This is an essay I wrote back in 2009, and there’s a sense in which my having written it may have been futile. I refer not to the disdain which the general public may so sadly direct at the noble art of the philosophy essay, but to the possibility that I am the only person capable of reading it –not because of my idiosyncratic and altogether inept use of language– but because I may, for all I know, be the only sentient person in the universe. We have all considered this possibility at some point, but because it is so absurd it is easy enough to dismiss. The philosophical problem of demonstrating why we are right to dismiss it is, however, far from easy.
This ‘problem of other minds’ is very similar to the radical (and highly general) sceptical hypothesis of solipsism, the view that my mind is the only thing that exists. Because of this, the two terms are used equivalently below, along with the expression ‘different minds’ to denote the more general issue of ‘minds significantly different to how we ordinarily believe them to be, possibly without consciousness.’
I will begin by considering some conventional responses to the problem in the philosophy of mind. But I ultimately contend that because the problem of other minds is a fundamentally epistemological issue,  its character is such that it cannot be adequately resolved by any theory of the nature of the mind, but requires an epistemological solution. A contextualist theory of justification (as articulated by Michael Williams) is thus elaborated in order to provide (without requiring us to refute Wittgenstein’s insights) the best solution that we have.
First we shall consider the ‘solution’ from analogy in its classical rendering by John Stuart Mill. Mill accepted the Cartesian view that we are aware of our own minds via direct acquaintance with them, but because we have no such acquaintance with the minds of others –but only their (presumably corresponding) behaviour– if we are to be justified in believing that others have a mental life like our own we must make a generalising inference from our case to theirs:
I conclude that other human beings have feelings like me, because, first, they have bodies like me, which I know, in my own case to be the antecedent condition of feelings; and because, secondly, they exhibit the acts and other outward signs, which in my own case I know to be caused by feelings.
Mill maintains of his generalisation that “in doing so I conform to the legitimate rules of experimental enquiry.” This is highly questionable, as Ludwig Wittgenstein points out: “If I say of myself that it is only from my own case that I know what the word ‘pain’ means – must I not say the same of other people too? And how can I generalize the one case so irresponsibly?” This point is decisive in refuting Mill’s argument from analogy for it is clearly the case that the ‘legitimate rules of experimental enquiry’ do not allow for a conclusion of this strength and scope (as it predicates this of an entire numerous species) to be inferred from the observation of just one case.
Wittgenstein had initially shown sympathy to the intentions of solipsism; he neither contested nor accepted the thesis, but in his distinctive style, clarified the issue. He claimed that there was truth in the notion that the world is my world, because the limit of intelligibility is boundary of the world, and this limit is set by my language, which I alone fully understand. He maintained that this truth could only be shown, never said; that is to say that it could be indirectly implied, but not be meaningfully expressed. Wittgenstein qualifies of this way of understanding solipsism saying that its implications are consistent with a direct realist understanding of reality because, while our metaphysical self is ‘the self of solipsism’, it remains the case that there is an external reality which corresponds to our experiences.
Wittgenstein later dealt with the problem of other minds more directly. In his Philosophical Investigations he used the example of pain to illustrate his notion of different minds: I can know that others are in pain because they exhibit learned pain behaviour on the same basis that I do, but I cannot know that they feel pain to anything like a comparable extent.8 To even make the assertion that we might know how others feel pain lacks coherence because the notion of pain cannot be divorced from the concept of conscious sensation. To imagine the pain of another we need to imagine a “pain which I do not feel on the model of a pain I do feel.”
He claimed that whenever we define an ‘inner state’ such as pain we have not truly identified anything, because this is contingent on the reliability of our memory which is not independently justified.11 He says that such a definition is like “a wheel that can be turned though nothing else moves with it is not part of the mechanism.”12 The preceding ‘ascribability’ and ‘idle wheel’ arguments support Wittgenstein’s claim that mental states such as pains can only be known via public behaviour; John Cottingham summarises this claim as follows:
In order to have meaning in our language, terms for mental states must be employed on the basis of public criteria, or rules for their correct application; and if this is so, then it cannot be right that expressions such as ‘I am in pain’ get their meaning from referring to a private event accessible only to the subject.
