David Sosa

Any developed Millianism is forced to make an arbitrary choice. Some Millian theories are profligate: it suffices for believing that Clark flies that you assent to some way of taking that proposition. But Lois no more believes that Clark flies than she fails to believe that Superman flies. An abstemious Millianism requires for believing that Superman flies that you not refrain from assenting to any way of taking that proposition. Profligate Millianism gives subjects beliefs they do not seem to have. Abstemious Millianism takes from subjects beliefs they seem to have. The fact that Millianism confronts such a choice is itself philosophically significant

These three papers, each constituting a chapter, lie at the intersection of philosophy of mind and philosophy of language. Chapter 1 reviews and reassesses Kripke's puzzle about belief. I argue, contra Kripke, that the puzzle shows Millianism to be inadequate . It must be supplemented with a Fregean theory. But Millianism and Fregeanism need not be opposed. Developing a distinction between mental representation and linguistic representation, I divide the notion of proposition. It is one thing to be the object of a propositional attitude, another to be the meaning of a declarative sentence. This perspective on entrenched difficulties in philosophy of mind and language discloses a unified theory accounting for all relevant intuitions. ;Chapter 2 discusses externalism in the philosophy of mind. I investigate the relation between the issue of Fregeanism and Millianism in philosophy of language and the issue of internalism and externalism in philosophy of mind. Having characterized externalism precisely, I argue that it faces a serious difficulty: it cannot correctly maintain the distinction between ignorance and incoherence. Mere empirical ignorance will be enough to make an agent logically incoherent. Worse, if externalism is correct, internal duplicates might differ with respect to logical coherence. ;In Chapter 3 I defend the Russellian conception of names from Kripke's modal argument. I offer a way to accommodate Kripke's claims about the rigidity of names without abandoning the Russellian theory