There have been many criticisms of the Cogito and many versions of it (some preceding Descartes’ version). Lichtenberg’s criticism has been influential: all we know with certainty is that there are thoughts occurring, not that there exists someone having those thoughts. We introspect our thoughts and thereby come to know that they exist, but we don’t introspect any subject that has those thoughts. Maybe there are thoughts without thinkers. We cannot with certainty infer a substantial self from the mere occurrence of thoughts. To this objection it can be replied (following Frege) that there cannot be thoughts without a bearer of thoughts: it logically follows from the existence of thoughts that someone has those thoughts. Descartes might argue that thoughts are “accidents” and that all accidents need substances in which to inhere; so if thoughts are occurring they must be occurring insomething. There is no thinking without a subject that thinks—there are no “free-floating” thoughts (doubts, pains, etc).
I am sympathetic to this kind of reply to Lichtenberg, but I think there is another gap in the argument that I have not seen addressed. Descartes does not conclude from the Cogito that someone exists; he concludes that heexists. The Cogito is not, “There are thoughts, so someone exists” or “I know that thoughts exist, so someone exists”; it is “I think, therefore I exist” or “I know that thoughts exist, so I exist”. If we eliminate the question-begging occurrence of “I” in the premise, the proper form of the Cogito should be: “There are thoughts, therefore I exist”. But now that looks like a patent non sequitur, because those thoughts might not belong to me. It is true that for Descartes’ purposes the impersonal form of the Cogito is good enough, since he can still prove that there are things other than thoughts, viz. selves; but he cannot prove his own existence this way, because the bearer of the thoughts he knows to exist might be someone else. He knows that thoughts exist but he doesn’t know whose thoughts they are. So he can’t prove that he exists by this form of argument. The skeptic will ask how he knows that it is him that is thinking, i.e. how he knows that the thinker of the thoughts he knows to exist is identical to himself. There is a logical gap here that needs to be plugged, however strange the gap may sound.
I can imagine a reply to this point along these lines: if the thoughts belonged to someone else I couldn’t know with certainty that they exist, because I can’t know other minds with certainty; so they must belong to me, given that I know of their existence with certainty. I can rule out the possibility that the thoughts belong to someone else by observing that I am certain of their existence, which I couldn’t be if they belonged to someone else. That is, I use my certainty of the existence of thoughts as a premise in the argument to my own existence: “I am certain that thoughts exist, so it must be me that is their bearer, since I can’t be certain of the existence of other people’s thought”. It is not the known existence of thoughts that leads to the conclusion that I exist—that only proves that someone exists—but the certainty I have in my knowledge of their existence, since this is incompatible with the possibility that the thoughts belong to someone else.
This, however, is not the Cogito as we have come to know and love it, in all its simplicity and straightforwardness; we now need to build in some assumptions about knowledge of other minds and how it differs from knowledge of one’s own mind. The skeptic can then ask us to justify these assumptions: Is it really inconceivable that I could know another’s mind with the certainty I know my own mind? Couldn’t God have hooked my introspective faculty up to another mind so that I directly know the thoughts in that other mind? If so, I could know of the existence of thoughts with certainty but it is false that those thoughts belong to me—it just seems to me that they do (or perhaps it doesn’t even seem that way). The mere fact of certainty about the existence of thoughts isn’t enough to justify the claim that it is my thoughts that I know about. Still, it may be that this hidden assumption is what enabled Descartes (and others) to overlook the possibility to which I am drawing attention. He just assumed that any thoughts he could know about with certainty had to be his, but the resourceful skeptic sees a gap here—how can Descartes be so sure that those thoughts don’t belong to someone else? The Cogito therefore only gives us certainty that someone exists (pace Lichtenberg) not that we ourselves exist. My existence remains unproven.
This suggests that the basis of my knowledge that I exist is not the Cogito or anything like it. I do know with certainty that I exist, but the Cogito cannot yield such first-person knowledge; so I must have this knowledge in some other way. It is not an inference from the existence of thoughts.