What influence do systems of classification have on the construction of world-views? How far do such systems form the basic and unchallenged presuppositions of such world-views, or to what extent are they the subject of conscious reflection and criticism on the part of those who use them? Classificatory systems have been extensively studied by social anthropologists, cognitive scientists, philosophers, linguists, historians. What light can an examination of their debates throw on the fundamental philosophical issues we have been tackling throughout this set of studies, namely the conflict between realism and relativism, the commensurability of belief systems, and the relation between science and popular belief? The comparative data of classifications both of natural and of cultural phenomena can be used to provide further illustrations of the multidimensionality of reality and of the diversity of styles of enquiry. This will allow us to follow up and elaborate the arguments advanced in the last four chapters on the problems of a common logic, the search for truth, the challengeability of belief, and a common ontology.
We may start with a well-known controversy, one that stems from a dilemma, or at least from two conflicting intuitions. The first is that natural kinds are cross-cultural universals. What 'natural kinds' include may be disputed: but certainly the biological ones, animals and plants, may be taken as paradigmatic. So the cross-cultural universalist view would insist that lions and tigers, for instance, are lions and tigers in zoos the world over.
But the second intuition is that classifications of natural kinds spring from, or are the work of, culture. That was a prominent theme in Durkheim and Mauss (1901-2/1963), and more recently Tambiah (1969) put it, following Lévi-Strauss (1962/1969), that animals are good to think—that is, good to think with. This view insists, then, that the ways in which animal codes are used to think about other things are enormously diverse world-wide.