We argue that a theory of the evolution of Empathizing (E) and Systemizing (S) needs first to clarify that these are personality traits, as distinct from cognitive abilities. The theory should explain both the observed reciprocity of, and the sexual difference between, E and S in a context of the historical emergence of these traits and their balance in relation to local selection pressures. We suggest that the baseline state is that (since humans are social animals) ancestral human hunter gatherers are assumed to be relatively High Empathizers, lower in Systemizing: thus more interested in people than in things. Changes related to the development of agriculture and technology meant that it became economically useful for some men to become more interested in ‘things’ than in people, as a motivation for them to learn and practice skills that were vital to personal and (secondarily) social survival, reproduction and expansion. This selection pressure applied most strongly to men since in the sexual division of labour it was typically men’s role to perform such tasks. We further hypothesize that High Systemizing men were rewarded for their socially vital work by increased resources and high status. Because marriages were arranged in traditional societies mainly by parental choice (and the role of parental choice was probably increased by agriculture), it is presumed that the most valued women, that is young and healthy women thereby having high reproductive potential, were differentially allocated to be wives of economically successful High Systemizers. Such unions of economically successful High Systemizing men with the most reproductively valuable women would be expected to lead to greater-than-average reproductive success, thereby amplifying the population representation of genes that cause high systematizing in the population. This hypothesis makes several testable predictions.