In this contribution, we argue that Moore’s paradox has its roots in the semantics of first-person. We build up on some of Frege’s concerns about the first-person, recently revived by Kripke as a criticism of the position according to which Kaplan’s two-dimensional semantics is all is needed for an adequate semantics of the first-person. First, we discuss the so-called pragmatic approach to Moore’s paradox, discussing its possible limitations, in accordance with some ideas expressed on the matter by Shoemaker. Second, we show that sentences where a predicate expressing a ‘phenomenal’ property combines with a first-person pronoun are bound to express Stalnaker’s diagonal proposition, and are true a priori. Crucially, the proposition expressed does not correspond to the ascription of a property to an independently established object. Finally, we provide significant empirical evidence to the effect that this emerging ‘subjective’ layer of meaning is actually what is needed to solve some of the puzzles around Moore’s paradox: the status of Moore’s sentences essentially depends, in fact, on the interaction between the ‘objective’ and the ‘subjective’ layers of meaning.