It can be plausibly asserted that Wittgenstein’s conclusion is a behaviourist one. He does not strictly reduce mental states to dispositions to behave in particular ways, but does contend that we can have no knowledge of our own mental experiences. Moreover, a behaviourist solution could not be adequate because it could not preclude the existence of mental states distinct from behavioural tendencies. Such a solution would be too quick and equivocates the standard concept of mental properties to one that identifies them with behaviour.
Furthermore, Wittgenstein’s move has implications even more radical than solipsism: we do not even have reliable experience of our own mental life and therefore a fortiori, nor do we have it of those of others- since we wouldn’t be sure what it is to have a mind. But we are sure of what it feels like to have a mind thus by modus tollens, behaviour can’t be necessary to pain itself, but only to the common usage that we give to the term ‘pain’ in our language games. Therefore e can’t be more certain of the sensations of others than we can of our own, and therefore scepticism stands. So while Wittgenstein’s arguments are fairly plausible there is a desire to mitigate the forms of scepticism they leave open.
The possibility of a more adequate solution springing from physicalism may seem promising because there strong intuition in our scientific world-view that minds are supervenient (subject to a one-way relationship of causal dependency) on brains. It looks like physicalism would resolve the problem of other minds, and if integrated with property dualism would also provide us with a means to predicate qualia to others’ minds. But physicalism does not, and could not solve the problem of other minds because it only provides an explanation of how others have minds (i.e. because their minds are dependent upon part of their physical bodies), and not a justificatory reason for believing that others have minds.
This is because physicalism is a hypothesis and we do not yet have all that much evidence for correlation between physical states and mental states, let alone proof with which to verify the hypothesis. Whichever theory of the metaphysics of mind we hold we cannot negate the possibility of others not having minds, and therefore to make progress we require an epistemological justification. We can now see that approaching the problem with mainstream theories of the metaphysics of mind is fundamentally flawed due to the epistemological character of the solipsist hypothesis, this is such that an epistemological justification is required. I turn now to consider the possibility of such a justification.
Contextualism is a theory of epistemological justification in which the conditions for attributing knowledge are relative to context; they have circumstantial variation. A similar sentiment is evident in the early Wittgenstein’s contention that what truth there is to solipsism cannot be meaningfully expressed. Contextualism addresses the concern that we need to somehow separate the radical sceptical scenarios such as solipsism from the demands of everyday knowledge, finding a way to deny that the possibility of such scenarios needs to be completely ruled out in order for knowledge to be possible. Contextualists maintain both that such everyday knowledge (e.g. perceptual knowledge) is justified and that we cannot rule out the possibility of sceptical scenarios being true; the crucial point is that these possibilities are considered in different contexts. Within an appropriate context our default entitlements to knowledge are liable to be challenged, whereas in other, inappropriate contexts, questioning them will prove such a question to be irrelevant or even senseless,  as Williams illustrates:
To be intelligible at all –and not just to be reasonable – questioning may need a lot of stage-setting…. this is true of the sceptic’s attempt to call into question our most ordinary and obvious judgements about the world around us.
Thus, there is a sense in which radical theses such as solipsism are incoherent. This is because in order for us to think through the stages of their argument, to understand them, we have to accept rather than challenge their premises and/or ‘stage-setting criteria.’ Our acceptance of these can be construed as inconsistent, if not with their form of scepticism, then with a similar program of hyperbolic doubt that yields the same result.
Furthermore, the absurdity of the possibility of solipsism is such that the metaphysical framework required for it to be the case (e.g. a god willing to create universe in which the only sentience among a large population of similarly intelligent beings is mine), is not only more unlikely than it being the case that others do have similar minds, but so much so that it [solipsism] is untenable for the sceptic. For the sceptical position is surely of the greatest sensitivity to such a ludicrous hypothesis.
This is an example of exercising what Williams calls ‘intelligibility constraints’ on challenges to our knowledge, which is one method of contextualism. This has much overlap with methodological factors, those concerning the logic of inquiry. These are constraints on what we can question without undermining beliefs necessary for the particular inquiry at hand, for wider procedures, and for our lives more generally.
