Some ten years ago, I reflected in a series of papers (1972a; 1972b; 1972c; 1974) on the limitations of cognitive anthropology as then practiced: “ethnoscience,” “ethnographic semantics,” “the new ethnography.” I argued that in its preoccupation with inductive rigor and cultural uniqueness, cognitive anthropology had followed the premises of American descriptive linguistics, even as the Chomskyan revolution had made these premises untenable in the realm of language. Analyzing the semantics of folk classification on the promise that the methods and models so developed could be extended so as eventually to produce a “cultural grammar” was, I thought, naïve. The focus on artificially simple and simplified problems (especially the semantics of kin terms in their genealogical senses) had deflected our attention from the deep complexities of meaning and context and deep questions about the rule-governedness of social behavior. I pointed to the incredible complexity of the organization of mind and brain, as then beginning to emerge in the neurosciences and artificial intelligence research. In this light, anthropologists' attempts to describe cultural knowledge seemed curiously unrealistic: