WILLIAM CHILD
We can distinguish two uses of the terms “inner” and “outer” in Wittgenstein’s writings on philosophy of mind. In one use, the term “inner” – like the term “mental” – is a general label that gathers together a particular class of phenomena: sensations, thoughts, intentions, emotions, and so forth. To say that something is an “inner” phenomenon in this sense is equivalent to classifying it as a mental phenomenon; it carries no particular philosophical commitments about the nature or status of the mental. It is in this vein that Wittgenstein says that “the inner […] is sensations + thoughts + images + mood + intention, and so on” (LW I §959) or that “the inner differs from the outer in its logic” (LW II 62).
In other contexts Wittgenstein uses the terms “inner” and “outer,” often in scare quotes, in a different way: to express a particular philosophical picture of the mental, a particular way of conceiving of the mental and its relation to behavior and the nonmental world. It is in this sense that he writes that “the ‘inner’ is a delusion” (LW II 84). He does not mean it is a delusion that we have sensations, thoughts, and so on. What is a delusion is “the outer–inner picture” (LW II 69): the picture of sensations and thoughts as internal, mental phenomena that are hidden behind the outer surface of people’s behavior. He regards this “picture of the inner and the outer” (LW II 28) as a distinctively philosophical way of thinking of the mental. It plays no part in the ordinary practice of applying mental terms to ourselves and others; but it naturally suggests itself – it “forces itself upon us” (PI II iv) – when we step back from the practice and consider it reflectively. But the inner–outer picture, Wittgenstein argues, fundamentally misrepresents the phenomena. “The whole complex of ideas alluded to by [the word “inner”],” he writes, “is like a painted curtain drawn in front of the actual word use” (LW II 84).
The first section of this chapter discusses the inner–outer picture: it explores Wittgenstein’s account of the origin and appeal of the picture, his reasons for rejecting it, and his own – very different – way of thinking of common‐sense psychology. The second section considers his account of our relation to our own experiences and attitudes, and discusses his suggestion that utterances like “I’m in pain” or “I want an apple” are avowals or expressions of a person’s experiences and attitudes. The third section discusses Wittgenstein’s positive view of the relation between “inner” mental states and “outer” behavior.
The inner–outer picture, as Wittgenstein describes it, has a metaphysical and an epistemic dimension. The metaphysical dimension is the image of a person’s mental and non‐mental properties as belonging to separate, ontologically distinct realms. There is an outer realm of behavior, utterances, facial expressions, external circumstances, and so forth. And, lying behind that outer realm, there is an inner realm of sensations, thoughts, intentions, and the rest. There are causal relations between the two realms: a person’s outer behavior is produced by the thoughts and intentions in her inner world; and changes in her inner world (sensations, for instance) can be caused by events in the outer world. But there are, as Wittgenstein puts it, no “logical relations” between the two realms; the inner and the outer are “logically independent”.
The epistemic dimension of the inner–outer picture involves two ideas. First, each of us is directly acquainted with our own thoughts and sensations. So we all have certain knowledge of what we ourselves are thinking and feeling. Second, no one can be acquainted with another person’s thoughts or experiences. So one person can never really know what another person is thinking or feeling; others’ thoughts and feelings are hidden behind their outer behavior. As Wittgenstein puts it: “One has to guess at [the mental] in someone else using external clues and is only acquainted with it from one’s own case” (LW II 61).
What makes the inner–outer picture so natural? We can begin with the obvious fact that there are asymmetries between first‐person and third‐person uses of mental terms. Our ascriptions of experiences and mental properties to other people are based on observing their circumstances and behavior. But we normally ascribe experiences and mental properties to ourselves without observation and without reference to our own behavior. Wittgenstein illustrates the contrast with an example:
What is the criterion for the redness of an image? For me, when it is someone else’s image: what he says and does. For myself, when it is my image: nothing. (PI §377)
Furthermore, there is normally no room for error about one’s own present attitudes or experiences:
If I ask someone ‘whom do you expect?’ and after receiving the answer ask again ‘Are you sure that you don’t expect someone else?’ then, in most cases, this question would be regarded as absurd. (BB 21)
The same goes for the questions “What are you feeling?” or “What sensation are you experiencing?” But there is nothing at all absurd about the equivalent questions in the third‐person case. For one person can easily be mistaken about another person’s attitudes or experiences.
