The social structure of any species is an emergent biological phenomenon and as such it has an evolutionary history. The human social structure is no exception to that rule but it has an important peculiarity: it is hidden from view by its numerous cultural expressions. To circumvent the problem and characterize the deep structure of human society one must carry out a comparative analysis of human and nonhuman primate societies and employ evolutionarily significant categories. Using that approach I define human societies as nested associations of multifamily groups, a structure made up of a specific set of features, notably, strong ties between groups stemming from the linkage of kinship bonds and pair bonds, a uniquely human trait. I also argue that primatology makes it possible to define a five-step model of the maximally parsimonious evolutionary sequence that led to human social structure. Finally I show how the present phylogenetic perspective informs functional analyses of human behavior by pointing to the ‘adaptive suite pitfall’ and the importance of phylogenetic constraints.