Kripke finds in Wittgenstein an argument for the conclusion that there are no meaning facts and considers the consequences of this outcome for the meaning of meaning-ascribing sentences. One immediate consequence is that their meaning cannot be given by their truth conditions. Kripke proposes instead that meaning ascriptions obtain their meaning from (i) their assertibility conditions and (ii) the non-representational function that the practice of asserting these sentences in these conditions plays in our lives, accepting that these sentences can’t play the role of representing the world. I present a strategy for avoiding this outcome. Meaning ascriptions obtain their meanings from their assertibility conditions, but they successfully perform the function of representing the world. The states of affairs they represent can be singled out with definitions by abstraction, using the synonymy conditions generated by their assertibility conditions. When meaning facts are construed in this way, the argument that Kripke finds in Wittgenstein does not establish that they don’t exist.

Most readings of the meaning skepticism Kripke ascribes to Wittgenstein understand it as metaphysical. The threat to meaning is supposed to follow immediately from the impossibility of citing facts in which meaning consists. I offer an alternative, epistemological, reading that is closer to Wittgenstein. What threatens meaning is the worry that, when I use an expression on any given occasion, I cannot know that my use conforms to previous uses of the expression; instead, in Wittgenstein’s terms, I go on “blindly,” without the understanding which is necessary for meaningful use. This reading makes for a stronger skeptical argument, in that it blocks the non-reductionist response of taking meaning facts to be primitive. But the argument, on this reading, can still be answered: not by citing meaning facts but by showing that I can know how to go on with an expression without needing to appeal to what the expression means.