Part of the appeal of the inner–outer picture, Wittgenstein thinks, is the sense that it explains these first‐person/third‐person asymmetries. Suppose we think that each of us is directly acquainted with the contents of our own inner world, and that no one is acquainted with the contents of anyone else’s inner world. That seems to explain how we can have certain knowledge of our own experiences and attitudes without attending to our behavior, and why we are more prone to be wrong about other people’s mental states than about our own. As Wittgenstein puts it: “It is obvious what justifies [the] picture” of “something inner […] which can be inferred only inconclusively from the outer”; it is “the apparent certainty of the first person, the uncertainty of the third” (LW I §951).
According to Wittgenstein, however, the inner–outer picture fundamentally misrepresents our relation to our own and to others’ minds. In the first place, he thinks, it is a mistake to think that the appeal to acquaintance does anything to explain a subject’s knowledge of her own sensations. He puts the point with characteristic economy:
‘How do you know that you have pains?’ – ‘Because I feel them’. But ‘I feel them’ means the same as ‘I have them’. Therefore this was no explanation at all. (BB 68)
His thought is this. The inner–outer picture depicts my knowledge of my own sensations as a form of perceptual knowledge: I know that I have a mole on my arm because I see the mole; analogously, it is supposed, I know that I’m in pain because I feel the pain. But the two cases are not analogous at all. When I see a mole on my arm, there are two things involved: the mole, and my perceiving the mole. So we can explain my knowledge of the mole by appeal to the fact that I perceive it. When I feel a pain, on the other hand, there are not two things involved: the pain, and my feeling the pain. On the contrary; feeling pain just is being in pain. So “I know I’m in pain because I feel pain” says no more than “I know I’m in pain because I’m in pain.” And that does not explain how I know that I’m in pain. It simply records the fact that I do.
Wittgenstein is equally dismissive of the idea that our knowledge of our own propositional attitudes – our beliefs, intentions, wishes, and so on – is a kind of observational knowledge. In the first place, he thinks, it is evidently not true that I have to introspect or observe myself in order to tell what I believe, what I intend, and so on. In normal circumstances, I can say what I believe or intend without considering any evidence at all: either “inner evidence” (feelings, images, occurrent thoughts, etc.) or “outer evidence” (my behavior). Furthermore, the observational model misrepresents our relation to our own attitudes, in a way that Wittgenstein brings out in connection with the point – noticed by G.E. Moore – that there is something paradoxical about propositions of the form “I believe it will rain but it won’t” or “It will rain but I don’t believe it will.” Such a proposition may be true. But one cannot coherently judge that it is true. What Moore’s paradox highlights is that, when someone judges “I believe it will rain,” she does not simply commit herself to a claim about herself and her beliefs; she also commits herself to a claim about the weather. But that would be hard to understand if self‐ascribing a belief were simply a matter of observing and reporting on the presence of an inner state. For why should the claim that I am in a certain inner state commit me to any particular claim about the world beyond my beliefs – in the way that the claim that I believe it will rain does commit me to the claim that it will rain? (For Wittgenstein’s discussion of this point, see PI II x.)
Wittgenstein, then, rejects the observational conception of our relation to our own minds that comes with the inner–outer picture. He also rejects the associated conception of our relation to others’ minds: in particular, the idea that one person can never really know what someone else thinks or feels. Of course, I cannot always tell what other people are thinking or feeling. But facts about one person’s mind, he thinks, are not as a class hidden from, and unknowable to, others: “It is only in particular cases that the inner is hidden from me; and in those cases it is not hidden because it is the inner” (LW II 33). Similarly, he insists, “My thoughts are not hidden from him, but are just open to him in a different way than they are to me” (LW II 34–5). And again: “If we are using the word ‘know’ as it is normally used (and how else are we to use it?), then other people very often know if I’m in pain” (PI §246). But Wittgenstein’s response to the claim that we cannot know what someone else is thinking or feeling is not simply to insist dogmatically on the opposite claim: that we can know it. He criticizes the reasoning that seems to support the inner–outer picture’s skeptical view.
