The Problem of Other Minds, Old and New

Sanford C. Goldberg

Northwestern University

How do you know that there are minds other than your own? This is the standard “Problem of Other Minds.” It is a problem (and not just a question) if we make two further assumptions, both of which can be traced to Descartes. The first is the assumption that each of us has mental access, not to the world and its inhabitants themselves, but to exclusively one’s own thoughts or ideas about the world and its inhabitants. The second is the assumption that when it comes to others’ minds, our thoughts or ideas are always directed towards others’ overt behaviors, from which we infer that others have minds. The standard “Problem of Other Minds” is the problem of justifying the transition, from the claim that one has thoughts or ideas about other people (with minds of their own), to the claim that other people (with minds of their own) actually exist.

The “Problem of Other Minds” persisted in more or less this form throughout much of the Twentieth Century. It is taken up in Russell’s Problems of Philosophy (1912) and in Ayer’s Problems of Knowledge (1956), both of which have been reprinted as recently as the 1970s and 80s. And when I was a graduate student in the 1990s, it was familiar to have the “Problem of Other Minds” (in roughly its Cartesian form) still on the syllabus for epistemology courses.

Though I am an epistemologist, I have never been particularly exercised by this problem. In part, this is because I reject a good deal of the Cartesian framework itself. But it is also because, while I think that the topic of other minds is a rich one for epistemology, the traditional “Problem of Other Minds” utterly fails to capture the problem other minds pose to epistemology.

To get at this problem, we do well to start off with the assumption that knowledge of other minds is not especially problematic: not only do we know that other minds exist, we also sometimes know how another mind – or rather the person with the mind – takes the world to be. When one person knows what another person believes about the world, I will call this content knowledge of another mind. From an epistemological point of view, content knowledge of other minds would appear to be very special. Suppose you know what another person – say, Smith – believes – say, that it is currently raining. For you to know this is for you to be in a position to appreciate that Smith takes it that there is sufficient evidence to settle the question regarding the current weather. Insofar as Smith is an ordinary, cognitively healthy human being, she forms her beliefs in accordance with what she takes the evidence to support.   As a result, the fact that Smith believes that it is raining is a consideration that tells very strongly in favor of the truth of what she believes: so long as you have no reasons to doubt Smith’s competence or to question whether it is raining, your knowing that she believes this ought to lead you to form the same belief. And if this is correct, then other minds have an epistemic significance that is very different from the epistemic significance of ordinary evidence. And herein is what I take to be the new Problem of Other Minds: to characterize the epistemic significance of other minds.

This “problem” is less a single problem than it is a collection of topics to be addressed. Front and center is the question about the epistemic significance of another person’s speech. The main source of our content knowledge of another mind comes from observation of how another behaves, and in particular what she tells us; but of course it is possible that what she tells us is not really what she believes. What epistemic significance should we ascribe to the fact that another person said such-and-such? This question, which is the core issue in the epistemology of testimony, is one we must address as we seek to characterize the “new” Problem of Other Minds.

But when it comes to the epistemic significance of other minds, testimony is only the tip of the iceberg. In forming beliefs about the world, we rely on one another in a host of further ways. When we rely on experts, we rely by extension on the institutions which certify expertise and determine who are the experts among us, as well as on the people who run such institutions. When we rely on professionals, we rely by extension on the professional organizations that regulate their intellectual conduct. When we rely on scientists, we rely by extension on the many individuals who participate in the construction, validation, and upkeep of the scientific instruments on which the scientists themselves rely. When we rely on journalists, we rely by extension on the social practices through which the media try to ensure reliability and relevant coverage. When we rely on our friends and neighbors for information, we rely by extension on the informal social practices that have evolved whereby we share information with one another, as well as on the methods through which we enforce the norms of those practices. In these and many other cases, we also rely on the teams of people involved in the participating institutions, social practices, and epistemic groups. All of this (and no doubt more) should be included when we think about the epistemic significance of other minds.

What epistemic significance should be ascribed to the cognitive and intellectual work of others, when one implicitly relies on them to do the sort of background work just described?

Traditional epistemology offers a simple answer: in itself, this work has no epistemic significance whatsoever. According to the traditional view, what has epistemic significance are the subject’s beliefs about her reliance on others, together with the inferences that she draws from those beliefs. On such a picture, it is as if each of us is an epistemic island: if we are to acquire knowledge through our reliance on others, each of us needs to have enough information in our own heads about the social practices on which we are relying, and/or about the reliability with which those practices yield truths.

