There is a form of reduction in genetics that attempts to explain phenotypic properties from a genotypic basis without attributing any particular structure to the genotype. Therefore, if successful such a reduction would be weak [that is, of type (a) from Chapter 3, § 3.2]. It consists of the use of concepts of “heritability” as measures of genetic influence. This technique has been routinely invoked for many putative human traits including IQ (which has had a long and occasionally sordid history), vocational interests, “religiosity,” “openness,” “agreeableness,” “conscientiousness,” neuroticism, and extroversion. The basic idea is that if some trait has a high heritability, its origin or, at the very least, why it varies from individual to individual can be explained from a genetic basis. The stronger claim is the more interesting one. Unfortunately, it is not even a distant approximation to the truth. However, the weaker claim is occasionally plausible. These are the two basic points that this chapter will make. These points are nothing new; however, this chapter will attempt to synthesize the past conceptual analyses of heritability.

The roots of heritability analysis go back to what was called “biometry,” largely the work of Galton and, especially, Weldon and Pearson in the 1890s, which partly explains why it is relatively neutral with respect to the details of the structure of the genome. After Mendelism was recovered around 1900, starting with the work of Yule (1902) and ending with a classic paper by Fisher (1918), biometry was reduced to Mendelism.