Our Knowledge of Other Minds

Our Knowledge of Other Minds

How do we know the contents of other people’s minds? By what method do we know about other minds? The options are the following: by sense perception, by inference, by introspection, and by a priori reasoning (I exclude telepathy, at least as the standard method). We can rule out the last two: I do not introspect your state of mind, only my own; and I do not know your mental state as I know mathematical truths, i.e. without recourse to experience. So there appear to be two possible theories: by direct perception and by indirect inference. Both views have been maintained. The perceptual view assimilates knowledge of other minds to our knowledge of the world of material things: just as I know by looking that there is a tree in front of me, so I know by looking that you are in pain or are deep in thought. I simply see that you are in pain, as I see that there is a tree yonder. My eyes are directed towards your body and your body expresses your pain—makes it manifest to my senses. Such a view goes naturally with philosophical behaviorism, since states of mind are identified with episodes of behavior (or dispositions to behavior). By contrast, the inferential view denies that we can perceive other people’s mental states, supposing instead that they are to be conceived as theoretical entities, posited to explain observed behavior, rather like atoms and fields of force. Then our knowledge of other minds is like our knowledge of the unobservable things that we postulate to make sense of what we observe. We know other minds by means of something like inference to the best explanation. This view goes naturally with a causal conception of mental states: they are unobservable causes of observable effects (episodes of behavior). Again, there is nothing special or distinctive about our knowledge of other minds—we know about them as we know about unobservable material entities. I see your behavior and I infer that it is caused by mental states of specific kinds, as I see meter readings and infer that atoms are the cause. Thus, if there is a skeptical problem concerning other minds, it is not peculiar to the case of other minds. It is just a special case of a more general skeptical problem—the problem of perceptual error and the problem of knowledge of unobservable entities, respectively. The other person might give a misleading perceptual impression of being in pain and not be in pain, or he might not in fact harbor the theoretical entity I posit to explain his behavior. This kind of assimilation doesn’t solve the skeptical problem, but it does give it a name: it tells us what kind of problem we are up against. The problem of our knowledge of other minds is essentially the same as the problem of our knowledge of the external world.

            Neither of these familiar views is satisfactory: we don’t see pains in people as we see colors and shapes in objects, and we don’t we infer pains as we infer atoms or fields. But I won’t go into why; instead I will offer an alternative. We should not attempt to reduce our knowledge of other minds to some other paradigm of knowledge. We need, rather, to identify what is distinctive of knowledge of other minds—what sets it apart. And what sets such knowledge apart is the role of self-knowledge in generating knowledge of others. Neither perceptual knowledge nor inferential knowledge (of the standard scientific sort) rests upon a basis of self-knowledge, but knowledge of other minds does. I know other minds by knowing my own mind. So I will contend.

            Self-knowledge includes a number of different things. Suppose that I am now thinking about playing tennis: what kind of knowledge do I have about that mental state? I know (a) that I am now thinking about playing tennis, (b) what this thought consists in, and (c) what kinds of things this thought inclines me to do. The first kind of knowledge is episodic (what is happening in my mind now); the second is constitutive (what it is that I am undergoing); and the third is dispositional (what dispositions to action I have in virtue of having the thought in question). Focusing on (c), we can say that for any mental state M that I know I have I also know the dispositions associated with M—I know what I might do in virtue of having M. If I know that I am in pain, say, then I know that I might complain about the pain or wince or cry out. How I arrive at this first-person dispositional knowledge is an interesting question: do I know it by observing correlations between my inner mental states and my outer bodily behavior, or do I know it innately by having the knowledge programmed into my genes, or is it in some sense a type of conceptual knowledge? I won’t go into the question; I will just assume that we have such knowledge in our own case. So we have knowledge, not just of what occurs inside of us mentally, but also knowledge of how this inner thing might be expressed publicly—we know our dispositions to behavior. I know that I’m in pain now, I know what pain is, and I know what pain makes me do–I know the bodily expression of pain. This dispositional knowledge is part of my self-knowledge. I could have it whether or not I knew anything about other minds.

