There are two readings of how Sellars analyzed ought-statements: a first-order expressivist reading and a metalinguistic reading. In this chapter, I argue that unpublished draft material related to the creation of a key text in Sellars’s practical philosophy, the last chapter of Science and Metaphysics, shows that Sellars at least seriously considered adopting a metalinguistic analysis of “ought.” I also explore the relation between these exegetical questions and the broader problems raised by Sellars’s thought. Which reading of Sellars on “ought” we adopt can affect our outlook on the place of norms in Sellars’s scientific naturalism. I argue that a metalinguistic reading of “ought” permits the view that norms can become perfectly implicit at Sellars’s ideal end of inquiry.