It is tempting, Wittgenstein thinks, to reason in the following way: I cannot know whether S has a gold tooth, say, because I cannot see into her mouth; analogously, I cannot know whether someone else has toothache, because I cannot feel her sensations. (For this analogy, see BB 49.) But, Wittgenstein argues, the two cases are not analogous. In the case of the gold tooth, there is a kind of evidence about another person’s teeth that I could have but do not; the evidence I could acquire by looking into her mouth. So my inability to see into S’s mouth can make the difference between knowing that she has a gold tooth and not knowing that she does. The case of toothache is different. It is true that I cannot feel S’s toothache. But the reason is a conceptual or grammatical one: that any toothache I feel is my toothache, by virtue of the fact that it is me who is feeling it. For that reason, there can be no such thing as feeling someone else’s toothache. But in that case, the fact that I cannot feel S’s toothache does not mean that there is some deficiency in my evidence that S has toothache; it simply reflects the fact that the person with toothache is S, not me. And, Wittgenstein insists, there is nothing to prevent me from knowing that someone else has toothache. “If I see someone writhing in pain with evident cause, I do not think: all the same, his feelings are hidden from me” (PI II xi 223). And I am obviously right not to think that. In such a case – as in many others – I can know perfectly well what someone else is feeling, on the basis of what he says and does.
At the same time, Wittgenstein recognizes that there is an ineliminable or “constitutional” uncertainty (RPP II §657) in the relation between a person’s behavior, on the one hand, and the mental concepts we apply on the basis of that behavior, on the other. “The uncertainty whether someone else” is really in pain, say, or is really irritated, “is an (essential) trait of all these language‐games” (LW I §877). There is an “indeterminacy” in “the logic of the concept of pain” (LW II 94) and of other mental concepts. One dimension of this uncertainty or indeterminacy is that there are no “exact rules of evidence” for mental concepts: no exact rules for ascribing mental states to others on the basis of their circumstances and behavior (LW II 94). Similarly, the evidence on the basis of which we ascribe mental states to others includes “imponderable” evidence: evidence that cannot be weighed systematically against competing evidence (PI II xi 228).
This “constitutional uncertainty” in the relation between a person’s behavior and her thoughts and feelings, Wittgenstein thinks, is another source of the inner–outer picture. For the conception of mental states as states of an internal, mental mechanism that produces behavior seems to promise an explanation of the uncertainty: for example, that the internal mechanism is too complex for us to be able to infer a person’s inner mental states in detail from her behavior; or that the mechanism works indeterministically. But such explanations, according to Wittgenstein, get things back to front.
It is not the relationship of the inner to the outer that explains the uncertainty of the evidence, but rather the other way around – this relationship is only a picture‐like representation of this uncertainty. (LW II 68)
Wittgenstein’s idea is that the uncertainty in the relation between mental concepts and behavioral evidence is a primitive feature of the mental scheme. The idea that our behavior is produced by an internal mental mechanism whose operation is either indeterministic or impossible to systematize is just a picture: a reflection of this feature of common‐sense psychology. It is not an independent truth about the relations between the states of an inner mental mechanism and the behavior it produces. As before, he thinks, we treat the inner–outer picture as if it explained the central features of our practice. But the explanation it offers is illusory.
If we reject the inner–outer picture, how should we conceive of the mental, and of the language‐game of ascribing mental states to ourselves and others? Wittgenstein writes: “I look at this language‐game as autonomous. I merely want to describe it, or look at it, not justify it” (LW II 40). Common‐sense psychology, he thinks, is an autonomous, sui generis scheme of description and explanation. It is a scheme that we all use of ourselves and of others. And it does not stand in need of explanation or justification; we should simply accept it at face value, and understand it on its own terms.
But how, exactly, should we understand the common‐sense mental scheme? What are we doing when we say that someone is in pain, or is irritated, or intends to do such‐and‐such, if we are not talking about a system of inner mental states and events that causally produce her behavior? One answer to that question would be the simple and flat‐footed one. When we say that someone is in pain, we are saying that she has a sensation of a particular kind; when we say that she is irritated, we are describing her mood; and so on. And if our aim as philosophers really is just to describe the common‐sense scheme, the language‐game, perhaps there is no more to say than that.
However, Wittgenstein does in places offer a way of thinking of the common‐sense mental scheme that is not completely pleonastic: an alternative picture that, he thinks, does not misrepresent the phenomena. In Wittgenstein’s picture, life is portrayed as a “weave” or “tapestry” containing patterns that “vary in a multiplicity of ways” and are “interwoven” with each other (see Z §§568‐9). When we say that someone is irritated, or that he feels grief or joy, say, we are identifying patterns in the “tapestry of his life” (Lebensteppich) or the “weave of his life” (Band des Lebens). So, for example, Wittgenstein says that:
‘Grief’ describes a pattern which recurs, with different variations, in the tapestry of life. If a man’s bodily expression of sorrow and of joy alternated, say, with the ticking of a clock, here we would not have the characteristic course of the pattern of sorrow or of the pattern of joy. (PI II i 174)
And he describes pretending to be in pain as a “very special pattern in the weave of our lives” (PI II xi 229; cf. LW II 42).
