Communication and thought attribution

Listen to audio

The modes of presentation we've been talking about so far (mental files) are psychological modes of presentation, corresponding to the way the subject represents the objects he or she is thinking about. They are to be distinguished from linguistic modes of presentation, which are an aspect of the meaning of the linguistic expressions referred to.

By virtue of its conventional linguistic meaning, a referential expression encodes certain properties of the entity to which the speaker is referring. It is these linguistically encoded properties that guide the public in identifying the reference. The mode of linguistic presentation thus acts as a constraint on the respective mental files of the interlocutors: when the linguistic expression used to refer to an object encodes property F (mode of linguistic presentation), the mental files that the interlocutors associate with this expression must contain the information that the object to which the file refers possesses property F. In this conception, communication does not involve the reduplication or sharing of thoughts between different subjects, but only the coordination and referential convergence of thoughts.

In addition to the coordination of interlocutors' respective mental files in referential communication, there's another kind of coordination between mental files : the coordination of two subjects' mental files when one of them reports the other's thoughts or utterances about an object. To account for this, we need to introduce a new distinction between two types of presentation mode (two types of psychological presentation mode, in this case) : the file deployed by the person reporting the word or thought when referring to the object that this word or thought concerns, and the file attributed by him to the person whose word or thought he is reporting. The notion of attributed record is relevant when the utterance that reports a subject's thoughts or words about a particular object receives a " opaque " reading, i.e. is seen as specifying not only the object in question but also the way in which the subject represents it to himself.

We know, by a simple application of the general rule, that the mode of linguistic presentation constrains the mental file deployed by the interlocutors (the speaker and the one addressed). Is the same true for the mental file assigned - does this mental file necessarily contain a piece of information corresponding to the property of the referent encoded by the expression used ? A negative answer is proposed, based on the existence of a non-canonical opaque type of reading for thought attributions. The discussion also focuses on apparent counterexamples to the thesis that the deployed mental file necessarily contains a piece of information corresponding to the property of the referent encoded by the expression used. It is shown that these counter-examples involve indexical files, and that the allegedly refuted thesis can be maintained provided that we admit the possibility of the reporter deploying an indexical mental file (and not a file of his own) to specify the object to which the attributed thought relates.