The question of how all of us know that there are other beings besides themselves who have thoughts, feelings, and other mental attributes has been widely discussed, especially among analytic philosophers in the English-speaking world. At least three of the most influential German philosophers—namely, Edmund Husserl, Max Scheler, and Martin Heidegger—have also dealt with this problem. The problem of "other minds" becomes a serious and difficult one because the traditional and most obvious solution to it, the argument from analogy, is open to grave objections. At the present time it would seem that a majority of the philosophers who have concerned themselves with the question consider the traditional solution—that our belief in other minds can be adequately justified by an analogical argument—at least inadequate, if not radically and unremediably defective.
Argument from Analogy
In general terms to argue by analogy is to argue on the principle that if a given phenomenon A has been found to be associated with another phenomenon B, then any phenomenon similar to A is very likely to be associated with a phenomenon similar to B. In the particular case of other minds, it is said, I observe that there is an association between my mental states, on the one hand, and my behavior and the physical state of my body, on the other. I then notice that there are other bodies similar to mine and that they exhibit behavior similar to my own. I am justified, therefore, in concluding by analogy that mental states like the ones I experience are associated with those other bodies in the same way that my mental states are associated with my body. I notice, for example, that when I have a pain in my tooth, it is likely to be decayed and that I am likely to groan, complain, and hold my jaw. Observing another body like my own that has a decayed tooth and behaves as my body behaves when I have a toothache, I conclude that this body, like mine, is the body of a being that has a toothache.
objections to the analogy argument
The first and least radical objection to the argument from analogy is that it does not establish its conclusion with an adequate degree of certainty. The argument, it is said, would be relatively strong if the correlation of the mental and the physical was observed to hold in a large and varied collection of instances before it was concluded that it also held in other similar cases. But this is not so. If I use the argument from analogy, I have only one case, my own, as a basis for my inference. Moreover, the characteristics and behavior of the other bodies vary markedly from my own. How can I be sure that the differences between myself and others are not associated with the presence of mental attributes in my own case and with the absence of them in other cases?
The other difficulties in the argument from analogy concern two features of that argument—first, that it is logically impossible to check up on the correctness of the conclusion of the argument and, second, that the argument's validity implies that one must learn from one's own case alone what it is to have a mental attribute. Let us elaborate a little on each of these points.
In the case of a normal analogical argument, it makes good sense to suppose that one might check up directly on the conclusion of the argument; in principle one could always dispense with reasoning by analogy, even though this may not be practicable in some cases. Of course, one who says that we know of the existence of other minds by analogy must deny that we can check up on our conclusion in some more direct way, for if we could, the argument by analogy with ourselves could be dispensed with. It also seems that he cannot say that our inability to check up is merely a practical matter. Such checking up cannot consist in making further observations of a person's behavior and body; this we can often do sufficiently well in practice. It would have to consist in some other operation that we cannot in fact perform but which we can conceive of ourselves performing; perhaps it would be something like telepathy.
But aside from any difficulty in making clear sense of the notion of telepathy, why should telepathy be regarded as a more direct way of checking up than ordinary observation of behavior? Indeed, it seems that one's grounds for thinking that one has telepathic knowledge of another person's state of mind must include the knowledge that what one seemed to know telepathically generally correlates well with what one knows as a result of ordinary observation. The same would also seem to apply to any other extraordinary but conceivable way of knowing about another's mental state. Granted, then, that the supporter of the argument from analogy must hold that the impossibility of checking the conclusion more directly is not any variety of empirical impossibility, why is this held to destroy the argument? Perhaps there is a difference here between this argument and other valid analogical arguments, but why does this difference make this argument unacceptable? The answer given is that this difference renders the conclusion of the argument senseless. What can the phrase "He is in pain" mean to me if no conceivable observation I could make would show that it was true or false, if I have no criterion for its truth, and if I have no idea of what would count for or against it? It will not do to say that the sentence means that he has the same as I have when I am in pain, for, again, what counts as being the same here?
The other main difficulty in the analogical argument centers, as we have said, on the necessity, implied by that argument, for each of us to learn from his own case alone what it is to have a mental attribute. Two arguments have been advanced to show that this is impossible.