I reject this traditional view outright. Knowledge itself is often the result of a very complicated division of intellectual labor. There is little reason to think that the participating individuals need to know very much about the division of labor on which they are relying. On the contrary, as individuals we are the recipients of an epistemic good – knowledge – in whose production we may have played little role, just as in our role as consumers we are the recipients of economic goods and services in whose production we typically play little role. The task is to understand how we relate to one another as knowers, such that we can epistemically benefit in this way from the intellectual work of others. That, I submit, is the overarching question raised by the “new” Problem of Other Minds. Any epistemology that purports to address this I call Seriously Social Epistemology.


In the next decade or so I fervently hope that an increasing number of thinkers from a variety of different disciplines become interested in pursuing Seriously Social Epistemology. More investigation is needed if we are to understand the social dimensions of knowledge. A good part of this will involve an attempt to understand the social practices whereby we acquire, process, preserve, transmit, and assess knowledge. To attain such an understanding, we will of course need philosophers – above all, epistemologists, philosophers of science, political philosophers, and ethicists. But we will need insights from a great many other disciplines as well, including Political Science, Psychology, Sociology, Economics, History, Library and Information Sciences, Anthropology, and no doubt many others. Given the increasing ease and efficieny of social sources of information, as well as our recent worries about “information bubbles,” allegations of “fake news,” and the like, this sort of research is extremely timely.

9 thoughts on “The Problem of Other Minds, Old and New

  1. Thanks Sandy, for a really interesting post.

    I wondered whether it might be possible to get a little more elucidation on the question of what is meant by ‘traditional epistemology’. I can think of three broad categories of answers to the question of what the epistemological significance of other minds is:

    None whatsoever—people have other minds and so they produce deliverances in ways that are different to, say, instruments. But this doesn’t show us anything interesting about epistemology, any more than the fact the way that my eyes work is different to the way that my ears work shows anything interesting about epistemology.

    Some limited amount—people have other minds and so they produce deliverances in ways that are different to, say, instruments. This does show us something interesting about epistemology. People producing statements, for example, might do so reliably because they *know* things and the reliability of instruments can’t be explained in this way, for instruments do not know things. Equally, we can have reasons for thinking that people *know* things and we can’t have such reasons about instruments. So other minds show us that we need to think about epistemology in new ways, but they don’t press us to introduce any new epistemic notions or categories; we’re still broadly in the domain of reasons and reliability (and so on…)

    A large significance—people have other minds and so they produce deliverances that are different to, say, instruments. Moreover, we can’t even adequately account for these things in terms of the traditional epistemic categories because acknowledging the significance of other minds introduces ways of thinking that don’t fit into the traditional categories. In addition to, say, evidence, we also need a notion of *assurance* (that can’t be reduced to evidence) and in addition to, say, reliability, we need a notion of *transmission* (that can’t be reduced to reliability). So other minds show us not only that we need to think in new ways, but that we need new categories to do an accurate job of describing those ways of thinking.

    I’ve largely modelled this on the epistemology of testimony, which is what I know best. My apologies if some of this is opaque. In my head, it’s fairly clear that (1) is a version of traditional epistemology and (3) is not. I was wondering about (2). Is (2) a version of traditional epistemology, in virtue of its claim that we can sort out epistemology perfectly well without the idea of assurance and transmission and the like, or is it a version of Seriously Social Epistemology, in virtue of its claim that epistemology would be different if there weren’t any other things that had minds?

    Plausibly the answer is *neither of the above*—the positions in (1) and (3) are so unlike one another that it’s probably not surprising that there’s middle ground between them by combining bits of one with bits of the other. But I’d be interested to hear a bit more about where, to other people’s minds, traditional epistemology stops and Seriously Social Epistemology starts…


    1. Stephen,

      Thanks for the reply. I am inclined to take «traditional epistemology» to be committed to individualism about the epistemic statuses (warrant, justification, knowledge), where individualism is the view that the only mind whose states and properties are relevant to the epistemic assessment of a given individual S is S’s own mind. In this sense your (2) is a version of Seriously Social Epistemology. I should add, though, that I think that without some explanation for why other minds are relevant to epistemic assessment, we are left with a mystery — and that the explanation will involve things that might commit us to elements of (3) — here I am thinking of transmission or something like it. (Why would we think that the reliability of the speaker’s testimony is relevant to justification, unless we think that somehow this is ‘transmitted’ to the recipient in some way — the interpersonal analogue of memory? (Or this anyway was what I tried to argue in RELYING ON OTHERS.) In any case, that a position like (2) will tend to incorporate elements of (3) remains a speculation, but one worth thinking about.

