ABSTRACT
In this paper I shall argue that cross–cultural data is weak, but that it is valuable. It is valuable, furthermore, because it is weak. Such data, I shall contend, presents us with problems of inference whose solution has a more general validity than might at first sight appear plausible.
