Sunday, May 24, 2009

Descartes Contra Wittgenstein

“It may easily look as if every doubt merely revealed an existing gap in the foundations; so that secure understanding is only possible if we first doubt everything that can be doubted, and then removed all these doubts.”

--Ludwig Wittgenstein, Philosophical Investigations, section 87


I. I think, therefore I am

Descartes' famous principle, cogito ergo sum ("I think, therefore I am," also known simply as the cogito), is probably the most widely known philosophical statement, and it is often considered to exemplify the only thing one can know for certain: that one exists.

The cogito might be considered an instance of modus ponens: "If I am thinking, then I exist; I am thinking, therefore I exist." However, Jaako Hintikka (1962) compellingly argues that Descartes' cogito ergo sum is not a logical inference at all, but a performative act: that a certain cognitive act instantiates direct knowledge of one's existence--that the act of thinking makes one's existence self-evident. On this view, "I think, therefore I am" is not a logical result, but expresses an inevitable consequence of thought itself. And because it is thought which manifests the necessity of the existence of the I, Descartes concludes that the I exists by virtue of thought alone. The I is, by definition, a thinking substance.

Descartes believed that he could doubt everything, including logical and mathematical judgments, but not that he was doubting. Whether it was a logical or an existential consequence, the fact remains that, by virtue of his thinking, he concluded that he must exist. His claim, moreover, was that the cogito established to himself that he existed as “a substance whose essence, or nature, was nothing but thought” (Discourse On Method, Part 4). His conclusion was explicitly dualistic: His soul was pure consciousness (res cogitans); his body was extended matter (res extensa).

It appears that Descartes did not thoroughly doubt all judgments, for even the interpretation of an existential judgment relies on some use of language and logic. If Descartes had doubted the possibility of all logical inference, he would have had to stop doing philosophy altogether. We might thus be tempted to think that doubting the cogito would require abandoning all hope of philosophy. For what could be more certain than the fact of one's existence?


II. Grammar

Descartes could not doubt the logic of the word "I." The cogito does not establish or elaborate upon the meaning of the word "I." All it does is remind us of how we use the term. The cogito can only serve as a reminder that the sentence "I do not exist" is not a valid proposition in our discourse. It is a reminder of the rules of our grammar, and nothing more. It is not a foundation for knowledge, and it does not represent a bedrock of philosophical certainty. It is no more certain than the rules of grammar which it presupposes.

To understand the meaning of the term "I" and how we know ourselves, we have to understand how the language we use is learned, and that requires first learning the language.

This takes us to the quote I placed at the top of this post. Wittgenstein’s point is that there is no gapless foundation to be revealed, because the foundations are not there ahead of time, waiting to be discovered. We may doubt as far as want, but this will only lead us to exhaustion. Foundations are to be built, not uncovered. We could thus say that Descartes was working in the wrong direction.


III. Mind and Body

Descartes claimed that his mind was distinct from his body, because he could doubt the latter but not the former. But how could one doubt that one had a body? Just try—in a real case, with discernible consequences—try to doubt that you have a body. What could such “doubt” consist in, if not just the words, repeated either to yourself or out loud: “I don’t have a body . . . I am just a mind?”

Those are empty words, no different than, “I don’t have a mind . . . I am just a body.” Repeating them does not constitute doubt, because these words have no discernable consequences. They are insignificant. It would make as much sense to say, “all logic is invalid . . . there are no valid inferences,” or perhaps, “there are no thoughts, only words; no feelings, only functions.” Why not, "there are no words, only thoughts; no functions, only feelings?" What's the difference?

Such mantras can produce interesting psychological results, but they are not to be taken seriously as philosophical propositions.

When I think (or feel), I think (or feel) something particular. And while it may be said that I cannot doubt that I am thinking (or feeling), I also cannot doubt that I am thinking (or feeling) that which I am thinking (or feeling).

I might doubt that what looks like a chair is really something I can sit on; or that the middle-C I hear was struck by a piano, as opposed to a recording. But I cannot doubt that I am seeing something that looks like a chair, or that I am hearing something that sounds like a middle-C. Interestingly, it is only when I have more than one kind of sensory organ that I can know different things can look like a chair and sound like a piano. If I was a creature with only one form of sensory input, I would not make such distinctions, and so I would have no occasion to doubt my perceptions.


IV. Feelings

The cogito seems to us to indicate something strange about consciousness. We are tempted to think that it reminds us of more than just the grammar of the word "I." We want to say that it reminds us just how different consciousness is from the rest of the world--that thinking and feeling are purely subjective, making up an inner world which is wholly different from the rest of nature.

The mistake here is in thinking that, when we introspect, we discover a unique property of consciousness, a feeling of feeling, which captures the essence of what we are.

Here's an exercise to help see what is going on here.

Close your eyes and hold a sponge in your hand. Try to ignore any sounds or smells that might indicate the presence of a sponge. Only feel it with your hand. Now, do you also feel that you feel it?

How could you?

Maybe you try to focus on the awareness of the sponge. But what could your awareness of the experience be, apart from the feeling itself? The feeling is the awareness.

When you are aware that you are feeling a sponge, you do not feel the feeling of feeling a sponge. You simply feel the sponge. Your knowledge of the feeling is not divisible into two parts: 1) the feeling of the sponge; and 2) the feeling of feeling the sponge.

Of course, while we are feeling a sponge, we can talk about it. We can say to ourselves or out loud, I am feeling a sponge. We might be tempted to think that such verbal activities represent some other kind of knowing, some higher level of knowledge of the feeling. But this is a mistake. For one thing, how do we know that we are thinking I am feeling a sponge?

We experience those words, but this experience is just another feeling, like the feeling of the sponge. We could be mistaken about the meaning of the words, just as we could be mistaken about the meaning we attribute to the feeling of the sponge. Is your certainty about the meaning of your thoughts any clearer than the certainty of the meaning of your other sensations? How could it be?

This is the point Descartes missed, and which Wittgenstein got.

We want to say that the statement “I am feeling a sponge” expresses the fact that I am feeling a sponge. And it surely does. But it also feels like something else, and this other feeling is not the feeling of feeling the sponge, nor is it the feeling of feeling in general. It is rather the verbal expression of the act of feeling the sponge. The verbal activity does not indicate a unique feeling of feeling.


V. Conclusion

In sum, there is no special knowledge to be found in the cogito, no ground for regarding one's own existence as the only philosophical certainty; no epistemic distinction between knowing that we are thinking/feeling and knowing facts about the world.

If we want to distinguish between two types of facts here, we can distinguish between grammatical facts and empirical facts. (The former are often called analytic and the latter synthetic). But here we must be careful not to assume that the line between the two is always so clear. And in any case, this distinction is not between self-knowledge and world-knowledge, or between certain knowledge and uncertain knowledge. It is rather between knowledge of the rules of discourse and knowledge of how those rules should be applied in new cases.