Consciousness and the Cogito

Consciousness and the Cogito

I think it is fair to report that people take the Cogito to reflect something important about consciousness, thinking, and the existence of the self. Roughly, they take it to demonstrate (or purport to demonstrate) that conscious thinking entails personal existence: if you know that you consciously think, then you know that you exist—and you do know that you consciously think. The validity of the inference depends on these three concepts and reflects their specific character. I will argue that this is wrong: in fact the Cogito is a special case of something far more general, having nothing specifically to do with consciousness, thinking, and the self.

            First, consider the unconscious: is there a version of the Cogito that applies to it? Not “I unconsciously think, therefore I am”, because there is no Cartesian certainty with regard to the unconscious: I might be wrong about the existence of the unconscious and the presence of thoughts within it. But what about the unconscious itself—can’t itformulate a Cogito? Couldn’t you unconsciously think, “I think, therefore I am”?  That is, can’t the unconscious be certain that it thinks, just like the conscious, and hence draw the inference that it therefore exists? The conscious Cogito concludes that the conscious self must exist, based on the certainty that conscious thoughts exist, and the unconscious Cogito concludes that the unconscious self exists, based on the certainty that unconscious thoughts exist. Why can’t the unconscious mind be as certain of its thoughts as the conscious mind is of its? It presumably has knowledge of its own thoughts, desires, emotions, etc., just as the conscious does, so why can’t this knowledge amount to certainty? If so, the Cogito has nothing intrinsically to do with the conscious mind: you can run it on the unconscious mind. In fact, since Descartes believed in unconscious knowledge in the form of innate ideas, it would be quite natural for him to put forward a Cogito for the unconscious; certainly, Leibniz’s rather robust view of the unconscious would be conducive to such a move. It is thinking that implies personal existence not consciousthinking specifically.

            But is the Cogito restricted to thinking? Evidently not, since we can formulate it using other types of mental state: “I desire, therefore I am”, “I see red, therefore I am”, “I’m in pain, therefore I am”. The existence of the self follows from the truth of any mental predication, not just thinking specifically. I can be certain of these predications too; so they form a solid premise from which to infer the existence of the self (assuming such an inference is valid for the special case of thinking). All psychological states require the existence of a bearer—something that has them. But can’t we go a step further: what about computational states of the mind or brain? Couldn’t the mind/brain engage in a Cogito of its own—“I compute, therefore I am”? If it has access to its computational states, comparable to that of the conscious mind and the unconscious mind, why can’t it proceed to formulate a Cogito of its own? If it is certain that it computes (or possesses the computational analogue of certainty), it has a premise from which it can infer its existence. Of course, I can’t enunciate such a Cogito because I (a conscious self) don’t have this kind of access to my computational states, but my computational brain might have such access—there seems to be nothing logically impossible about this. We can conceive of a being with such resources: one day it occurs to this being, in the depths of its computational explorations, that its existence follows from its indubitable acts of computing. So the Cogito is not specifically about the mind, conscious or unconscious; it applies to any system that has propositional content—any representational system.[1]

            But what about the contents of thought: can’t we infer that the contents of thought exist too? I can say, “I think the earth is round, therefore the concepts of the earth and roundness exist”. The existence of these concepts is presupposed in thoughts that employ them, so we can infer the existence of the concepts from the existence of the corresponding thought. Admittedly, this is a pretty trivial result, but it shows that the type of inference we are dealing with is not limited to the self: thoughts need selves, but they also need concepts. Likewise, we can infer from the truth of “I am in pain” that pains exist, since being in pain presupposes the existence of pains. Anything whose existence is embedded in the premise can be extracted in the conclusion: not just the subject of the thought but also the thought itself, as well as the concepts employed in the thought.[2] If we can be certain of the premise, we can be certain of its logical consequences; but some of these are existential in form—various existences follow from “I think”. The traditional Cogito is just one instance of this general principle, focusing on the existence of the subject of the thought; but there are other instances. It works, if it does work, by relying on that general principle—that you can infer the existence of anything that logically follows from the truth of “I think”. The inference has nothing specifically to do with the special nature of the self. Thoughts need existing subjects, but they also need existing concepts, as well as their own existence. The existence of the self is not the only existence to follow from the fact of thinking. Given that we can be certain that we think, we can be certain of everything that follows from that; and some of these consequences are existential in form. Maybe we could also infer that time and space exist given that thoughts do, so that thinking is even more existentially pregnant than we supposed; but it is certainly not limited to the self. Indeed, the inference to the self has notoriously been questioned, so it might be less apodictic than these other possible inferences. The point is that the deductive structure of the Cogito is not specifically about the self, nor specifically about thinking: it is about the certainty of any indubitable proposition and its existential commitments. The proposition that 2 + 2 = 4 is not about thinking and implies nothing about selves, but it too provides an instance of the general principle in question: it can be known to be true with certainty and it implies the existence of numbers as well concepts of numbers. Conventionalists about analytic truth are committed to the idea that such truths imply the existence of language and conventions, as well as being known with certainty, so they too provide an instance of the general principle employed by the Cogito. The Cogito is not the unique existence-generating proposition it has been thought to be. And let’s not forget that by itself it tells us nothing about the nature of the self—it could be a momentary, solitary, non-substantial, entirely mental entity. So it gives us very little with which to combat the skeptic, hardly more than the other existential implications I have cited. It is quite unclear why it alone is supposed to form the foundation of Cartesian epistemology.[3]

[1] We can envisage the computational Cogito being expressed in the language of thought: we conscious beings don’t have access to this language (still less infallible access), but we can suppose that a homunculus module does have such access. This being can therefore output the sentence, “I compute, therefore I am”—or its Mentalese equivalent.

[2] It should be noted that the move from “I think” to “Thoughts exist” is not tautological, since the premise says nothing explicitly about existence: it would be consistent to suppose that “I think” is true but “Thoughts exist” is not. It might be held that the mere predication of thinking does not imply that there are certain entities that exist: the former is merely adverbial, say, with no implications about an ontology of things called “thoughts”. That might be regarded as illicit reification.

[3] The move from “I think” to “Concepts exist” (or “Thoughts exist”) is not obviously more trivial than the move from “I think” to “A subject of thinking exists”. Both conclusions add something to the premise, signaled by the use of “exists”, so it is not clear that the Cogito provides more leverage against the radical skeptic than those other existential implications. In neither case do we get anything substantive about the nature of the entities whose existence is implied.