If we see things in this way, we will still say that the way someone behaves is evidence of what she is feeling. But what it is evidence for is not, as in the inner–outer picture, the presence of an inner state that is hidden behind the behavior. It is, rather, the existence of a particular pattern in the person’s life.
Imagine it were really a case of patterns on a long ribbon.
The ribbon moves past me and now I say ‘this is the pattern S’, now ‘This is the pattern V’. Sometimes for a period of time I do not know which it is; sometimes I say at the end ‘It was neither’.
How could I be taught to recognize these patterns? I am shown simple examples, and then complicated ones of both kinds. It is almost the way I learn to distinguish the styles of two composers. (LW II 42–3)
We should think of learning to apply mental terms, Wittgenstein thinks, in a similar way: as a matter of pattern‐recognition.
The inner–outer picture encourages the view that our use of language in self‐ascriptions of current experiences and attitudes is analogous to its use in talking about the non‐mental world. Someone who says “I’ve got toothache,” “I want an apple,” “I believe it will rain,” and so on is, on this view, describing inner phenomena that she observes within herself; she is stating or reporting how things are in her inner realm. Wittgenstein rejects that view. He insists, as we have seen, that we do not normally observe our own thoughts and experiences. And in the normal case, he thinks, our first‐person, present‐tense psychological utterances are not descriptions; they are expressions or avowals of our experiences and attitudes. That idea has several dimensions and plays a number of roles in Wittgenstein’s discussion.
A first point of Wittgenstein’s insistence that mental self‐ascriptions are typically expressions is to draw a contrast between expressing an experience or attitude on the one hand, and describing or reporting it on the other. He thinks of describing as a definite activity, which involves observation and the assessment of evidence. But utterances like “I’ve got toothache” or “I hope he’ll come” normally require no observation or assessment of evidence at all. So, he thinks, it is wrong to regard them as descriptions. For the same reason, it is misleading to call them “statements.”
To call the expression of a sensation [e.g., the utterance “I’ve got toothache” or “That tickles”] a statement is misleading because ‘testing’, ‘justification’, ‘confirmation’, ‘reinforcement’ of the statement are connected with the word ‘statement’ in the language‐game. (Z §549)
And it is similarly misleading to treat avowals of sensation or attitudes as reports:
When someone says ‘I hope he’ll come’, is this a report about his state of mind, or a manifestation of his hope? – I may, for example, say it to myself. And surely I am not giving myself a report. (PI §585)
Of course there are some cases where first‐person, present‐tense psychological utterances really do involve a kind of observation and assessment of evidence. For example, we sometimes come to realize what we want or believe only by reflecting on our own behavior. And Wittgenstein describes various other cases in which it is appropriate to regard mental self‐ascriptions as descriptions. He says, for example, that in the normal case “the exclamation ‘I’m expecting him – I’m longing to see him!’ may be called an act of expecting”; it is not a description of oneself or one’s attitude.
But I can utter the same words as the result of self‐observation, and then they might amount to: ‘So, after all that has happened, I’m still expecting him with longing.’ (PI §586)
In this latter case I am describing myself. But such cases are the exception. In the normal case, I can say what I am experiencing, what I want, what I believe, etc., without the need for any self‐observation. And when I do so, Wittgenstein insists, I am expressing or giving voice to my experiences and attitudes; I am not describing or reporting them.
A second theme in Wittgenstein’s discussion of the expression of sensations and mental states concerns our acquisition of mental concepts. He writes:
How do words refer to sensations? – There doesn’t seem to be any problem here; don’t we talk about sensations every day, and give them names? This question is the same as: how does a human being learn the meaning of the names of sensations? – the word ‘pain’ for example. Here is one possibility: words are connected with the primitive, the natural, expressions of the sensation and used in their place. A child has hurt himself and he cries; and then adults talk to him and teach him exclamations and, later, sentences. They teach the child new pain‐behaviour.
