I have a silly idea that one could derive what we think of as "sociology" from a model of individual human psychology after the fashion of statistical mechanics. Derived laws about behavior in bulk by analysing the distribution of individual behavior.