Abstract
Conceptual engineering (CE) is sometimes presented as (A) a radically novel metaphilosophical stance; (B) either superseding or providing the foundations for all other philosophical projects. My presentation takes issue with both claims. Its objects of comparison are, on the one hand, a venerable discipline, namely metaphysics, on the other hand, a recently rehabilitated method, namely conceptual analysis (CA).
Concerning (A), I shall argue that (i) the older label of concept formation is more apposite, since it includes ways of introducing and modifying concepts, notably in science, that are driven by the attempt to capture extant features rather than the ambition of engineering something novel to suit one's own practical purposes; (ii) traditional metaphysics shares the revisionary impetus of CE, while insisting that our notions should mirror the nature of reality; (iii) CE arises historically out of CA, and, as in the case of experimental philosophy, moderate versions of it (logical explication) are close to liberal versions of CA ('impure CA').
Concerning (B), improving conceptual schemes in a controlled and fruitful way prima facie presupposes an understanding of the conceptual status quo. Consequently, dismissive condemnations of conceptual analysis by conceptual engineers constitute a paradox that needs to be taken far more seriously than has been the case. One reason why the paradox has escaped notice is even more paradoxical: the most vocal and influential champion of CE, Cappelen, does not believe in the existence of concepts, let alone the possibility of changing them. By employing the tools of CA, I will show that this position is untenable. At a minimum, projects trading under the label of CE must aim at improving the way we categorize and reason about phenomena, rather than altering our specific beliefs and theories (as in scientific theory-formation). This enterprise presupposes a semantic self-understanding of the kind provided by a version of CA that is both connective and impure: it elucidates concepts not in isolation, but through their interconnections and their applications outside of philosophy. My arguments for this conclusion will draw on a loosely speaking pragmatist semantics. Yet this does not constitute a petitio principii, since that semantics captures the established notions of meaning and understanding that are presupposed even in semantic revisionism.
