Abstract
The philosophical literature on so-called first-person thinking or sethinking is based on a battery of examples that contrast two types of case. Reflexivity occurs in both cases because the subject thinks of himself and attributes a certain property to himself, say the property of being lucky. In the first type of case, if the subject were to express his thoughts linguistically, he would use the first person: "I'm very lucky". The subject is thinking of himself as such, as we do when we say "I". In the second type of case, the subject thinks of a person who happens to be himself, but without necessarily knowing it. Reflexivity is accidental or de facto, rather than de jure as in the first case. For example, on learning that the winner of the lottery is going to be a billionaire, the subject thinks "He's very lucky", without realizing that he himself is the winner, and therefore that he is the person he thinks is very lucky.
The de jure reflexivity that goes hand in hand with the use of the linguistic first person can be explained by the fact that the word "I", by virtue of the rule of use that governs it, encodes the identity of the referent with the speaker. But we cannot explain in this way - by appealing to the linguistic meaning of the word "I" - the de jure reflexivity inherent in first-person thinking, unless we accept the idea that thinking is nothing other than internalized language, and that we think in the first person when we mentally pronounce the word "I".
In the second lecture, I mentioned Benveniste's idea that it is thanks to the linguistic mastery of "I" that subjects can think of themselves as such. Can we accept this idea? I don't think so. The rule according to which the speaker uses "I" to refer to himself can only be applied by a speaker if he alreadypossesses the first mental person, i.e. a representation of himself as himself. So we can't explain the de jure reflexivity inherent in the mental first person by appealing to the rule of use that governs the linguistic first person.
How, then, can we explain de jure reflexivity at the level of thought? What does it mean to think of oneself as oneself? I propose the following answer. Just as, in the linguistic case, the identity of subject and object is guaranteed by the rule of using the word "I", in the mental case the identity of subject and object is imposed by the nature of the "mental file" we deploy when thinking in the first person: this mental file Ego serves to represent oneself, just as the word "I" serves to designate oneself.
But can't the circularity objection that's fatal to Benvenistian strategy be turned against the theory I've just outlined? In order to put information in the Ego folder , where we put information about ourselves, don't we have to know that the information in question is information about ourselves ? This suggests that we can't reduce the concept of self to this mental Ego folder, since we can only operate the mental folder in question if we already have the concept of self.
Among the information we acquire about ourselves, however, some is immediately first-person. The existence of such intrinsically or primitively first-person information circumvents the circularity objection: we need to already have the self-concept in order to integrate into the Ego file information that isn't first-person from the outset, but this doesn't mean that in general the Ego file can't function without a prior self-concept.