This leads on to the consideration of economic factors: possible defeaters for our knowledge claims must be rational; they must have a value construable in a cost-benefit analysis, a reason why they might obtain. Solipsism, however, works not on reasons, but on possibility. “If a challenger implies that we might be making a mistake, we are entitled to ask how. If the challenger has nothing to say… then no real challenge has been entered.”
By upholding contextualism we can therefore perceive that the possibility of solipsism is irrelevant due to: its unintelligibility, its lack of value to the debate (as a thesis that we have no reason to adopt). This is just as well, because for practical reasons even if solipsism were still true we would still have to act the same, not simply because we could not be certain of its truth, but because we are part of a system of interactions with others in which we have particular needs and concerns and these would not be satisfied if we behaved otherwise. We could not, for example, behave as if we were the only sentient being, not just because we almost certainly are not, but because people behave in a sentient manner and we could not live treating them as if they did not. The potential truth of solipsism is out of context, indeed, irrelevant, so long as we are integrated into this system of interactions and the responses that result from our actions with it. These are just the sort of considerations that would otherwise incline us towards a physicalist solution, and very similar to the conclusions that Wittgenstein comes as a result of his view of the embedding of linguistic meaning in practice. But unlike the those solutions contextualism gives us a solid reason with which we can support our conviction.
In summary, the argument from an analogy is quite implausible as a solution, even if it is the method by which we imagine the mental lives of others most of the time. Wittgenstein’s considerations, as their influence shows, are highly valuable, but cannot solve the problem with the necessary scope. This can, however, be achieved by the contextualist theory of justified knowledge, which allows us to mitigate the relevance of scenarios such as solipsism as possible defeaters to our standard knowledge. Yet it must be remembered that we have focused on minds significantly different to how we ordinarily believe them to be, and as Wittgenstein demonstrates so effectively in his idle wheel argument, there is much potential for variance between the mental experiences of individuals. Thus, even if we do not have any reason to think that such variances would be radical- for others to have mental lives unrecognisable to our own, they may well vary enough to account for some of the mistakes we make when trying to understand others. The pessimist may be forgiven for saying that every human being is born alone and dies alone, for we each do this alongside each other, in both mind and body.
Conee, Earl & Cohen, Stewart, ‘Is Knowledge Contextual?’, In Contemporary Debates In Epistemology, Steup and Sosa (eds.), (2005), Oxford, Blackwell Publishing Ltd.
Maslin, K. T., , (2007), An Introduction To the Philosophy of Mind, 2nd Edition, Cambridge, Polity Press
Mill, John Stuart, , ‘An Examination of Sir William Hamilton’s Philosophy,’ In Western Philosophy: An Anthology, Cottingham, John (ed.), (1996), Oxford, Blackwell Publishing Ltd.
Williams, Michael, (2001), Problems of Knowledge, Oxford, Oxford University Press
Wittgenstein, Ludwig, , (2001), Tractatus Logico-Philosophicus, McGuiness, B. F., & Pears, D. F. (trans.), 3rd Edition, Abingdon (UK), Routledge
Wittgenstein, Ludwig, , (2001), Philosophical Investigations, Anscombe, G. E. M. (trans.), 3rd Edition, Oxford, Blackwell Publishing Ltd.
1 Maslin, (2007), p. 215
2 Cottingham (ed.), (1996), p. 165
3 Mill, , p. 167
4 Mill, , p. 167
5 Wittgenstein, , §293
6 Wittgenstein, , 5.6 – 5.62
7 Wittgenstein, , 5.64
8 Wittgenstein, , §246
9 Wittgenstein, , §246
10 Wittgenstein, , §302
11 Wittgenstein, , §258
12 Wittgenstein, , §271
13 Cottingham (ed.), (1996), p. 165. Cf. Wittgenstein, , §§246, 293, & 303
14 Maslin, , p. 234
15 Maslin, , p. 217
16 Williams, (2001), p. 159
17 Conee & Cohen, (2005), p. 59
18 Williams, (2001), p. 159
19 Williams, (2001), p. 160
20 Williams, (2001), p. 160
21 Williams, (2001), p. 169