September 01, 2006

Why we need more anthropometry

Kerim of Savage Minds, an anthropological blog, writes a post titled Anthropometry: Alive and Kicking. In it, he dismisses the 1988 All India Anthropometric Survey, North Zone : Basic Anthropometric Data, although it is not at all clear why he objects to anthropometry. I guess, anthropometry is what those bad old colonial racists did, so dismissing it requires no justification.

Anthropometry is the measurement of the human body: its shape, size, color; the collection of quantitative data about human beings. Anthropology is the study of mankind. Without anthropometric data there is no anthropology, or at most there is something like modern social or cultural "anthropology", a data-lite, theory-heavy field which generates papers with titles like "Coetzee, Agamben, and the Passion of Abu Ghraib", "The End of the Soul: Scientific Modernity, Atheism, and Anthropology in France", "Piro Designs: Painting as Meaningful Action in an Amazonian Lived World", "Agency, Personhood and the 'I' of Discourse in the Pacific and Beyond"

It is often erroneously claimed that modern science has superseded traditional physical anthropology. The old typologists, it is said, were misguided, their conclusions wrong, their caliper-wielding methods naive, or even dangerous, leading to racism, intolerance, genocide, and assorted very bad things.

On the contrary, many of the conclusions of traditional physical anthropology, such as the taxonomic division of Homo sapiens into at least four subspecies have stood the test of time, having been amply confirmed by genetic studies, as well as the multivariate analysis of craniometric and anthropometric data.

The old physical anthropology was simply limited by its lack of computational power and statistical methodology. Most of it was done before the pioneering work of R. A. Fisher's discriminant functions, before the use of classification trees, clustering algorithms, or Principal Components Analysis. Without computers, multivariate analyses of large samples were unthinkable. Anthropologists relied on their expertise, and on simple comparisons of means and standard deviations of single traits. Nonetheless, at least for the major problems, their conclusions were essentially correct.

Nowadays, computational power is abundant, statistical methodology is well established, and tools for collecting data about human personality, intelligence, genes, and appearance are there for anyone to use. Unfortunately, though, no one is using them. The Human Genome Diversity Program was shot down, and the Genographic Project, privately funded, is a much toned down version, which aims only to reconstruct human origins.

As a practical matter, anthropologists could easily do a comprehensive study of human variation. Imagine a study where 100,000 people from around the world are given a thorough anthropometric measurement, an IQ test, a personality test, a number of medical tests, and a comprehensive genetic test which can be refined down the road as gene sequencing becomes cheaper. Imagine thousands of attributes, phenotypic or genotypic, for each one of the 100,000, the first large sampling of human variation in all its measurable aspects. Such a database wouldn't cost much, but would allow us to really speak authoritatively about human variation.

What is stopping us from doing such a study? Bizarre notions about the rights of indigenous groups, a dogmatic rejection of any correlation between physical (especially racial) traits with morality, personality, or cognitive function, a pre-occupation with medical applications and pathological variation.

We need more anthropometry, not less. More psychometry too, more genetic testing, more data of every kind. Above all, more willingness to study man quantitatively, rather than a reliance on politically-correct preconceptions about ourselves.

No comments: