Proof of an External World

Proof of an External World

Kant famously (and ruefully) remarked that it was a scandal of philosophy that it has been unable to come up with a proof of the external world. He was right: it is a matter of some embarrassment that philosophy should be unable to prove something so obvious, so commonsensical. What good is philosophy if it can’t even prove something that elementary? The proof need not be simple or obvious (that also would be to the detriment of philosophy as an interesting enterprise); it could be intricate and convoluted, with spots of uncertainty. I am going to offer such a proof: it has a Kantian ring, but is not to my knowledge to be found in Kant (or anywhere else). This should remove the scandal and prove the worth of the discipline of philosophy. It should also be personally satisfying (I myself feel a great sense of relief).

Let’s start with a simple thought, which will point us in the right direction. Suppose the skeptic says that our perceived world might be pure projection—a figment of the human imagination, corresponding to no further reality. After all, we already agree that much of it is projection—as with the perception of color and other secondary qualities. Why not all—why shouldn’t primary qualities also be subjective projections? We might think there is an obvious reply to this: projections need a screen onto which to project, which is not itself a projection. Thus, material objects in space provide the screen onto which colors (etc.) are projected; they are the equivalent of the movie screen that pre-exists the pattern of light thrown onto it. So, the perceived world can’t all be projected image; it must include a non-mental background. If so, we have a proof of the external world: it follows from the fact of subjective projection that something other than projection must exist, viz. material objects in space. But, of course, the skeptic will not be deterred by this simple-minded maneuver: he will suggest that the alleged non-mental screen is really just a virtual world, an imaginary world, a fictional world. So-called objects in space are non-existent objects, or may be for all we know. It only seems to us as if such objects exist; they might all be non-existent intentional objects, like objects in dreams or works of fiction or hallucinations. It is that hypothesis that needs to be disproved in order to prove that there is an external world. For example, there is an appearance of a square object in my visual field, but this could be a non-existent square object not one that really inhabits objective space. How can we rule this possibility out? I could be dreaming of a square object in front of me, this object being a mere figment of my imagination.

Here is the problem with this alternative skeptical hypothesis: we normally think there is a definite number of things that fall under a perceived (or conceived) attribute, but this will not be so if its extension consists only of non-existent objects. If lions and square things exist, then there is a definite number of them, known or unknown; but if they don’t exist, then there is no definite number of them. The point is familiar: there is no definite number of moles on Hamlet’s back or unicorns or angels or fairies. Such things are numerically indeterminate. But we normally think that ordinary objects of perception come in definite quantities, so they can’t just be non-existent entities. It follows from the fact of numerical determinacy that the objects of perception are not non-existent. Indeed, it is their existence in space that accounts for their numerical determinacy, since material objects are individuated by their location in space. Since non-existent objects do not exist in space, they can have no spatial principle of individuation that underpins their numerical determinacy. So, the skeptical hypothesis can be ruled out and our normal conception accepted. However, the skeptic is not beaten yet: why not say that there is no definite number of square things or lions since they are non-existent intentional objects? Why not bite the bullet and accept that consequence?

First, we should note that even if we do bite the bullet, we are still accepting that there are non-mental objects, because non-existent square things are not mental entities, any more than existent square things are (same for lions). We can quantify over them and they are not mental, so we have still proved that there are non-mental things (that don’t exist). But second, it is not so easy to give up on the numerical determinacy of attribute extensions: for attributes like these (sortal attributes) provide principles of counting, criteria of individuation, and these will generate assignments of cardinality. It is easy to miss this when an attribute applies to both existent and non-existent objects, but what sense does it make to suppose that an attribute that applies to pluralities of objects applies to no definite plurality of objects? If we claim that the attribute lioncorresponds to no definite number of lions, how can it be said to distinguish one lion from another? Not in virtue of position in space, to be sure, because non-existent lions don’t exist in (real) space. We lose the idea of a totality of individual lions standing in spatial relations to each other and adding up to a specific number of lions. That idea requires existence; it can’t survive in the realm of non-existence. The notion of non-existent lions is parasitic on that of existent lions, but then we are back with the external world as naively conceived. A fictionalist about minds (a mental eliminativist) has a problem about the individuation of minds—how many non-existent fictional minds are there?—and a fictionalist about bodies has the same problem about theirquantity. There really must be a definite number of minds and bodies for those concepts to have any intelligible content, but that idea goes out the window once we give up on existence altogether. Even the concepts of identity and difference begin to wobble when we enter the land of the non-existent (when are non-existent gods identical and when different?).

