What kind of concept do we want for conceptual engineering?

Bojin Zhu

University of Vienna

In the past decade, conceptual engineering (CE) has become a topic of attention. The idea is that apart from analysing our concepts, we could also actively take control of them, assess them, and improve upon them. However, there is much debate in the CE literature on what concepts are, and philosophers seem to disagree from the very beginning. This paper makes two contributions to resolving this debate within the context of CE.

First, inspired by Tim Crane’s treatment of propositions (of which concepts are constituents), it advocates a methodological view that treats concepts as models, in the sense of scientific models (instead of model theory). According to this view, concepts are metaphysically lightweight; they are models we adopt to better make sense of people’s behaviours (in a very broad sense). Models are crucially incomplete, imperfect, and purpose-relative. Hence we should treat different accounts of concepts as each modelling a different aspect of our cognitive life and behaviours, without saying one of them is uniquely correct. More importantly, in the case of CE, it recommends a methodological shift: we don’t first decide what concepts are and then come to CE; we take conceptual changes, both natural and philosophical, as our phenomena (as obviously they do happen), and ask what model of concepts allows us to best account for such phenomena.

With this methodological point in mind, the paper then argues that inferentialism acts as a much better model for CE than representationalism, both very broadly construed. Bracketing the debate between the two in a larger context, the current point is a narrower one: it only says that, given the purpose of modelling CE, treating concepts as inferential devices rather than intensions and extensions helps us make much better sense of the aim and practice of CE. That is, one need not adopt inferentialism in general to adopt it as a good model of CE. Crucially, for most cases of CE, in-/extensions are not the kind of thing that could be better or worse. Instead, inferential connections are: they are what conceptual engineers aim to improve. In the end it also shows that inferentialism naturally gives us an answer to the problem of topic continuity in CE, and how, for this purpose, it improves upon both Herman Cappelen’s notion of ‘topic’ and Amie Thomasson’s and Sally Haslanger’s notion of ‘function’.

Chair: tba

Time: September 12th, 10:00 – 10:30

Location: SR 1.005