According to the first, which derives from Ludwig Wittgenstein, the analogical argument requires that one be able to pick out something (for example, a pain or a state of anger) and thereafter to identify it, when it recurs, as a pain or a state of anger. The trouble is, however, that this account leaves no room for a distinction between a correct and an incorrect identification. Behavioral and other checks are ruled out, leaving no conceivable means of deciding whether a mistake has been made. But a distinction between a correct and a mistaken identification is surely essential to the very notion of identification itself. In this way the analogical argument, which requires that we be able to make correct identifications of our inner states, also deprives the notion of identification of any meaning.
The second argument, which has been advanced by P. F. Strawson, is more complex. According to him, the idea of a predicate involves the idea of a range of individuals to which that predicate can be significantly applied. In the case of mental attributes, this range includes both oneself and others; one cannot have the notion of a mental attribute unless one has a notion of oneself and a notion of another. Since the notion of oneself is the notion of a subject of mental and other attributes, one cannot have the notion of oneself without the notion of some mental attributes. Therefore, one cannot have a notion of oneself without also having the notion of another subject of mental attributes. This notion, however, can be possessed only if one knows how to ascribe mental attributes to such subjects. Hence, until one knows how to do this, one has no notion either of oneself or of another. But the argument from analogy requires that one should first have a notion of oneself, of one's own case, and then discover how to ascribe mental attributes to others by arguing analogically from correlations that are found to hold in one's own case. A person without a notion of his own case could indeed argue analogically. He could find that pain was to be expected when a certain body (his own, as we say) was branded with a hot iron. He could infer that there would also be a pain when another similar body was similarly affected. But he would soon find out that he was mistaken in this conclusion, for he would detect no pain when the hot iron was applied to any body other than his own.
defenses of analogy argument
Some persistent attempts (especially by A. J. Ayer) have been made to defend the argument from analogy against the charges laid against it. To counter the charge of weakness, the following suggestions have been made. Emphasis has been laid upon the special feature of the argument from analogy—that people can speak and that their descriptions of their mental states are very like those I would give of some of my own. This, it is claimed, is something more telling than a mere similarity of behavior. Against this it is pointed out that speech can be regarded as something understood by the speaker only if it is accompanied by the appropriate nonverbal behavior.
Another defense is that conclusions drawn analogically from behavioristic similarities are powerfully reinforced by like conclusions drawn by arguments based on similarities in the state of the nervous system. This consideration hardly meets the main complaint—namely, that I base my inference on one case only, my own.
According to a rather more convincing attempt to meet this complaint, no more can be asked of any method of inference than that I be able to test its conclusion more directly in some cases and that when I do so, the conclusion usually turns out to be correct. The argument from analogy satisfies this test. I can suppose that there are, as there seem to be, other people besides myself and that these people argue analogically that I have certain thoughts and feelings. I can check on these imagined inferences and find that their conclusions are generally true. Whether these inferences are in fact made is neither here nor there; I can see that the method would work if it were used. Nor need I be worried because I can check only those cases in which the conclusion is about myself. In all or most inferences there will be a restricted class of cases that I can check up on. It is, for instance, logically impossible that I should make a direct check on a change of color that occurred where I could not observe it. But it would be a mistake to argue that any analogical argument that a color change had occurred was weak because it was based upon one sort of case only—the sort that I was able to observe. Why should it make a difference to the strength of the other minds argument that the relevant class of case is my own mental states as opposed to what I myself observe?
An argument similar to this one can also be used to rebut the charge that there is no conceivable means of checking up on the conclusion of the argument from analogy. There are in fact some cases in which I can make a check—namely, those cases that concern myself. Moreover, although it is logically impossible for me to be some other person and hence to make a direct check on that other person's mental states, this is unimportant, for it is never logically impossible that I should check on the truth of a psychological statement when the subject is referred to by a descriptive phrase, even though that description fits someone other than myself. It is logically impossible, perhaps, that I should be Robinson, but it is not logically impossible that I should now be the man flying a certain aircraft, even though Robinson is in fact that man. Moreover, it is claimed, when I make a statement about Robinson, what is stated is, in effect, that someone who answers to such and such a description has had such and such an experience. To this it has been objected that the only interpretation of this claim that yields the desired conclusion is untrue, namely, the interpretation that "Robinson has a pain" means the same thing as some sentence of the form "The so and so has a pain." However, this objection clearly fails to settle the matter, as can be seen by considering the following statements:
- The man sitting in this chair is angry.
