Abstract
According to the "véhicularist" perspective, two co-referential concepts are the same concept if and only if the vehicle (the thin mental file) is the same. The conception (the content of the concept) may vary, but this does not affect the identity of the concept, which is individualized not by the content but by the vehicle. It remains to be seen how, from this perspective, the mental files themselves are individualized. What makes two co-referential mental files the same or different? Obviously, we can't answer this question by invoking the reference or conception conveyed, as this would be a return to the previously rejected modes of individualizing the concept.
As a vehicle, the thin mental file is a "continuant": a particular that persists through time, and notably through the variation in conceptions that the file conveys at different times. A widespread thesis concerning the individualization of particulars invokes origin: two particulars are the same if they have the same origin. Drawing on Sainsbury and Tye (2012), we can adapt this conception to the case of mental records, and argue that a mental record is individualized by the cognitive event that presides over the opening of that record.
The fact remains that concepts (and thoughts, which are assemblages of concepts) can be shared between different individuals. There would be no possible agreement or disagreement between different subjects if this were not the case. This is why Sainsbury and Tye argue (as did Frege and most philosophers) that concepts are a supra-individual reality. I, on the other hand, propose reversing the terminology and calling the concept the basic psychological reality, i.e. the mental file, individualized by its origin in the subject's mental life. Concepts in this perspective cannot be shared, since they belong to the mental life of a particular individual. But we can define a notion of "shared concept" as an equivalence class of private concepts. Two distinct private concepts belong to such an equivalence class, and are therefore instances of one and the same shared concept, when there is a certain relation Rbetween them , to be defined.
A first possible theory defines the R-relation as belonging to the same network of interconnected mental files thanks to a word in public language. This way of conceiving the R relation has certain advantages, but it comes up against a major objection: it is not necessary to share a language in order to share a concept. In the following lecture, we'll look at other ways of conceiving the R relation.
