my payment of a debt to you consisted in my writing a cheque and handing it to you, or in these things plus your cashing the chequeor other things along these lines. 13urning to the position of the token-identity theorist, there will be two questions facing one who adopts such a position: first, whether there is such a thing as an investigation of someone's "being in mental state S", and second, whether such an investigation yields the sort of answer predicted by the theorist.It is the first question which is likely to be really problematic, in particular in cases where it seems that a "connection" is made, e.g. between thought and object, by the subject's subsequent statements or actions.But even if we can safely assume, of some mental process, that the surface grammar of the relevant psychological statement(s) is in fact a suitable guide to when it occurred, it is still entirely possible that there is no such thing as an investigation of that process, an investigation of the sort that could tell us what it consisted in.
There is no such thing as an investigation of the white king in blindfold chess, nor of Samantha's inheriting a house on her eighteenth birthday.You can, to be sure, describe the background facts, the surroundings, in virtue of which there is a white king and in virtue of which Samantha inherited the house; but the phrase 'in virtue of', used thus, takes us to criteria and the context principle, rather than to constitution.Using Aristotelian terminology, you could say that this "in virtue of" indicates a formal cause rather than a material cause.It is true that, quite often, we derive what shall count as a material cause from formal causes: thus, the background facts which together formally (i.e.conventionally) determine that I have given you a book may include my saying "Take this; it's yours", something which in turn can be called the material cause of my giving you the book, i.e. that in which my giving it you consisted.(For this reason, to talk of investigating the event to see what constitutes it, as I did in connection with Napoleon's coronation, is admittedly to use 'investigate' in a rather extended sense.)But when the identity-theorist says of some mental process that it takes place in virtue of the occurrence of a brain process, the phrase 'in virtue of' is intended to signify a material cause not simply derivable from formal causes, or criteria for that mental process; the putative material cause is conceived of as discoverable independently of human customs and institutions.And the fact is, there might be no material cause of this kind.
The idea that there must be a material cause, a something in which the mental process consisted, derives in part, I think, from a species of false analogy, as I mentioned in my introduction.Impressed by the apparent contextuality of many psychological phenomena, by the fact of their necessary dependence on "external factors", behavioural and non-behavioural, we may fall into the trap of thinking them akin to those phenomena -of which there are many -to which the context principle applies and for which more "brute" descriptions can be found, yielding truths of the form "X consists in Y".In many cases, the context principle does only apply because of some convention, custom, or what have you, according to which Fs are counted as Gs -e.g.such-and-such shapes all count as the word 'rat'. 14This means that an F "by itself" will not be a G, but will only be a G given the right background; which is to say, the context principle applies to Fs.Hence, in these cases, the application of the context principle is tied up with the fact that Gs are constituted by Fs.But the contextuality of my thinking of Margaret Thatcher is not like this, and there is no analogous reason for saying that it must consist in anything at all.

Sensations
But what about sensations?What about pain?Surely there is little temptation to think that it is my subsequent statements or deeds that determine my having been in pain, in any sense of 'determine'?The central status of the first-person, present-tense expression of pain is admitted and even stressed by Wittgenstein himself.This status remains crucial in the face of all those facts to do with behavioural criteria, the learning of the word 'pain', etc.And as Wittgenstein's interlocutor says at PI 296: "Yes, but there is something there all the same accompanying my cry of pain.And it is on account of that that I utter it.And this something is what is important -and frightful."Isn't there a possibility of investigating this something?If it "accompanies" someone's cry of pain, and if the scientists are right, then won't we be investigating someone's pain if we put them in a brain scanner and observe their brain when they cry out in pain?And if we keep finding the same sort of thing in experiments like that one, won't we have good grounds for saying things like "A person's migraine will generally consist in such-and-such a brain process"?
The notion of X's consisting of (or in) Y is, as I have argued, flexible, protean even.Cases of constitution include paradigm cases, peripheral cases, and stipulative cases, all these cases sitting on a continuum, or on more than one continuum.We should always ask, "What is achieved by adopting this way of talking?-by saying that X consists in Y?" And we may well find that there is point in saying that Jane's migraine consisted in the electrochemical activity in part of her brain.Saying such a thing does not, after all, rule out other ways of speaking, such as the one sketched earlier, according to which it is changes in Jane, regarded as a unitary subject, which constitute her migraine, or her suffering a migraine. 15Adoption of such ways of speaking need not betoken succumbing to any philosophical prejudice.But it is a philosophical prejudice to think that there must be something that constitutes a person's pain, a something about which that person is talking, in a robust sense of "about", and to which she is referring when she says "I have a headache", or "That hurts!" -or for that matter "Ouch!"And this prejudice is surely false.It is false in the same way as the corresponding prejudice about the white king in chess.A white king can in fact just be a white king, for example in a game of blindfold chess.Likewise, a person with water in his head, who talked and acted just as I do and who, on stubbing his toe, cried out "Ouch!My toe!" would mean just what I mean when I say those words in that sort of situation -and would, consequently, be in pain just as I am.For the brainless person, being in pain, we might say, just is being in pain. 16Of course we do not believe that brainless people who act "just like us" are biologically possible, and if we read about such a case, we would no doubt discount the story on the grounds that it was too much in conflict with too much of our belief system.But there is nothing conceptually amiss, so to speak, in the hypothesis of the brainless person who is one of us, in the important sense of that phrase. 17he philosophical prejudice of thinking that there must be something that constitutes a person's pain is fed by the sort of expostulation Wittgenstein's interlocutor gives us, which I will repeat: "Yes, but there is something there all the same accompanying 15 An analogy might help: one could either say that Emma's crouching consisted in her legs bending a certain way, or in her whole body adopting a certain posture. 16And the question "Where is his pain?" would have a straightforward answer, just as it has in my case -viz."In his toe".-It may be added that the possibility of phantom-limb pain does not force us to "relocate" people's pains, e.g.into their skulls.This move is not even necessary in the case of those actually suffering from phantom-limb pain. 17See Wittgenstein, Zettel 608-610; also Teichmann 2001.my cry of pain.And it is on account of that that I utter it.And this something is what is important -and frightful."Wittgenstein's response to this is very instructive; he does not disagree or demur, but says: "Only whom are we informing of this?And on what occasion?"(PI 296) It is one thing to assure your friend or your doctor that your cry of pain was authentic (not feigned, not hysterical, etc.), to beg them to take it seriously, and so on -it is another thing to wield the philosophical picture of a psychological or semantic connection between two somethings, viz.pain-language and what pain-language is "about".
It is significant that our attempts to express this picture so often take the form of empty statements.
The very fact that we should so much like to say: "This is the important thing" -while we point privately to the sensation -is enough to shew how much we are inclined to say something which gives no information.(PI 298)   Part of Wittgenstein's point in this passage relates to the imagined use of the demonstrative "this" -a futile clutching of air, as it were.But there is more to the point he is making than that.Thus it is empty to assert, as anything other than a protest of sincerity, "When I said I was in pain, I was really talking about my pain!It was because I was in pain that I said I was in pain!" Empty and true, if you like, but empty all the same. 18It is similarly empty to say "When I said my king had taken your pawn, I was really talking about my king!"We can imagine the person who says this trying to give substance to her claim by carrying on: "Only my king could take that pawn -the queen couldn't, for instance.And without the white king the game would fall to the ground, you would be unable to checkmate me, our tactics would be thrown into disarray…" But all this is, after all, compatible with our playing a game of blindfold chess.