- Robinson is the man sitting in this chair.
- Robinson is angry.
Statement (1) cannot be said to be unintelligible to me on the ground that I, not being the man in question, cannot check up directly, for it is conceivable that I might have been sitting in the chair; statement (2) can also be checked on by me; statement (3) follows from (1) and (2). It is surely quite implausible to hold that statement (3) is unintelligible to me, whereas statements (1) and (2) are not.
There is, however, another possible difficulty in the argument from analogy that is usually not at all clearly distinguished from the one just considered—namely, that it is in principle impossible for more than one person to check directly on the conclusion. It is often said that publicity is the essential requirement. But does this mean that it must be logically possible for each person to make the check, or is it the more stringent requirement that it be possible for everyone, or at least more than one, to do so? If the latter, then the difficulty has not been overcome. Equally it has not been shown clearly why publicity should be required in the more, rather than in the less, stringent form.
This brings us to the reasons given for holding that one cannot understand psychological predicates from one's own case alone, which is a requirement of the argument from analogy. One of these reasons, as we have seen, is that there is no sense in the idea of an identification that is subject to no check, where there is no criterion of correctness. This view has been questioned on two grounds. Strawson has argued that a criterion of correctness is not needed in all cases of identification, and according to Ayer, an identification of a sensation can be satisfactorily checked, without recourse to anything publicly observable, by means of other private sensations.
Other Solutions to the Problem
Assuming that the argument from analogy is unacceptable, the most obvious alternative is to adopt some form of that variety of behaviorism according to which all psychological expressions can be fully understood in terms of behavior. If behaviorism is correct, there is clearly no room or need for the argument from analogy. In ascribing a pain to someone, for example, one is asserting something that is in principle subject to a public check—something about the way the individual is behaving, about how he would behave in certain circumstances, about what the circumstances in fact are, or the like. There is no need to make any inference from the publicly observable to something radically different.
This is not the place for a general discussion of behaviorism. Any objection to a given form of behaviorism will, of course, be an objection to that form of behaviorism as a solution to the problem of other minds. There is, however, one difficulty that has given rise to a number of closely related attempts to deal with the problem—namely, that it is implausible to give a behavioristic account of some first-person psychological statements. When, for example, I say that I have a terrible pain, I do not say this on the basis of observation of my own behavior and the circumstances in which I am placed. Nor am I speculating about how I would behave in other, hypothetical circumstances.
This difficulty has become of central importance for many philosophers who are impressed by some or all of the arguments that purport to refute the argument from analogy. They regard such arguments as showing, not only that this argument fails, but, more positively, that the connection between mental states, on the one hand, and behavior and circumstances, on the other, is logical or conceptual, not contingent. What is needed to remove the difficulty about our knowledge of other minds, it is thought, is to clear away the obstacles that prevent us from seeing clearly that this connection is a conceptual one. The primary obstacle in this instance is the peculiar nature of first-person psychological statements. It is this obstacle that prevents us from wholeheartedly accepting the true view and that makes us always hark back to the picture of mental states as objects to which the owner has privileged access.
There are at least two points involved here. First, if my own statements about my mental states are not about private happenings to which only I have access and if they are not about my behavior either, then what account is to be given of them? Second, the statement "I am in pain," made by me, contradicts the statement "He is not in pain," made about me by someone else. If one admits that the former is not about my behavior, how can one avoid the conclusion that the latter also is not about my behavior? But if the latter is not about my behavior, how can it be maintained that the connection between my pain and my behavior is a logical one?
In dealing with the question "How do words refer to sensations?" Wittgenstein suggested, "Here is one possibility: words are connected with the primitive, the natural, expressions of sensation and used in their place" (Philosophical Investigations, Sec. 244). This suggestion, which is not elaborated much by Wittgenstein, has sometimes been treated as an attempt to deal with the first point stated above and has had certain merits ascribed to it—for example, by Norman Malcolm. It explains how the utterance of a first-person psychological statement can have importance for us; such an utterance has the importance that natural expressions of sensation and emotion have. It is also said to explain certain features of the logic of psychological statements, the absurdity of someone's concluding that he has a pain from the observation of his own behavior, and the impossibility of someone's being mistaken about whether he has a pain or of wondering whether he has a pain. However, whatever its merits, this stress on the likeness of first-person sensation statements to natural expressions of emotion and sensation merely sharpens the second of the difficulties noted above—namely, that "I am in pain" can contradict "He is not in pain." It even makes it hard to see how the former can be a statement at all; a cry of pain is not a statement.
This difficulty is obviously insuperable for one who, unlike Wittgenstein, adopts the extreme position that apart from being verbal and learned responses, first-person sensation statements are exactly like natural expressions of sensation. Wittgenstein, however, appears to hold that a statement like "My leg hurts" is never in all respects like a cry of pain but is sometimes more like it and sometimes less, depending on the context of utterance. There seem to be three main likenesses that he wanted to stress in all first-person present-tense expressions of sensation and in many such expressions of emotion—namely, (1) the impossibility of these expressions being mistakenly uttered; (2) the possibility of their being insincere or pretended; and (3) the fact that such statements can justifiably be made without a basis of self-observation. The problem that arises in formulating a successful defense of his views is showing how a statement that bears the above likenesses to a cry of pain can yet be different enough to contradict another statement for which the criteria of truth lie in the realm of the publicly observable—that is, in the behavior of the speaker.
It cannot be said that Wittgenstein himself made a serious attempt to cope with this difficulty. Others have made the attempt, but no attempt has been very convincing. The second and third points of likeness present no great difficulty (see Douglas Gasking, "Avowals"). Any statement can be made insincerely, and there are many nonautobiographical statements that a person can justifiably make without observing that the criteria for their truth are satisfied. For example, some people can tell you that a certain note is middle C without first carrying out the tests that determine whether it has the appropriate frequency. For such statements to be justified, it is necessary only that those who make them are usually right in such cases.
The first difficulty, which arises from the alleged incorrigibility (as it is termed) of first-person present-tense statements, is not so easily disposed of. The most hopeful approach—indeed, the only approach—is to exploit the fact that the natural expressions of sensation and emotion can be feigned. An insincere groan is akin to a lie, and a lie is a false statement. Perhaps a verbal expression can reasonably be called false if it is insincere and true if it is sincere, the distinction between sincerity and insincerity being a matter of the behavior of the speaker. In this way a plausible account could be given of how something very like a groan could also in some ways be like a statement and be regarded as such. The incorrigibility of such statements would then be accounted for.
But this is not enough; it does not explain how such a "statement" can be the contradictory of another statement that is logically connected with statements about the behavior of the maker of the "statement." For (1) "I have pain," said by me about myself, is the contradictory of (2) "I have not a pain," said by me about myself. Therefore, since (3) "He has a pain," said about me by someone else, is also the contradictory of (2), (1) and (3) must both be the same statement. Consequently, if (3) is logically connected with certain behavioral statements, (1) must also have these connections. This makes it difficult to see how (1) can be incorrigible. If I can be mistaken about my own behavior, as is the case, and if there is a logical connection between my pain and my behavior, then, it would seem, I can be mistaken about my pain. This difficulty is not overcome by assimilating the truth of a first-person pain statement to the sincerity of a groan. For (4) "I am sincere in saying I have a pain," said by me about myself, is the same statement as (5) "He is sincere in saying he has a pain," said about me by someone else. Therefore, if (5) is logically connected with statements about my behavior, so is (4), and, if (4) is so connected, it must, it seems, be corrigible. For to claim sincerely that p is to think that p when one makes the claim, and to claim insincerely that p is to think that not-p when one makes the claim. If (4) is corrigible, then someone might think he is sincere in claiming he has a pain when in fact he is insincere—that is to say, he might think that he thinks that he has a pain, although in fact he thinks that he has not a pain. If, however, one cannot be mistaken about one's own pain, then to think that one thinks one has a pain is to think one has a pain, and to think one has not a pain is not to have a pain. It follows that if (4) is corrigible, someone might think that he has a pain although, in fact, he has not a pain. In short, if (4) is corrigible and (1) is not, then (1) is corrigible.
