The “problem of other minds” goes like this. I have direct access to my own thoughts and experiences, but not to yours. I can perceive only your body and behavior. So how do I know you really have any thoughts and experiences? Maybe you merely behave as if you had them, but in reality you are a “zombie” in the philosophy of mind sense, devoid of conscious awareness. And maybe this is true of everyone other than me. How do I know that any minds at all exist other than my own?
One traditional answer to the problem is the “argument from analogy.” I know from my own case that when, for example, I flinch and cry out upon injury, there is a sensation of pain associated with this behavioral reaction, and that when I say things like “It’s going to rain” that’s because I have the thought that it is going to rain. So, by analogy, I can infer from the fact that you also flinch and cry out, and also say things like “It’s going to rain,” that you too must have thoughts and sensations of pain. In modern philosophy this argument is put forward by writers like John Stuart Mill, and it is sometimes suggested that one can find something like it in Augustine’s On the Trinity, Book 8, Chapter 6.
The standard objection to this argument is that it amounts to the weakest possible kind of inductive inference, a generalization from a single instance to every member of a class. There are eight billion people, and I have observed in only one case, namely my own, a correlation between thoughts and experiences on the one hand and bodily properties and behavior on the other. So how does the inference from my own case to the rest of the human race not amount to a fallacy of hasty generalization?
But the idea that there is a “problem” here, and that the solution to it is a kind of inductive generalization, are artifacts of modern philosophical assumptions. I don’t think Augustine was in fact giving an “argument from analogy” in the sense in which modern philosophers have done. He was not offering a philosophical theory in response to a philosophical problem. He was just noting how, in his view, we do in fact in everyday life know that other minds exist. Here is the relevant passage:
For we recognize the movements of bodies also, by which we perceive that others live besides ourselves, from the resemblance of ourselves; since we also so move our body in living as we observe those bodies to be moved. For even when a living body is moved, there is no way opened to our eyes to see the mind, a thing which cannot be seen by the eyes; but we perceive something to be contained in that bulk, such as is contained in ourselves, so as to move in like manner our own bulk, which is the life and the soul. Neither is this, as it were, the property of human foresight and reason, since brute animals also perceive that not only they themselves live, but also other brute animals interchangeably, and the one the other, and that we ourselves do so. Neither do they see our souls, save from the movements of the body, and that immediately and most easily by some natural agreement. Therefore we both know the mind of any one from our own, and believe also from our own of him whom we do not know. For not only do we perceive that there is a mind, but we can also know what a mind is, by reflecting upon our own: for we have a mind.
End quote. Note that Augustine attributes to non-human animals the same kind of knowledge that other things are alive (and, presumably, conscious too) that he attributes to us. But non-human animals do not engage in inductive reasoning. So, it isn’t an inductive generalization, fallacious or otherwise, that he is attributing to us either.
I submit that there are at least three assumptions more typical of modern philosophy than of ancient and medieval philosophy that underlie the idea that there is such a thing as a “problem of other minds,” and that our knowledge of other minds is grounded in a kind of inductive generalization.
The first is that genuine knowledge is always a kind of “knowing that” or propositional knowledge, as opposed to a kind of “knowing how” or tacit knowledge. (See Aristotle’s Revenge, pp. 95-113 for discussion of this distinction.) Not all modern philosophers take this view, but it is criticized by thinkers like Wittgenstein, Heidegger, Michael Polanyi, Hubert Dreyfus, et al. as typical of a Cartesian conception of rationality. The idea is that knowledge can always be analyzed into the possession of explicitly formulated propositions and inferences. If ordinary people don’t seem consciously to be entertaining such propositions and inferences, then it is sometimes claimed in response that the propositions and inferences are nevertheless present below the level of consciousness (say, as the rules and representations of a computer program implemented in the brain).
It is natural, on this model of rationality, to think that knowledge of other minds must involve some kind of inference. Wittgenstein’s critique of the whole debate over the problem of other minds might be read as claiming that knowledge of other minds is in fact a kind of tacit knowledge rather than a propositional or inferential kind of knowledge. On this view, the debate simply misunderstands the nature of our knowledge in this case.
