In prior work, we characterized two-sex marriage functions for socially structured populations as multiplicative perturbations of heterosexually random/proportionate mixing. These perturbations were expressed in terms of the preferences/ affinities of males for females and vice versa. Male and female preferences/affinities are obviously not independent as they depend on the availability of male and female behavioral "genotypes." We show that knowledge of the preferences/affinities of one gender can characterize the preferences/affinities of both genders m socially-structured populations; in other words, it takes two to tango. This is the basic content of the T 3 Theorem. In this chapter, we revise our results for socially structured populations and extend them to situations where the population is characterized by continuous variables such as age. It is shown that different sets of preferences/ affinities, that is, distinct behavioral "genotypes", may give nse to identical mixing/mating probabilities, the determinants of the behavioral "phenotypes." Hence, different sets of individual decisions can lead to identical social dynamics -a fact well established in genetics. The importance of the incorporation of mating systems at the population level is a neglected but central area in evolutionary biology.-2-
L IntroductionMarriage functions are solutions to the two-sex mixing/pairing problem.Despite their importance m areas such as population genetics (mating functions), demography (population projection), cultural anthropology (preservation and dissemination of cultural traits), and evolutionary biology (life history), their application has been quite limited. Most researchers have addressed theoretical issues in these areas through the use of single-sex models or highly simplified two-sex models. A basic premise being ignored is that "it takes two to tango." The difficulties involved are quite evident from the pioneering work of Kendall (1949), Keyfitz (1949), Fredrickson (1971), McFarland (1972), Parlett (1972), Pollard (1973), and Caswell and Weeks (1986.We have developed an axiomatic framework to conduct a systematic study of marriage functions Castillo-Chavez 1989, 1991;Castillo-Chavez and Busenberg 1991;Hsu Schmitz 1994;Hsu Schmitz et al. 1994, Blythe et al. 1991. Our work has been applied in areas as diverse as cultural and parameter estimation , Hsu Schmitz and Castillo-Chavez 1994, 1995.We provide a summary of our characterization of marriage functions for populations defined through fixed characteristics such as race, language, biological species, religion, level of education, and socioeconomic level. We then provide a detailed characterization of age-structured marriage/mixing functions. The discrete framework described in this chapter can be incorporated into finite dimensional deterministic or stochastic models while the continuous framework is easily incorporated into age structured models.In earlier work, we (Castillo-Chavez and Busenberg 1991) characterized two-sex marriage functions as mul...