There are apparently only two ways out of these difficulties that do not involve abandoning the thesis of the incorrigibility of first-person psychological statements and thus ceasing to attach much value to the assimilation of such statements to natural expressions of emotion and sensation. One might deny that (1) and (3) are the same statement, or one might maintain that although (1) is logically connected with behavioral statements about which I can be mistaken, yet I cannot be mistaken about (1). The first of these alternatives would involve finding a satisfactory explanation of why I cannot assert the same thing that someone else does when he asserts (3). The second would require an account of the notion of a logical connection that would allow for the existence of statements that, when made by myself, are incorrigible, but which are logically connected with other statements that, when made by myself, are not incorrigible.
In fact it has been argued by some that there are no psychological statements that are incorrigible and that the problem we have just been discussing is therefore an unnecessary one. It seems to be quite true that there are some ways in which one can be mistaken when one says one has, say, a pain. But the matter has not yet been clarified sufficiently for anyone to be justified in saying with confidence that this renders the problem unnecessary. Even if first-person present-tense pain statements are corrigible, this does not show that they are corrigible in all the ways that other statements are corrigible. Nor has it been shown convincingly that they are corrigible in such a way as to obviate any difficulty that may arise from the fact that "I have a pain," said by me, contradicts "He has a pain," said about me.
In addition to the above objections to Wittgenstein's views on the subject of psychological statements, there is another one that is of a less definite character and to which Wittgenstein himself alludes when he puts into the mouth of an imaginary objector such words as "and yet you again and again reach the conclusion that the sensation itself is a nothing " (Philosophical Investigations, Sec. 304). He protests, of course, that this is not the sort of impression he wishes to create and that it arises from his "setting his face against the picture of the inner process." Nevertheless, it cannot be said that he altogether succeeds in dispelling this impression. His problem might indeed be described in just these terms—to set his face against the inner process picture without creating the impression that he wishes to deny the existence of sensations. It does not seem that he succeeds in this.
p. f. strawson
It is perhaps Wittgenstein's failure that in part gives rise to another attack on the problem—namely, that of P. F. Strawson. Strawson, like Wittgenstein, is convinced that the argument from analogy is mistaken and that skepticism about other minds is senseless or at least empty and pointless. Like Wittgenstein, he holds that the relation of the behavior of other people to their mental states is not contingent: "the behavior-criteria one goes on [in assigning P-predicates—that is, psychological predicates] are not just signs of the presence of what is meant by the P-predicate, but are criteria of a logically adequate kind for the ascription of the P-predicate" (Individuals, p. 106).
In spite of this he is out of sympathy with Wittgenstein in many ways. He considers that the assimilation of first-person present-tense psychological statements to the natural expressions of sensation and emotion "obscures the facts and is needless" (Individuals, p. 107). He is unconvinced by Wittgenstein's reasoning against the idea of a private language that might serve as a basis for the argument from analogy. He sees little difficulty in the notion of a person's inventing for himself a private language in which he has names for his sensations even when such sensations have no outward expressions: "He might simply be struck by the recurrence of a certain sensation and get into the habit of making a certain mark in a different place every time it occurred" (Individuals, p. 85). Nor does he consider the notion of a person's continuing to exist in a disembodied state as logically absurd (Individuals, pp. 115–116). He accuses Wittgenstein of hostility to the idea of what is not observed and of a "a prejudice against the inner" ("Critical Notice," p. 91).
All these criticisms of Wittgenstein suggest that Strawson holds the view that the connection between behavior and mental states is, after all, a contingent one. But this, as we have seen, is not so. How, then, does Strawson reconcile these apparently conflicting aspects of his thought? His line of thought appears to be approximately that general agreement in judgment is necessary before it is possible to have a common language. Such general agreement exists about, for example, "what it looks like here," and this agreement makes possible our common impersonal language of, for example, color. There is no such general agreement about "whether or not 'it's painful here,' " and there is thus no possibility of a common impersonal pain language. However, there is something available (namely, pain behavior) on which general agreement is possible, and if we are therefore to have a common pain language, we must each ascribe pain to others on the basis of their behavior. In this way a common personal language becomes possible.
