Self and Others

To provide a full account of the ability to think “I”-thoughts, we need an explanation of the transition from implicitly self-related information to explicit self-representation.

In the previous post, I argued that world-directed action and perception do not require explicit self-representation. This raises the question of when explicit self-representation does become necessary. I believe that this need arises only in the context of intersubjectivity. Just as explicit reference to a place or time only makes sense for a being that is aware of the existence of other places or times, so explicit reference to the self only makes sense for a being that is aware of the existence of other subjects who have their own perspective on the world.

Moreover, I take it that the transition from implicitly self-related information to explicit self-representation proceeds in degrees, which suggests that we need an account that admits of different levels of explicitness (thereby transcending the dichotomy associated with the distinction between nonconceptual and conceptual representations).

A helpful model for how this transition might proceed can be found in Karmiloff-Smith’s (1996) theory of “representational redescription”, which she developed on the basis of her studies of infant cognition. According to this model, we can distinguish different levels of representation. At the first level, information is encoded in procedural form (or in terms of “knowledge-how”; see my previous post). As such, it remains implicit and is unavailable to other operations in the cognitive system. In a subsequent reiterative process of representational redescription, the information is transformed into increasingly abstract, explicit and less-specialized but more cognitively flexible formats of representation (culminating in conceptual representations).

If we apply this model to the problem of self-consciousness, we can see that there are different levels of self- and other-representation. Thus, my suggestion is that self-consciousness arises as the result of an increasingly complex process of self-other-differentiation, during which information about self and others that is implicit in early forms of social interaction becomes gradually redescribed into increasingly explicit formats.

As I aim to show in some detail in the book, this is borne out nicely by findings from developmental psychology regarding the development of our social-cognitive abilities. For example, while abilities for shared attention (i.e. sharing attention to an object with the caregiver) or social referencing (i.e. using the emotional expression of the caregiver to regulate one’s own emotions and behaviour), which emerge between 9-12 months, suggest some awareness of others and their intentional relations, they do not yet demonstrate the ability for an explicit distinction between self and other. While the child has de facto access to the mental states of others and engages with them, she need not explicitly represent these states as belonging to the other, for this access occurs during episodes of shared intentionality in which the child herself also experiences the mental states she shares with her partner.

The distinction between one’s own mental states and those of others begins to emerge around 14-18 months, when children show empathy, level-1 perspective taking and mirror self-recognition. For example, in order to show the comforting behaviour associated with empathy, the child needs to recognize the emotional distress of another as belonging to the other. Likewise, the child now learns to represent perceptual states as belonging either to herself or to others (e.g., she understands whether another can or cannot see something that she sees).

Yet, even here we do not yet have a conceptual understanding of mental states as such. This seems to develop much later — at around 4 years –, when children begin to master explicit false-belief tasks (i.e. they begin to understand that mental states can misrepresent, such that someone will search for an object where they last saw it rather than where it actually is), show level-2 perspective taking (e.g. they understand that a drawing looks upside down for someone who is sitting across from them), and are able to engage in intentional deception, among other things.

So, in a nutshell, my proposal is that explicit self-representation develops in parallel with the representation of others as subjects and agents. The recognition that others are intentional agents like oneself leads to a contrastive differentiation between the intentional states of others and one’s own intentional states, and thus to self-consciousness in the sense of having the ability to think “I”-thoughts.

In my next and final post I will sketch some of the implications of this proposal and questions that arise from it, for example with respect to our theorizing about social cognition.



Karmiloff-Smith, A. (1996). Beyond modularity. MIT Press.


  1. Thanks, Kristina, for another interesting post.

    I wonder what you think of the idea (which I suppose may have been Strawson’s?) that there’s a role for a self-concept or some such form of self-representation in making a distinction between self and world, which might be more fundamental than the distinction between oneself and other selves. Do you think this is mistaken, so that your proposal is in competition with it? Or are the two ideas compatible (maybe because the forms of self-representation required for these two distinctions are different in some way)?

    • Kristina Musholt

      Another excellent question, which I’m not quite sure how to respond to.

      I think that both for our ability to self-ascribe mental and bodily states we need some sense of other subjects (because I take self-consciousness to be a contrastive notion, and because, even in bodily experience, we experience ourselves as subject and not as object; hence the relevant contrast space must consist of other – embodied – subjects). I am also sympathetic to the view (though I don’t argue for it in the book) that concepts derive their meaning from the role they play in the intersubjective practice of communication and reason-giving (and so concept possession – including the possession of a self-concept – isn’t really possible in the absence of intersubjectivity).

      That said, might there still be some room for a more basic representation of a self-world-distinction? Maybe. Perhaps you could say more about what you take the role of this representation to consist in?

      In any case, I think I should re-read Strawson…

  2. Thank you Kristina for this clear introduction to such a complex subject.
    Just one comment: you seem to focus on an ontogenetic perspective of self-consciousness . Maybe a few lines on a possible phylogenetic/evolutionary aspect of self-consciousness would have been welcome (As you may remember, this subject is of some interest to me, but with a different usage of inter-subjectivity
    Perhaps you address such aspect in your book as a concept-less limited self-consciousness could exist for animals.
    Let me go through your book

    • Kristina Musholt

      Hi Christophe,

      Yes, I do in fact have a (brief) chapter on self-conciousness in animals, in which I suggest that there might indeed be basic forms of self-consciousness that can be attributed to (some) non-human animals.

      • I’m happy Kristina that you address this subject in your book. An evolutionary nature of self-consciousness has, I feel, a lot to bring to philosophy of mind (see ASSC/TSC presentations &
        It is unfortunate that xxth century philosophy of mind has somehow considered evolution as “irrelevant”. But things are changing. And I want to hope that your teaching as
        Professor of Cognitive Anthropology will cover that point which is probably key for an understanding of human mind.

  3. Pengbo

    Hi Kristina
    Thanks for introducing your book! Fascinating project!
    I wonder if you can say a little more about the development from implicit forms of self-representation to more explicit forms. Here are some more specific questions.
    First: I am curious about the role language (or at least some words, like first person pronouns) plays in this process. You mentioned that “around 14-18 months… children show empathy, level-1 perspective taking and mirror self-recognition”. This is also the period they start to use first person pronouns to self-refer (and the older they get, the more often they use these terms to self-refer). So do you think the acquisition or use of these (or other) words contributes to the development of (at least some) explicit forms self-representations, or they are just manifestations of the self-representation that children acquires independently?
    Second: I wonder whether, on your view, non-human animals can have explicit self-representations too. I have in mind those who pass the “mirror tests”, like chimpanzees. One might argue (and some psychologists do, I think) that passing the test shows that they recognize themselves in the mirrors, which in turn suggests that they possess self-concepts of some sort.
    Third: I’m a little unclear about the idea that self-representations have different “levels of explicitness” (and I’m sure you explained it in more detail in the book), partly because I’m not entirely sure how the notion of explicitness is fleshed out in the book, and partly because I am a little (but not too much) concerned about the individuation of self-representation. So, for example, when you talks about the development towards more explicit forms of self-representation, do you mean that a) the subject acquires different (kinds of?) self-representations, some are more explicit than others; or b) it just acquires one self-representation, but that representation shows different degree of explictness as it develops ( incorporates more self-related information, figure in general forms of reasoning, be applied to a wider range of cognitive tasks,etc.)?

    • Kristina Musholt

      Thanks for your excellent questions and apologies for the delay in my response.

      1) The role of language is a very interesting issue indeed, though I haven’t made up my mind about it. In fact, your question is one of the open questions I raise in the last chapter of my book (also see my next post). My intuition is that the acquisition of words does play a role in developing more explicit representations/concepts, though how exactly it does so I’m unsure about. It’s one of the issues I’d like to address in future research.

      2) I address this question in chapter 7 of my book where I argue that although much more research is needed, the evidence to date (based on studies of mirror self-recognition, mindreading and metacognition) does suggest – as you suspected – that some non-human animals possess at least some basic forms of self-representation. (In the book I distinguish more precisely between different levels and argue that some non-human animals seem to show self-awareness at what I call level 2, which is to say that they can differentiate between themselves and others and ascribe basic intentional relations, such as perceptual states, but probably without an explicit understanding of the nature of mental states as such.)

      3) Interesting question! In the book I argue while explicit representations require a redescription of the original implicit representations, this doesn’t mean that the implicit representations are abolished or replaced in this process. So, for example, some of our social interactions are based on more explicit forms of representation (e.g., explicit mental-state ascriptions), while others are based on more implicit forms of representation, and this is still true in adults, where different forms of representation can exist alongside each other. The same is true for self-representations. That said, I also think that when we reach the level of acquiring an explicit self-concept, this can be used to incorporate different kinds of information about oneself, such that the way we think about ourself (or our ‘self-file’) becomes more complex/rich over time. However, I admit that the relation between different levels remains a tricky issue – it’s another one of those things I need to think more about! (And I’d be very interested to hear other people’s thoughts about this!)

Comments are closed.

Back to Top