Abstract  Formulations of the essential commitment of the etiological theory of functions have varied significantly, with some individual authors' formulations even varying from one place to another. The logical geography of these various formulations is different from what is standardly assumed; for they are not stylistic variants of the same essential commitment, but stylistic variants of two non-equivalent versions of the etiological theory. I distinguish these strong and weak versions of the etiological theory (which differ with respect to the role of selection in their definitions of function), draw out their respective implications, and argue that the weak version is to be preferred to the strong.