In discussing Strawson's thought, it is crucial to emphasize that until a person decides to ascribe pains to others on the basis of their behavior, he has not got and cannot have our concept of pain, for part of that concept is that a pain is something that someone possesses. Nevertheless, he can have a concept (or perhaps something more rudimentary than a full-fledged concept) that is akin to our concept of pain but does not involve the idea of something that is had or possessed by either himself or others.
Perhaps this can be made more intelligible by considering a conceivable though unlikely case, that of a young child who has not yet got our concept of pain but is on the way to getting it. When he falls and knocks his head or scrapes his knee, he says, "It hurts." He has learned this sentence, perhaps as a replacement for natural cries of pain, and he uses it to get picked up and otherwise comforted. However, when his twin brother or a brick falls off the table, and the child is asked, "Does it hurt?" he replies, "No." Nevertheless, he cannot be said to mean by "It hurts" what is meant by "It hurts me," even though he says the former only when the latter is true, for he attaches no sense to "Does it hurt John?," as opposed to "Does it hurt me?" Nor, with regard to what he calls hurting, does he see any difference between John and a brick. If John says, "It hurts," when he himself is feeling all right, he regards what John says as simply untrue. In order for this child to make the transition to the concept of pain as something that either he or someone else has, he must learn to say, "It hurts John," when John bumps his head and cries and to say, "It hurts me," when formerly he said only, "It hurts." Until this linguistic convention is acquired, the child cannot be said to have the concept of pain as a property of persons at all, not even as a property of himself.
Thus, the argument from analogy breaks down because it assumes not only that a person can have a private language but that this language contains our concept of pain (ascribed pain). But such a language could contain at best only a concept of what we may call unascribed pain. The connection between unascribed pain and my behavior is a contingent one, but the connection between behavior and ascribed pain is not. We can see now why Strawson says, "I have argued that such a … 'justification' [of our beliefs about others] is impossible, that the demand for it cannot be coherently stated" (Individuals, p. 112). To talk about other people's pains at all is to accept and use the concept of ascribed pain, and it is an integral part of this concept that behavior shows any person whether that concept applies to other people.
Criticisms of Strawson
Strawson's views are open to some of the criticisms that have been directed against opinions that are the same as his own. In addition, Ayer has directed a number of criticisms specifically against Strawson's positions, asserting that his notion of logical adequacy is obscure and arguing that this obscurity is irremediable. It is certainly true that Strawson does not make the notion of logical adequacy as clear as he might, but Ayer's reasons for thinking that this obscurity could not be remedied are themselves inconclusive. Ayer's other main criticism is directed against Strawson's reason for holding that neither the argument from analogy nor the philosophical skepticism that arises from this argument can be stated coherently. This criticism is based on a failure properly to understand Strawson's position, which in turn leads to the mistaken idea that Strawson cannot allow for the existence of someone with the concept of a person "who was invariably mistaken in ascribing states of consciousness to others" (The Concept of a Person and Other Essays, p. 106).
There is nothing in Strawson's position to prevent him from holding that analogy is used in the ascription of states of consciousness to others; the only thing that he rules out is analogical argument of the traditional pattern. To understand this, let us use the words "upain" and "utickle" for the concepts of unascribed pains and tickles. According to Strawson, in order to pass from these concepts to those of (ascribed) pains and tickles, I must adopt verbal rules according to which I say "I have a pain" when there is a "upain" and "He has a pain" when another body exhibits certain behavior, and so on. But what sort of behavior, and so forth? There is no reason that Strawson's answer should not be along some such lines as "behavior, etc., that is like the behavior, etc., that this body (i.e., mine) exhibits when there is a upain." In accepting such a rule, I am not arguing by analogy. Now, I can adopt such a rule and thus have the concept of a person, but I can still fail to realize that all the objects I regard as persons are in fact unlike myself in ways that I have not noticed.
