Your 15 March issue honouring Carl Linnaeus brings to mind what is probably his most significant contribution to modern life: the idea that groups of people can be regarded as naturally distinct taxonomic entities, or subspecies, in the same fashion as species, genera and higher categories.

In the first edition of Systema Naturae, published in 1735, before formalizing binominal species nomenclature, Linnaeus presented humans as sorting naturally into whitish Europeans (Homo Europaeus albescens), reddish Americans, dark Asians and blackish Africans. By the 10th edition, in 1758, these had become subspecies, colour-coded as red Americans, white Europeans, yellow Asians and black Africans. In addition, Linnaeus separately listed wild children (Homo sapiens ferus) and a non-geographical grab-bag category, Homo sapiens monstrosus.

The idea that humans can be understood as constituting natural taxonomic units has bedevilled anthropology ever since. In the eighteenth century, both Johann Friedrich Blumenbach and Georges-Louis Leclerc, Comte de Buffon, recognized that the principal empirical patterns of human diversity are geographically gradual, which frustrates the project of human taxonomy. We would now say that pattern is 'clinal'1.

Further, as anthropology matured, it clarified the fact that human groups principally differentiate themselves culturally – that is, by language, dress, principal deities, taboos and the like. The strongest antagonists are not the most biologically different, but simply the worst neighbours. Consequently, perceptions of group difference are local, political and ephemeral; but are nevertheless still commonly regarded as natural — witness the racialization of categories such as 'Hispanic' and 'Middle Eastern' in the United States, and the newsworthy discovery in the United Kingdom that the Irish and the Scots are not so different after all2.

Genetics has been multivocal on the subject. On the one hand, it has emphasized the extensive polymorphism in the human gene pool3, showing that there are all kinds of people everywhere — as fieldworkers had long known, but without quantitative support. On the other, it has focused on the small component of genetic variation that differs the most in the most widely separated peoples, and commonly interpreted the results in racial terms4.

It has taken two and a half centuries to shed Linnaeus's fallacy that the human species comes taxonomically organized into a few large, natural groups that are fairly homogeneous and fairly distinct from one another. We have come to understand, rather, that the predominant patterns of human variation are cultural, polymorphic, clinal and local.

This does not mean that everyone is the same, or that there is no biogeographic differentiation within our species. It means just that the effort to treat our own species taxonomically has considerably more social and symbolic than biological meaning.