‘So you are saying that the word “pain” really means crying?’ – On the contrary: the verbal expression of pain replaces crying and does not describe it. (PI §244)
Wittgenstein’s suggestion is that the word “pain” is learned as an addition to, or replacement of, our natural, pre‐linguistic expressions of pain. When we teach a child to utter the words “I’m in pain” in circumstances where she already displays non‐linguistic expressions of pain, we are teaching her to apply the word “pain” to herself in circumstances where she feels pain. That establishes a connection between her use of the word “pain” and her own feelings of pain. And crucially, for Wittgenstein, it establishes it in a way that does not depend on the kind of inner ostensive definition that is a central target of the private language sections of Philosophical Investigations. Of course, learning to say “I’m in pain” when one is in pain does not suffice for understanding the word “pain”; for that, one must learn the rest of its use, too, including its application to other people. But, Wittgenstein thinks, it is an important first step.
He makes a parallel proposal about our acquisition of the concept of wanting, or desire. There are natural, pre‐linguistic expressions of desire: reaching out for an object, for instance, or refusing to let go of it. And, Wittgenstein suggests, we learn the words “I want” in the first instance as an addition to, or replacement of, such natural, pre‐linguistic expressions. The child is taught to say “Apple!,” and then “I want an apple,” in circumstances where she already exhibits natural, pre‐linguistic expressions of wanting an apple. (For this idea, see e.g., LPP 23, 25, 141.) That shows how she can learn to apply the words “I want” to herself in appropriate circumstances, without supposing that there is any process of introspectively identifying an inner state of desire. As before, there is more to understanding the word “want” than learning to apply it to oneself in an appropriate way; one must master the third‐person use as well. But learning the first‐person use is evidently an important part of the whole.
The general idea, then, is that self‐ascriptions of experiences and attitudes are developments or replacements of more basic behavior that already expresses the experiences and attitudes in question. In the cases we have discussed so far, the first‐person utterances – “I’m in pain,” “I want an apple” – supplement or replace pre‐linguistic expressions of pain or desire. But Wittgenstein does not suggest that the same model can be applied in every case. Accounts of words for other sensations and other attitudes must respect the same principles: they must explain the meanings of mental terms in a way that does not appeal to inner ostensive definitions; they must not represent a subject as an observer of her experiences and attitudes; and so on. But within that framework there is room for significant differences. In particular, the primitive behavior from which the first‐person use of mental terms develops may already be linguistic behavior: a judgment, rather than a pre‐linguistic expression. Take the case of belief. Here, Wittgenstein thinks, we start with simple judgments like “It will rain.” When I judge “It will rain,” I make a judgment about the external world. But in making that judgment, I also indicate something about myself: that I believe it will rain. The judgment “It will rain,” then, is a verbal expression or manifestation of my belief. And, Wittgenstein suggests, I can learn the first‐person use of the word “believe” by learning to move from judgments that express my beliefs, like “It will rain,” to judgments that explicitly self‐ascribe those beliefs, like “I believe it will rain.” As before, that shows how I can learn to apply the word “believe” to myself, and simultaneously start to acquire the concept of belief, without supposing that there is any process of identifying inner states of belief within myself. (For Wittgenstein’s discussion of this point, see PI II x. That discussion has inspired a number of influential accounts of self‐knowledge; see particularly Evans, 1982, ch.7.4, and Moran, 2001.)
We said above that there is normally no room for error about one’s present experiences and attitudes. How should we understand that fact? In rejecting the inner–outer picture, Wittgenstein rejects one traditional explanation of the reliability or authority of a person’s judgments about her own experiences and attitudes: the idea that we have a faculty of inner perception that gives each person direct introspective access to the contents of her own mind. His own approach to the question builds on the idea that first‐person psychological utterances are expressions of the experiences and attitudes they ascribe.
That idea suggests a straightforward way of understanding the authority of self‐ascriptions in certain very simple cases. A natural, pre‐linguistic expression of pain – crying or wincing, say – is an automatic, unthinking response to the sensation of pain. When an infant cries in pain, she does not first identify her sensation as a pain and then decide to cry; she simply cries in response to the pain. Correspondingly, there is no possibility of her crying by mistake, because she has misidentified some other sensation as a pain. In a similar way, Wittgenstein suggests, the utterance “That hurts” or “I’m in pain” is in the simplest cases an immediate reaction to one’s pain, like flinching or crying out when one pricks one’s finger with a needle. These verbal expressions of pain are acquired reactions. But, like the pre‐linguistic reactions they supplement or replace, they are in these very simple cases an automatic, involuntary response to the experience. That is why there is no question of their being mistaken.
That is an important point. But even for the case of pain, it applies only in a very limited range of cases. For very few self‐ascriptions of pain are involuntary reactions to pain. When I tell the dentist about the character of the toothache I am feeling, my utterance is not an automatic, involuntary response; it is not like wincing when she touches an exposed nerve. So the authoritativeness or reliability of my judgment cannot be traced to the same feature that explains the reliability of the simplest pre‐linguistic expressions of pain: the existence of an automatic, involuntary association between the expression and the experience. The same is true for most other avowals of sensations; unlike simple pre‐linguistic expressions of sensation, they are not automatic reactions to the experiences they self‐ascribe. And similar points apply to self‐ascriptions of attitudes and emotions. In a limited range of cases, such utterances really are unthinking, involuntary expressions of the attitudes or emotions they self‐ascribe: the words “I don’t want you to go!,” cried by a small child to a departing parent, are a plausible example. But most self‐ascriptions of attitudes are not like that.
Wittgenstein, of course, acknowledges that. But he stresses that our self‐ascriptions of sensations and attitudes do quite generally resemble pre‐linguistic expressions in another respect; in the normal case, like pre‐linguistic expressions of sensation, they do not depend on observation or inference. That means that our mental self‐ascriptions are not vulnerable to two sources of error that affect our judgments about the external world and about others’ minds: misleading evidence; and mistakes in the inferences we draw from that evidence. Suppose I believe, on the basis of what someone says, that he intends to meet me at the pub in Kidlington. My belief may be false because I am wrong about what he said; he actually said that he would meet me in Kiddington, and that is what he intends to do. Or I may be right about what he said, but wrong to think he meant it; he said he would meet me in Kidlington, but he never intended to do so. But suppose I make the first‐person judgment that I intend to meet someone in Kidlington. In the normal case, I do not make that judgment on the basis of what I say, or any other evidence. So my judgment about my own intention cannot be wrong in either of those two ways.
That point reflects a crucial difference between psychological self‐ascriptions and other judgments. But it still does not explain the general reliability of our judgments about our own sensations and attitudes. For what the present point tells us is only this: that we have a general capacity to say what we are experiencing, what we want, what we believe, intend, and so on; and that these self‐ascriptions are typically made without observation and without reference to our behavior. In order for our self‐ascriptions to be true, however, they must cohere with the rest of our behavior. Merely judging that I believe that p, or want x, or intend to Φ, does not make it true that I do; the truth of such self‐ascriptions is answerable to my non‐linguistic behavior. And it is a striking fact that the self‐ascriptions we make without reference to any evidence generally do cohere with the rest of our behavior: that they are generally true. But it is natural to ask what explains that fact: how is that we are able to produce self‐ascriptions that reliably cohere with the rest of our behavior, without needing to rely on any evidence?
That is, indeed, a natural question. But, for Wittgenstein, it is a question that philosophy cannot – and need not – answer. If people could not learn to make groundless self‐ascriptions of attitudes that cohered with the rest of their behavior, he thinks, our ordinary mental concepts would have no application to them. For, he suggests, a word that people could not apply to themselves without reference to their behavior would not denote a kind of mental phenomenon at all. For instance:
One might distinguish between two chimpanzees with respect to the way in which they work, and say of the one that he is thinking and of the other that he is not.
But here of course we wouldn’t have the complete employment of ‘think’. The word would have reference to a mode of behaviour. Not until it finds its particular use in the first person does it acquire the meaning of mental activity. (RPP II §§229–30, emphasis added)
For Wittgenstein, then, the fact that we can generally ascribe sensations and attitudes to ourselves without evidence, in a way that coheres with our other behavior, is crucial; it is one of the “extremely general facts of nature” on which our concepts and our language‐games depend. (For the idea of such “extremely general facts of nature,” see the boxed comment associated with PI §142, and PI II xii 232.) But from the point of view of philosophy, at least, it is a basic fact: something we must simply accept as given.
What is the relation between our “inner” experiences and mental states and the “outer” behavior on the basis of which we ascribe them to one another? “The inner,” according to Wittgenstein, “is tied up with the outer not only empirically, but also logically” (LW II 63). “The connection of the inner and the outer,” he says, “is part of those concepts” (LW II 62). Those ideas run through his later philosophy of mind. And they are central to his rejection of the inner–outer picture. But what do they mean? And are they plausible?