Ayer describes an imaginary child who is brought up and taught to speak by lifelike robots and who never meets real people. He argues, quite correctly, that this child would have the concept of a person and yet always be mistaken when he ascribes mental attributes to anything. But no consequences fatal to Strawson's views follow from this. The child has adopted the verbal rule whose acceptance, according to Strawson, is necessary for the possession of the concept of a person. The child mistakenly thinks that the robots are persons because he believes that they are much more like himself than in fact they are. This gives no ground for the skeptical conclusion that I may here and now be mistaken in my belief that there are other people besides myself. If one accepts Strawson's position, such skepticism need be justified only if what I think to be other people are a great deal less like me in behavior, etc., than I take them to be. If there is a doubt left here, according to Strawson it can have nothing very specifically to do with other minds. The basis of Ayer's misunderstanding is his mistaken belief that Strawson "infers that any attempt to justify the belief that there are other persons by relying on the premiss that one knows oneself to be a person would be circular; the premiss would already assume what the argument is supposed to prove" (ibid., p. 104). But Strawson's objection to the argument from analogy is not that it is a circular argument. According to him, the trouble is that the argument both uses the concept of a person and rejects the verbal rule that is a necessary part of that concept, namely, the rule that mental attributes are to be ascribed to things on the basis of their behavior, and so on.
john wisdom's views
Finally, something should be said of John Wisdom's very important work on this problem. It is quite impossible to summarize Wisdom's contribution as another solution to the problem of other minds. This impossibility is inherent in his views about philosophy and in the method he used in conformity with these views. All that can be done here is to give some idea of what is to be found in his writings on the problem of other minds by sketching his method of dealing with it.
Wisdom was much influenced by Wittgenstein, especially in regard to the idea that the treatment of a philosophical problem is in some ways like the treatment of an illness. Such a problem or puzzle is a symptom of deep-seated intellectual disorder that consists in a persistent tendency to think about a certain area of thought and language in accordance with a misleading and partially inappropriate model. The puzzle is dissipated when one is "cured" of this tendency. Inattention, however, is not the only remedy, nor is the taking of drugs. The only "cure" available to a philosopher qua philosopher is a certain form of insight. The misleading model that distorts one's thinking is largely an unconscious one. Insight and freedom from its grip are obtained by bringing it into the open, by making quite clear in detail how our thought is governed by it, and by giving us a proper view of the nature of, for example, our knowledge of other minds.
Thus, Wisdom's first aim is to induce and sharpen philosophical perplexity by showing how it arises precisely out of the sort of position that is at first sight the most attractive to us. For example, the most natural answer to the question about other minds is the traditional one. But it is from this answer and the way of thinking that goes with it that philosophical skepticism most easily arises. Skepticism is satisfactorily removed only when we are brought to see that knowing about other minds is not altogether like other ways of knowing that are by analogy and that it need not be. It might be thought that the aim of a philosopher should be to find a correct model that does not mislead. But according to Wisdom, this is not so. Although every statement has its own logic, the logic of every statement is in some degree like that of every other. We cannot usefully create a limited set of pigeonholes into one of which goes our knowledge of other minds along with, say, our knowledge of the past, while our knowledge of any theoretical entity goes into another. The matter cannot come to this sort of a conclusion. There will be important differences that will make inappropriate any such pigeonhole, as well as the likenesses that make it possible. To get a true grasp of the nature of our knowledge of other minds, it is necessary to make a very large number of detailed comparisons between the various ways in which we know or might know things and between the logic of various types of statements. Only then will we see psychological statements and the ways in which we know of the existence of other people's thoughts and feelings in all their idiosyncrasies and in all their similarities to other statements and to other ways of knowing things. Until this is done, we cannot be entirely freed from our tendency to see things as they are not.
As may be deduced, Wisdom's writings about other minds are almost as much about induction, the past, perception, philosophy of science, and so on as they are about other minds. He used his method with subtlety, inventiveness, and imagination. Many points made by later writers on the problem of other minds are little more than elaborations or oversimplifications of points already made by Wisdom.
See also Private Language Problem.
Ayer, A. J. The Concept of a Person and Other Essays. London: Macmillan, 1963. Ch. 4. Criticizes Strawson's views.
