One of the most interesting questions analysed in philosophy of language concerns the hypothesis that the referential competence – that is, the ability to link words to the objects to which they refer – is based, not only on linguistically explicit properties of that objects, but also on cognitive contents not always exhaustively definable by means of linguistic instruments. For example, experiments on neural networks suggest the possibility that perceptive properties of objects influence in an unconscious way our categorization of that objects. Moreover, neuropsychological syndromes – in particular, Capgras and Fregoli syndromes – suggest that emotional contents could influence our decision to consider an entity as a member of a particular category: emotional contents could be considered, in such cases, part of the meaning of words referring to that objects.