A second relevant modern assumption is that there is a metaphysical gap between matter and consciousness that makes inferring the presence of the latter from facts about the former inherently problematic. The standard modern conception of matter inherited from the early modern “mechanical philosophy” facilitates this assumption. Matter is taken to be characterized by quantifiable primary qualities alone (size, shape, motion, etc.) and lacking anything like qualitative secondary qualities (color, sound, heat, cold, etc.), at least as common sense understands them. Hence it becomes mysterious how the qualia of conscious experience could be material, and knowledge of a thing’s material properties comes to seem insufficient to ground an inference to its having any mental properties. (This is, of course, a topic about which I’ve written many times.)
A third relevant assumption – and the one I want to focus on here – involves what we might call a Baconian conception of our knowledge of the natures of things, as opposed to an Aristotelian conception. For the Aristotelian-Scholastic position against which Francis Bacon reacted, common sense is more or less right about the natures of everyday natural objects, even if it isn’t very sophisticated. The ordinary observer correctly grasps what it is to be a stone, a tree, or a dog, even if it takes scientific investigation to understand these natures in a deeper and more sophisticated way. On this view, the senses are, you might say, friendly witnesses and just need an expert to ask the right questions in order to find out what else they know.
For Bacon, by contrast, the real natures of things are hiding behind false appearances, and sense experience is a hostile witness who must be tricked and threatened into revealing what it knows. Hence Bacon’s emphasis on slow and painstaking observations under artificial conditions, and the gradual working up from them to a general conclusion before one can claim to know what a thing of a certain kind is really like.
Also relevant is the modern idea that understanding the physical world is not a matter of uncovering the essences of things, but rather a matter of identifying the laws that relate observed phenomena. This involves formulating a general theory, deriving predictions from it, and then testing those predictions by way of a series of observations and experiments.
If that is the way that knowledge of the empirical world works, then naturally it comes to seem that whether other people really have minds is not as clear-cut as common sense supposes, and some kind of theoretical inference must be deployed in order to justify the supposition that they do. The meagerness of the evidential base in one’s own case in turn becomes a major problem, insufficient as it is for the construction of a Baconian inference.
Now, as contemporary neo-Aristotelian philosophers of science point out, this is not exactly how science in fact typically proceeds. To be sure, the Baconian emphasis on making observations under artificial experimental conditions certainly does feature in modern scientific method, but the idea of piling up instances before drawing a general conclusion does not. As Nancy Cartwright has argued, once the right experimental conditions are set up, multiple observations are not seen as necessary. She writes: “Modern experimental physics looks at the world under precisely controlled or highly contrived circumstance; and in the best of cases, one look is enough. That, I claim, is just how one looks for [Aristotelian] natures” (The Dappled World, p. 102, emphasis added).
Now, if few observations or even a single observation are all that is needed when the circumstances are right, then we have an essentially Aristotelian rather than Baconian approach to how many cases are needed in order to determine the basic nature of a thing. The difference is in the nature of the cases, not the numberof cases. The Aristotelian thinks that ordinary observation isn’t especially liable to get the nature of a thing positively wrong (even if it does not go very deep either) whereas the Baconian thinks that ordinary observation’s getting things positively wrong is a serious possibility.
Now, suppose the Baconian were correct about observation of things other than ourselves – stones, trees, dogs, etc. Suppose that common sense is indeed prone to getting the natures of these things wrong (which is, again, a stronger claim than just saying that common sense has only a superficial knowledge of their natures, which the Aristotelian would not deny).
Still, it wouldn’t follow that we are likely to get things wrong about our own nature, and prima facie that is unlikely, since in this case the knower and the thing known are the same. Of course, one could argue that there is a dramatic appearance/reality gap even here, but the point is that you would have to argue for such a claim. It is not prima facie what we would expect.
In that case, though, why couldn’t one know just from one’s own case that a physiology and behavior like ours go hand-in-hand with a psychology like ours as a matter of human nature? And then, applying this knowledge of human nature, why couldn’t one thereby know that other human beings too have mind’s like one’s own? “Problem of other minds” solved.
Again, some philosophers would of course argue that things are not as they seem even in the case of our apparent knowledge of our own minds. But the point is that that is simply not a plausible default assumption. There is a presumption in favor of our having a correct understanding of our own nature, and (given what I’ve just argued) there is, accordingly, a presumption in favor of other people having minds just like our own.
That suffices to undermine the idea that there is a frightfully difficult “problem of other minds.” The correct description of our situation is not that we don’thave good grounds for believing in other minds, and therefore have to cobble together some solution to this problem. It’s that we do have good grounds for believing that there are other minds, and therefore the philosopher who thinks otherwise can make things seem problematic only by making a number of tendentious modern (and, I would say, wrong) philosophical assumptions.