This chapter provides an in-depth discussion of some key aspects of Sellarsian inferentialism, addressing two main problems. The first problem is the language–world relationship: if meaning is the use of an expression, how exactly are extra-linguistic objects relevant to the meaningfulness of expressions? The correct use of an expression entails its use in the appropriate non-linguistic circumstances, but the correlation between a linguistic predicate and an object, even when ostensible, must not be understood as a meaning relation.The second problem concerns the ontological status of Sellarsian linguistic notions, such as roles and functions. Given that Sellars is a nominalist, how can we avoid concluding that he introduces new kinds of universals: linguistic? Although inferentialism requires the ability to categorise expressions, categorisation can be successfully done using linguistic distributive sortals, which translate property and kind talk into the talk of linguistic tokens. Therefore, there is no need for Platonic or linguistic universals. Explanation through human practice is sufficient.