            Then the thesis is that we use such self-knowledge in acquiring knowledge of other minds. The way we obtain knowledge of other minds is quite straightforward (which is not to say simple): we first observe another person behaving in a certain way; then we note that when we behave that way we are in a particular mental state; we then attribute that mental state to the other. So we put together two premises: (1) that the other person is behaving thus and so, and (2) that in our own case dispositions to that kind of behavior go with our having a certain mental state. We then conclude that the person has the mental state in question. That is, we generalize from our own case. Our only basis for the attribution we make is that in our own case a certain association obtains. We know it in our own case and we assume it for the case of others. If we did not have this kind of self-knowledge, then we would not be able to have knowledge of other minds. We rely on our self-knowledge to generate other-knowledge. This is not how it works according to the other two theories, in which self-knowledge plays no essential role in our knowledge of other minds—as it doesn’t in perceptual knowledge and knowledge by scientific inference.

            It is important to see what this account is meant to achieve and what it is not meant to achieve. It is meant to explain what is special about our knowledge of other minds—how such knowledge differs from other types of knowledge. We don’t reason in this way when forming beliefs about the external world: we don’t use our self-knowledge as a premise, assuming that what is true in our case must be true more generally. The reason is simply that in other cases we are not trying to gain knowledge of other minds, so we don’t use any premises comparing our mind with other minds. In perceptual and inferential knowledge (of the scientific kind) we don’t make any comparison between the object of such knowledge and ourselves. We don’t presume that ordinary material objects have a mind like ours! (Nor do we assume that other bodies resemble our own—we just have a look at those bodies.) We only do this kind of thing when aiming to acquire knowledge of other minds. So the present account sets knowledge of other minds apart from the run of empirical knowledge. What the theory is not intended to do is provide a reply to skepticism. In fact, it underscores skepticism about other minds, because this method of knowing is vulnerable to skepticism on several fronts. First, we are generalizing wildly from our own case, which is just one case among many, and there is a question whether we are entitled to draw such far-reaching conclusions from so slender a basis (it looks like a rash type of induction). Second, it may be that the association that holds in our case between mental state and behavior does not hold in the case of others—maybe the same disposition is associated with a different mental state (as in inverted spectrum cases). Third, how do we know that there is an association even in our own case? Maybe we have no body at all or make mistakes about how behavior and mental states match up. There is plenty of room here for the skeptic to stick her oar in. But replying to the skeptic was not the intention; indeed, I might even say that doing justice to skepticism about our knowledge of other minds is part of the point of the theory. According to the perceptual and inferential theories, we just get a version of the usual kinds of skepticism that afflict perceptual judgments and inference to the best explanation: but there is surely something special—and especially troubling—about other minds skepticism. We really are on shaky epistemic ground here—which is why mistakes about other minds are so frequently made. And other minds are surely hiddenin a unique way, not merely as atoms are hidden. We are compelled to reason from what is true in our own case to what is true for others—a risky move. We have only one mind to go on, our own, and we are forced to rely on it to provide knowledge of indefinitely many other minds. This is not an epistemologically happy situation; but it is the way things are and the way they must be. The world, after all, does not owe us an epistemological living. We are lucky we have the self-centered method I have sketched—we might have had no method at all.

            It is possible to imagine a being with no self-knowledge who nevertheless has both perceptual and inferential knowledge of the external world, but we cannot (if I am right) imagine a being that has knowledge of other minds but does not have knowledge of its own mind. Given that some minds are not self-aware in the manner sketched, such minds cannot attribute minds to others (this will be true of many animals). But if a creature can form the idea of other minds, then that creature must have self-knowledge: cats and dogs (not to mention apes) would seem to qualify as mind-readers, so they must know their own minds in some way—they must be projecting their knowledge of their own minds into the minds of others. Thoughts of other minds require thoughts of one’s own mind—and many animals give every sign of knowledge of other minds. Animals know how their own minds link to their behavior: a dog knows what growling means in its own case, and it extrapolates this knowledge to other dogs.