The attitude of sophisticated common sense is that we perceive a world of objects laid out in space, numerically distinct from each other, and forming totalities of specific cardinality. The skeptic tries to convince us that what we perceive are just non-existent intentional objects, but this involves abandoning the idea that we have concepts with definite cardinalities attached to them; and that is not a possible position, given the nature of our concepts (and associated attributes). Thus, an external world exists. The essential move in this proof is the observation that non-existence can provide no grounds for determining the number of things falling into the extension of a concept; only existence in space (in the case of material objects) can provide a basis for this determination. Things that don’t exist are not really countable in the way we normally (and rightly) take objects to be. Countability implies objectivity.[1]

[1] The proof here offered comes at the problem from a surprising direction. I think this is what we should expect, since no obviousmethod of proof has succeeded in removing the scandal. It would be surprising if the proof were not surprising.

Share
17 replies
  1. MJA
    MJA says:

    ‘If lions and square things exist, then there is a definite number of them, known or unknown; but if they don’t exist, then there is no definite number of them’. What if the skeptic is a Berkeleyan? He might be deny that there is an *external world* (i.e. a world existing independently of your thoughts or ideas), but not deny that there is a world. In this case the evil demon really isn’t far off the truth. The world is being caused in you–only by God and not a demon. If there is a mind-dependent world it might contain a definite number of lions, etc. But they too are not what you think they are. I’m not of course saying that Berkeley was a skeptic–far from it–but that a Berkeleyan skeptical argument is very close to an evil demon argument, and it avoids the indeterminacy objection above.

    Reply
    • Colin McGinn
      Colin McGinn says:

      I thought about putting a footnote in about Berkeley, so I’m glad you brought it up. My view is that Berkeley does believe in an external world–that is, a world external to human and animal minds (God’s mind). He takes ordinary objects to exist as fully as any materialist; he just thinks they have a divine nature. That is how he gets determinacy.

      Reply
  2. Giulio Katis
    Giulio Katis says:

    Could one state this argument in terms of constraints (the determinate number of some type of thing being such a constraint)? It seems even an ardent solipsist must admit some constraints on what is subjectively projected, otherwise there would be no structure to experience; and such constraints (or the data associated with such constraints) can be held to be external to one’s mind.

    Reply
    • Colin McGinn
      Colin McGinn says:

      The question is whether the constraints must be external–couldn’t they be internally derived? The solipsist thinks there is enough structure within the mind to give it definite content; no need for external objects.

      Reply
      • Giulio Katis
        Giulio Katis says:

        I don’t think so. The constraints (including the information content in them) must be external to the fabricating mental projection mechanism, otherwise they wouldn’t be constraints on that mechanism. Could the constraints exist unfabricated in another part of the mind, to be imposed on the fabricating part? If so, where did they come from? What or who imposed that structure on this other part of the mind (which is itself a constraint)? It just moves the question from one part of the mind to another.

        Reply
        • Colin McGinn
          Colin McGinn says:

          The constraints could come from the genes or the brain. We need an argument to prove that only the world outside the subject can constrain the mind appropriately. This is what my argument supplies.