On the conventional interpretation, Wittgenstein is making a metaphysical claim: a claim about what must be the case for creatures to have experiences and mental states at all. On this view, Wittgenstein takes it to be essential to pain, for instance, that in normal cases it has the kind of outer behavioral expression it does. He recognizes, of course, that there can be instances of pain without any behavioral manifestation. But such cases, he thinks, are necessarily the exception, not the rule. It is no more possible for there to be a world in which people experience pain but never manifest it in their behavior, or a world in which people have thoughts but never express their thoughts, than it is for there to be a world in which people play games but no one makes anything but false moves in any game. (For this analogy, see LPP 99 and PI §§344–5.)
The evidence for ascribing such a view to Wittgenstein includes passages like these:
“But doesn’t what you say amount to this: that there is no pain, for example, without pain‐behaviour?’ – It amounts to this: that only of a living human being and what resembles (behaves like) a living human being can one say: it has sensations; it sees; is blind; hears; is deaf; is conscious or unconscious. (PI §281)
Only of what behaves like a human being can one say that it has pains. (PI §283)
Wittgenstein talks here about the conditions under which one “can say that” a creature has pains, and so on. On the conventional interpretation, he does not mean merely that we can have no reason for saying that a creature is in pain unless it behaves like a living human being. He means that it does not make sense to say that something is in pain unless it behaves that way; and relatedly, that it is only possible for a thing to be in pain if it behaves like a living human being.
That might seem to rule out as unintelligible various possibilities that we appear to understand perfectly well: for instance, the possibility of someone suffering total paralysis but continuing to feel pain. But Wittgenstein’s position is more careful than that. He writes, for example:
I can perhaps even imagine (though it is not easy) that each of the people whom I see in the street is in frightful pain, but is adroitly concealing it. And it is important that I have to imagine adroit concealment here. That I do not simply say to myself: ‘Well, his mind is in pain: but what has that to do with his body?’ or ‘After all, it need not show in his body.’ (PI §391)
His point is that we can make sense of the thought that someone is in frightful pain but doesn’t show it; but in order to do so, we have to conceive of something preventing or inhibiting the normal behavioral manifestation of pain. Similarly, Wittgenstein imagines “a tribe” in which people are “brought up from early youth to give no expression of feeling of any kind” (RPP II §706), and seems to allow that we can coherently describe a situation in which adults never manifest their pain in any way at all. But it is an important part of this story that people exhibit the usual expressions of pain as infants, and must learn to suppress them as they grow up. So it remains true that the connection between feeling pain and the behavioral expression of pain is part of the nature of pain. Again, Wittgenstein allows that we can make sense of ascriptions of thought and sensation in circumstances where there can be no behavioral expression of thought or sensation, because the subject of our ascription is dying, or anaesthetized, or asleep. (For “dying,” see the remarks about Queen Victoria’s dying thoughts in RPP I §366 and LPP 32–3, 99, 152, 229, 274. For “anaesthetized” and “asleep,” see LW II 57.) But, he insists, such ascriptions make sense only against a background of other cases in which thoughts and sensations are manifest in behavior.
The lesson is that Wittgenstein’s view of the connection between thoughts and sensations, on the one hand, and behavior, on the other, is less straightforward and less open to obvious counterexamples than is sometimes supposed. Nonetheless, on the conventional interpretation, Wittgenstein still sees that connection as part of the metaphysics of mind. A world in which there was not, and never had been, any behavioral expression of thought or of pain would be a world in which there was no thought or pain.
But we might read Wittgenstein in a different way. For we could treat the dictum that the inner is “logically tied up with” the outer as a claim, not about the conditions for pain, say, to exist, but rather about the conditions for us to acquire and employ the concept of pain. On this alternative reading, Wittgenstein is not committed to the view that a world in which there was never any behavioral expression of pain would be a world in which there was no pain. His point is simpler and more mundane: that “if there were no characteristic expression of pain” we could not (or, more simply, would not) acquire and employ the concept of pain. In such a world, Wittgenstein says, “our normal language‐game” with the word “pain” would “lose its point” (see PI §142). But, on the alternative interpretation, he does not hold that the fact that we could not (or would not) talk or think about pain in such a world entails that pain could not exist in that world. He takes the fact that pain has a characteristic behavioral expression to play an essential role in fixing the reference of the term “pain”; but he does not think it is part of the meaning of the term that pain has that characteristic behavioral expression, or any characteristic expression at all. Nor does he think that pain is in its nature tied to any kind of behavior or behavioral disposition. (For readings of Wittgenstein along these lines, see the concluding “Postscript” in Albritton, 1968; and Koethe, 1996, ch.5.)