            It might be said that the self-knowledge theory is itself an inferential theory, because it involves inferring that others are like myself. That is a perfectly correct observation, but it is a special kind of inferential theory, quite unlike normal scientific inference, as noted above. Nor is it a case of inference to the best explanation: it is inference based on observed similarity of behavior, combined with self-knowledge of certain mental-behavioral links. It would not be wrong to describe it as a version of the “argument from analogy”, because it proceeds by noting a similarity between oneself and others and then attributing a further similarity: the other is similar to me behaviorally, so she must be similar to me mentally. The point I am adding is just that self-knowledge plays an indispensable role: first I must know how my own mental states are connected with my behavior; only then can I attribute mental states to others. This is what is characteristic of knowledge of other minds: knowledge of other minds depends on prior knowledge of one’s own mind, i.e. knowledge of how one’s own mental states are expressed bodily. In a certain sense, knowledge of other minds is more “self-centered” than other kinds of knowledge. Whenever I am thinking of the minds of others I am implicitly thinking of my own mind, because I have no other basis for such knowledge than knowledge of my own mind. Introspection is essential to knowledge of other people’s minds, though not because I introspect their minds. There is no escaping my own mind in my thinking about other minds. My mind is the model I use to build a picture of other minds. (If this sounds like a truism to some, or as naively pre-Wittgenstein to others, then I am encouraged.)

            There is a strange ambiguity or ambivalence in our knowledge of other minds: at one moment it can seem like the most immediate knowledge in the world, but at the next moment it can seem impossibly remote. Sometimes we think we know just what is on the other person’s mind, but then we reflect that this is really quite hidden from us. Perhaps this oscillation stems from the dual basis of this kind of knowledge: on the one hand, I do know immediately what my mental states are and what they incline me to do; on the other hand, it seems rash to suppose that others are the same way. If I focus on the self-knowledge premise, I seem to be on solid ground (for Iam certainly inclined to yelp when feeling a sharp pain); but if I focus on the analogy-with-others premise, then I seem to be overreaching (maybe others don’t yelp when they are in pain). The fact is that I am partly thinking of other minds as my mind in another body and partly recognizing that other people are genuinely alien subjects. I am torn in this way because the very nature of my knowledge of other minds reflects both (infallible) self-knowledge and (fallible) other-extrapolation. Our entire conception of other minds is frankly a kind of confused amalgam, caused by the epistemological necessity to combine facts about one’s own mind with facts about other minds. I observe the bodily expressions of others, then I reflect on my own mind and its relation to my body, and then I go back and attribute a mental state to the other. Nothing like that happens in other areas of knowledge. Our knowledge of other minds rests on a roundabout method of trying to overcome the fundamental fact that our own mind is the only one we can introspect. It would be so much easier if we could just peer directly into the minds of others, but the metaphysics of mind rules that out; so we have to do the best we can with our self-knowledge combined with some risky extrapolation. We have to work with the epistemic materials we have, feeble and fallible as they may be. That is just the way it is with our knowledge of other minds—we are condemned to play epistemological catch-up. We are trying to overcome a structural problem written deep into the nature of things. On the positive side, we necessarily unite ourselves with others in a common psychological family—we have no choice. Alien minds are unknowable minds, given the way we obtain knowledge of minds other than our own. If another mind broke all the rules of mental-behavioral association that hold in our own case, we could not know its contents. We are trapped in an epistemic corner created by our own psychophysical nature. We seek to resolve the mystery of other minds by reducing them to our own mind.  [1]

  [1] Surely our primordial attitude towards other minds is that they are a complete mystery, an arena of inextinguishable ignorance. Yet we must make inroads into this darkness because we are a social species; so we make stabs in the dark using knowledge of our own mind as our weapon. The result is hardly complete illumination, more like forlorn speculation. We hope we get things right, but we can never be sure.