Change in looping effects

This paper seeks an account of what types of reactions to being classified can trigger a demand to revise the original classifications and therefore instigate a feedback loop According to one view, such looping effects happen if and only if classifying a kind changes which properties are correlated with kind membership (correlationism). According to another, they happen if and only if classifying a kind changes its essential properties (essentialism). I argue that correlationism implied looping effects where there are none and that essentialism either collapses back into correlationism (on a broad construal of “essential property”) or implies that looping effects are impossible. In response, I develop and defend a view according to which a looping effect happens if and only if the effects of classifying a kind depend, in part, on the existing nature of that kind (explanationism). I show that explanationism forges a path between the above extremes and that it offers new insights about old examples of looping effects, not least about the sort of evidence we should need to justify the claim that a looping effect has occurred or that a kind is susceptible to looping effects.