That is a coherent philosophical position. But is it a plausible interpretation of Wittgenstein? A first reason for doubt is this. The interpretation essentially depends on the distinction between the conditions under which it would be true that creatures have pain and the conditions under which the concept pain would have a use. In contemporary philosophy, that is standardly treated as an obvious and crucial distinction. But in Wittgenstein’s framework, in which a concept’s sense is conceived as being a matter of its use, it is not clear what could be made of the suggestion that an application of the concept pain might be true in circumstances where the concept had no use. A defender of the alternative interpretation might respond by calling on another standard contemporary distinction: the distinction between a concept’s having a use in a world and its having a use with respect to a world. It is true, she might say, that people who lived in a world in which there were no behavioral expressions of pain could not acquire, and would have no use for, our concept pain. Nonetheless we, who live in a world where there are behavioral expressions of pain, and who have acquired the concept of pain, can intelligibly use our concept with respect to that other world: to speculate, for example, about whether people in that world feel pain despite exhibiting no behavioral expressions of pain. And the thoughts we express when we do so speculate will be straightforwardly true or false. As before, that is in itself a reasonable philosophical view. But it is doubtful whether Wittgenstein would accept the distinction on which it relies: between a concept’s having a use in a world and its having a use with respect to that world.
Second, even if this alternative interpretation of Wittgenstein were correct for the particular case of pain, it would be hard to extend it to his treatments of other mental phenomena. Take intention. In Wittgenstein’s view, someone’s having intentions with definite contents depends on her mastery of, or participation in, relevant techniques, practices, customs, or institutions. And the existence of those practices, customs, and so forth requires the actual existence of repeated patterns of overt behavior. We are prone, he says, to think otherwise: to conceive intention as a wholly inner, mental phenomenon, whose existence does not depend on the existence of anything outer or behavioral. As he puts it:
“[…] what is remarkable about intention, about the mental process, [is] that the existence of a custom, of a technique, is not necessary to it. That, for example, it is imaginable that two people should play a game of chess, in a world in which otherwise no games existed – and then be interrupted.” (PI §205)
But, Wittgenstein insists, that is a mistake. For what would make these people’s intention an intention to play a game of chess? Suppose the people say, “Let’s play a game of chess.” What makes their words the expression of an intention to play chess? “Chess,” says Wittgenstein, “is the game it is in virtue of all its rules (and so on)” (PI §185). So, he asks, “where is the connection effected between the sense of the words ‘Let’s play a game of chess’ and all the rules of the game?” He answers: “Well, in the list of rules of the game, in the teaching of it, in the everyday practice of playing” (PI §185). If there were no such practice, there would be nothing to give these people’s intention a content about chess.
An intention is embedded in a setting, 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)
And for the technique of playing chess to exist, he thinks, people must actually play chess. On this view, there could not be a world in which people intended to play chess but no one ever did play chess. And more generally, that people regularly act on their intentions is not merely a necessary condition for our acquiring and applying the concept of intention; it is a necessary condition for the existence of intentions. For the case of intention, then, the alternative interpretation of Wittgenstein’s dictum that the inner is “logically tied up with” the outer seems untenable. The same goes for other mental states with intentional content: belief, desire, hope, and so on.
It does not follow from what has been said about intention that the alternative interpretation is wrong for every kind of mental phenomenon. The crucial point about intention is that having states with intentional content depends, for Wittgenstein, on mastery of, or participation in, practices. But he does not think the same is true for every mental phenomenon. For example, one need not have mastered a technique in order to have toothache (see PI II xi 208). As we have seen, however, there are other reasons for doubting the alternative interpretation, even in cases where the considerations about practices and techniques do not apply. So even if the alternative interpretation is consistent with some of what he says, and even if it has attractions as a view in its own right, it is implausible to think that it captures Wittgenstein’s own understanding of the relation between the inner and the outer. When he says that “the inner is tied up with the outer not only empirically, but also logically,” he is advancing a view not only about the conditions for acquiring and employing mental concepts, but also about the nature of the mental: about what it is to be a subject of sensations, thoughts, intentions, and the rest.