Ayer, A. J. The Problem of Knowledge. London: Macmillan, 1956. Ch. 5. Defends the argument from analogy.
Bilgrami, Akeel. "Dummett, Realism, and Other Minds." In The Philosophy of Michael Dummett, edited by Brian McGuinness. Dordrecht: Kluwer, 1994.
Brewer, Bill. "Emotion and Other Minds." In Understanding Emotions: Mind and Morals, edited by Peter Goldie. Brookfield, U.K.: Ashgate, 2002.
Buford, T. O., ed. Essays on Other Minds. Chicago: University of Illinois Press, 1970.
Dennett, Daniel C. "Beliefs about Beliefs." Behavioral and Brain Sciences 1 (1978): 568–570.
Dretske, Fred. "Animal Minds." Philosophic Exchange 31 (2000–2001): 21–33.
Fodor, Jerry. "A Theory of the Child's Theory of Mind." Cognition 44 (1992): 283–296.
Gasking, Douglas. "Avowals." In Analytical Philosophy, edited by R. J. Butler, 154–169. Oxford, 1962.
Ginet, Carl. "Plantinga and the Philosophy of Mind." In Alvin Plantinga, edited by James E. Tomberlin. Dordrecht: Reidel, 1985.
Goldman, A. "In Defense of the Simulation Theory." Mind and Language 7 (1992): 104–119.
Heal, Jane. "Other Minds, Rationality, and Analogy." Proceedings of the Aristotelian Society, supplement 74 (2000): 1–19.
Hill, Christopher. "On Getting to Know Others." Philosophical Topics 13 (1985): 257–266.
Leslie, A. M. "Pretense and Representation: The Origins of 'Theory of Mind.'" Psychological Review 94 (1987): 412–426.
Malcolm, Norman. "Knowledge of Other Minds." Journal of Philosophy 55 (1958): 969–978. A radical criticism of the argument from analogy and the traditional viewpoint.
Malcolm, Norman. "Wittgenstein's Philosophical Investigations." Philosophical Review 63 (4) (1954): 530–559. Defends Wittgenstein's views on private languages.
McDowell, John. "Criteria, Defeasibility, and Knowledge." Proceedings of the British Academy 68 (1982): 455–479.
McGinn, Colin. "What Is the Problem of Other Minds?" Proceedings of the Aristotelian Society, supplement 58 (1984): 119–137.
Mill, J. S. An Examination of Sir William Hamilton's Philosophy. London: Longmans Green, 1865. Ch. 12. Contains a straightforward version of the argument from analogy.
Nichols, Shaun. "Mindreading and the Cognitive Architecture underlying Altruistic Motivation." Mind and Language 16 (2001): 425–455.
Sober, Elliot. "Evolution and the Problem of Other Minds." Journal of Philosophy 97 (2000): 365–386.
Strawson, P. F. "Critical Notice of Wittgenstein's Philosophical Investigations." Mind 63 (249) (1954): 70–99. A criticism of Wittgenstein's views on private languages.
Strawson, P. F. Individuals. London: Methuen, 1959. Ch. 3.
Wisdom, John. "Other Minds." Mind 49 (196): 369–402; 50 (197): 1–22; 50 (198): 97–122; 50 (199): 209–242; 50 (200) 313–329; 51 (201): 1–18 (1940–1942).
Wisdom, John. "Other Minds." Logic and Reality, PAS, Supp. 20 (1946): 122–147.
Wisdom, John. Other Minds. Oxford: Blackwell, 1965. Reprints all the above.
Wittgenstein, Ludwig. Philosophical Investigations. Oxford: Blackwell, 1953.
Husserl, Edmund. Cartesianische Meditationen. Husserliana, Vol. I. The Hague, 1950. Meditation V. Translated by Dorion Cairns as Cartesian Meditations; An Introduction to Phenomenology. The Hague, 1960.
Scheler, Max Ferdinand. Zur Phänomenologie und Theorie der Sympathiegefühle und von Liebe und Hass. Halle, 1913. Appendix.
J. M. Shorter (1967)
Bibliography updated by Benjamin Fiedor (2005)