Forward-Looking Concept Functions and the Function/Accident Distinction

Appealing to the functions of concepts in conceptual engineering projects has been hailed as a promising avenue to explain defects of concepts, the possibility of amelioration, and the continuity of topic post engineering. Some of the most prominent theories of concept functions build on what I call the Usefulness Thesis, i.e. what makes something the function of a concept is what makes it useful. In this paper, I argue that this approach does not meet some of the most fundamental adequacy criteria for a theory of function, that is, (1) they do not draw a stable function/accident distinction and (2) they cannot account for the normativity of functions. I show that theories of concept function that build on the Usefulness Thesis are all essentially forward-looking theories of functions and show that these theories face a problem of overreach and a vacuous normativity proviso. I further argue that they do not lend themselves to conceptual critique, so they are not fit for conceptual engineering purposes.