    1. This is intriguing, Matt, I will have to have a look at it, but at first blush it looks Seriously Social. (I guess my attitude towards SSE is: let 1000 flowers bloom.)

      1. Thanks, Boaz. Note however that my proposal is not that it should be the paradigm; I’m trying to give a minimal account of what knowing people ‘personally’ is, and to sort out how it’s related to other epistemological issues.

        I’m dismayed that I hadn’t yet come across that paper by Code (recent similar papers, by Matheson, Lauer, and Talbert, also do not cite it, unfortunately). Though I had looked at other work by Code wherein she invokes the importance of subjectivity, those didn’t attempt much discussion of what it is to know other people.

        Will have to cite it from now on.

  2. Hi Sandy. This seems like a somewhat restrictive framing of social epistemology as the study of communication interfaces between minds. Where so questions about social power relations and conflicts fit within this framing? The interfaces are affected by such power relations, but they are not located at the interfaces. How do social institutions and social facts fit in? They do not exist merely in the minds of people. A prison, for example, is made of stone and steel, and putting a person in prison may affect the generation and dissemination of knowledge. Can this effect be studied if we only examine relations between minds? It this a job for social epistemology?

    1. Hi Boaz,

      Thanks for the reply. I don’t think the proposal is quite as restrictive as you seem to fear. Two points are worth highlighting.

      First, when we rely on others, we rely on the institutions in which they act and social practices whose standards entitle us to expect things of them and by reference to which we assess their performances. As I tried to emphasize above, I see these as in the mix. Although I didn’t mention this here — I do so in other places — I think we can and should distinguish the institutions and social practices which are «legitimate» from those that are not, and I would say that discussions of social power relations and conflicts can be discussed in this connection (and elsewhere). This is why I suggested above that social epistemology ought to include elements of ethics and political philosophy — to address questions of «legitimacy».

      Second, I take the domain of social epistemology to be as extensive as the core task of social epistemology, which is «to understand how we relate to one another as knowers, such that we can epistemically benefit in this way from the intellectual work of others.» To be sure, this gloss puts on the positive spin — how we BENEFIT from others — rather than a negative one — how some EXPLOIT others. But if you like you can read «benefit from» to cover the variety of the ways in question.

      As for your comment about prison, you are no doubt right but I worry about having a domain that is TOO GENERAL. Consider: ensuring that people have a good diet can affect the generation and dissemination of knowledge (if you’re hungry or ill, you won’t be as good an investigator); so, too, ensuring that we are not constantly killing one another can affect this; etc.; but I worry that if we include these things in social epistemology there will be little to do with humans that we will exclude. It’s not that these things aren’t interesting, worth studying and pursuing, etc. — of course they are — but rather that I am trying to give some sense of the sort of research that (as I see things) would contribute to our understanding of the social dimensions of knowledge. I don’t pretend for a moment that my characterization gives it fixed borders (it doesn’t); but I do hope that it gives us a sense that not anything that involves other people and affects the generation or dissemination of knowledge is relevant.

      1. Hi Sandy. Many fields (sociology, economics) are imperialist, and try to account for every phenomena under the sun. Epistemology is curiously not, and (analytic) social epistemology has inherited the tendency to restrict its scope. It seems that some social epistemologists are not so much interested in the social dimensions of knowledge, but only in the social world as a new playground for playing their old logic game of «Smith believes that p and p is true but Smith does not know that p». I think the diet example is misleading. The epistemic influence of a good diet is probably mostly biological, and therefore not very interesting to a social epistemologist. But if prisons have a genuine epistemic influence, and it’s not trivial (I say «if» because I don’t know), why should it be outside the scope of social epistemology? In other words, what social factors interestingly participate in knowledge is an empirical contingent question. Some research questions may be discovered to be duds (maybe the affect of prisons is one of them), but I see no need to draw principled borders around the domain of social epistemology.
        Also, I agree with your substantive claim about social epistemology (though should it be still said in 2017?), but framing it around the problem of refuting idealism seems unnecessary to me.

Deja un comentario

Tu dirección de correo electrónico no será publicada. Los campos obligatorios están marcados con *