Skip to main content


"Inductive risk in framework shifts: the case of cultural evolution", Azita Chellappoo (University of Cambridge)

Egenis seminar series

Egenis seminar series. Non-epistemic values have been long-acknowledged to play a significant role in scientific inquiry: for example, in problem selection, and directing the use of scientific knowledge. Douglas (2000) provides a widely-applied account of another avenue for non-epistemic values to play a legitimate role: inductive risk. Inductive risk refers to the risk involved with the acceptance or rejection of a hypothesis: in the decision whether to accept a given hypothesis or not, there is always the risk of either accepting a false hypothesis (a Type 1 error, or ‘false positive’) or rejecting a true hypothesis (a Type 2 error, or ‘false negative’). When these errors have non-epistemic consequences, non-epistemic values will influence the ‘rule of acceptance’ (the level of evidence or statistical significance required to accept the hypothesis).

Event details

I argue that kind of inductive risk is present, not only when considering isolated hypotheses, but also when we consider whether to adopt or pursue a new framework or approach more generally. In particular, there is inductive risk in the adoption and pursuit of the cultural evolution project as a whole, and of certain research programmes within cultural evolution. We are faced with a choice about whether to adopt an evolutionary approach, which carries with it possibilities of ‘false positives’ or ‘false negatives’, in a similar way to the decision whether to accept or reject a hypothesis.

The decision to adopt these frameworks is partly dependent on the balance we are willing to strike between ‘false positives’ and ‘false negatives’, and as these errors have social and political consequences, social and political values enter into the calculation. I argue that consideration of these consequences means that we should exercise caution, and aim to clarify our explanatory goals and expected gains before applying cultural selection frameworks and models further.


Byrne House