Lewontin's Fallacy

From Wikipedia, the free encyclopedia

If height or weight (or single genes) are measured alone, the red and blue populations here would overlap strongly. If both traits are measured, however, natural clusters emerge.

Human Genetic Diversity: Lewontin's Fallacy is a 2003 paper by A.W.F. Edwards that criticizes Richard Lewontin's 1972 conclusion^[1] that because the probability of racial misclassification of an individual based on variation in a single genetic locus is approximately 30%, race is an invalid taxonomic construct.

Edwards argues that while Lewontin's statements on variability are correct when examining the frequency of specific loci between individuals, when one takes into account more loci, the probability of racial misclassification rapidly approaches 0%, due to the correlation of loci frequencies within a population. (See Figure 1 of the paper.) Edwards argues that the information which distinguishes races is "hidden in the correlation structure of the data."

Edwards argues that both ordination and cluster analyses can reveal the correlation structure of multilocus data.

A caricature of Lewontin's argument is that because humans share 50% of their DNA with carrots, we must be 50% the same. Lewontin certainly made no such argument, discussing the variability between groups based on only the variable DNA between them, not the absolute measure of all DNA (humans being 99.9% identical by DNA, but certainly not 99.9% identical by traits such as height or weight).

A caricature of Edwards' argument is that because we can measure enough loci (the entire DNA sequence) to make us all have our own individual "cluster", we must all be of different races (barring identical twins).

Whether or not the Fallacy is a fallacy depends on the question being asked.^[2] If differences between populations are regarded as "real" whenever the two populations can successfully be distinguished by use of a large number of arbitrary traits, then the Fallacy is a fallacy and Edwards is right and Lewontin is wrong. If the differences between populations are instead considered "real" according to the extent to which they differ in a randomly chosen trait, then Lewontin's argument is persuasive and there is no Fallacy.

[edit] References

^ Made in The apportionment of human diversity (1972) and again in the 1974 book The Genetic Basis of Evolutionary Change.
^ Chakraborty, R. (1982). "Allocation versus variation: The issue of genetic differences between human racial groups.". American Naturalist 120: 403-404.