          Reply
          • Giulio Katis
            Giulio Katis says:

            The genes and brain have structure, and are therefore subject to constraints. An infinite regress is required. My comment below also suggested your cardinality argument (cardinality being a structure) would seem to imply a constraint of some form. No?

  3. Giulio Katis
    Giulio Katis says:

    The reason why I raise constraints is that I think one may have to do a little more work to fill out why “countability implies objectivity”. For instance, consider being on Noah’s ark, and in front of you you see the only two lions to be in existence. Their count (two) could, of course, be just part of your projection in this particular case. If it really is two (and wont be 500 tomorrow), then this suggests a constraint is at work.

    Reply
  4. Free Logic
    Free Logic says:

    Have your argument proved that the famous Matrix movie scene where the heroes are showed to us, the viewers, as hallucinating in liquid filled tubes is impossible? If not what is the “external world” for these hallucinating people who can’t perceive what we the viewers can?

    Reply
  5. Marc
    Marc says:

    I am excited to read that you have a solution for Kant’s famous scandal of philosophy. Unfortunately, it is difficult for me to really understand the problem with the alternative skeptical hypothesis. You write:

    “Here is the problem with this alternative skeptical hypothesis: we normally think there is a definite number of things that fall under a perceived (or conceived) attribute, but this will not be so if its extension consists only of non-existent objects. If lions and square things exist, then there is a definite number of them, known or unknown; but if they don’t exist, then there is no definite number of them. The point is familiar: there is no definite number of moles on Hamlet’s back or unicorns or angels or fairies. Such things are numerically indeterminate. But we normally think that ordinary objects of perception come in definite quantities, so they can’t just be non-existent entities.”

    I have some trouble to really understand the point. Let me put it differently in simple terms. Does it mean, that we cannot count non-existing things (like unicorns)? On the other hand, real things can be counted, like 3 apples on a table; they are 3. And in that sense regarding the countability, existing and non-existing things differ. Is this the key point of the argument?

    Reply
  6. Tina Lee Forsee
    Tina Lee Forsee says:

    Hi Colin, I hope you don’t mind a stranger with virtually no knowledge of analytic philosophy intruding. I just finished your book, “The Making of a Philosopher”, and I found it to be a highly enjoyable, plain-language introduction to an area of philosophy I find hard to make sense of. (I tried to read Wittgenstein once and I don’t think I’ll try that again. Life is short.)

    My question here, and I apologize if it’s a stupid one, involves this:

    “… only existence in space (in the case of material objects) can provide a basis for this determination. Things that don’t exist are not really countable in the way we normally (and rightly) take objects to be. Countability implies objectivity.”

    But suppose we don’t assume a scientific understanding of space (as a thing ‘out there’), but instead we say that Kant may have been right about space, that it’s the a priori form of outer intuition? Wouldn’t the distinction you make between “real” countable objects and “unreal” uncountable objects still be maintained without implying anything about reality as it is in itself?

    I’m not a Kantian, but his arguments about space made a huge impression on me.

    Thanks!

    Reply
    • Colin McGinn
      Colin McGinn says:

      The difficulty is that Kant’s phenomenal world hovers between existence and non-existence. For him, it is a realm of existence, so objects are countable within it; he is rather like Berkeley. But then he goes on to discover another world, the noumenal world, and that also exists. He is a kind of double realist.

      Reply
      • Tina Lee Forsee
        Tina Lee Forsee says:

        Kant would roll over in his grave if he heard he was being likened to Berkeley, but I think I get your point. And that’s not to pick on Berkeley, whom I admire a great deal.

        Are you a double realist?

        Reply
        • Colin McGinn
          Colin McGinn says:

          Strawson says several times in The Bounds of Sense that Kant is more like Berkeley than he admits. Berkeley is much more of a realist than people tend to think (he just doesn’t like the concept of matter as used by his contemporaries). I too admire Berkeley.

          I am a double aspect realist.

          Reply

Leave a Reply

Want to join the discussion?
Feel free to contribute!

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.