include ones that modify inherited attributes directly rather than indirectly by natural selection. These “Lamarckian” effects are added easily to models, and the models remain evolutionary so long as rationality remains bounded. The degenerate case, of course, needs no recursion because everything happens in the first generation (instantly in a typical rational choice model). Evolutionary models are a natural extension of the concept of bounded rational choice. They help explain how the innate and cultural constraints on choice and on rationality arise (Boyd and Richerson, 1993).
Evolutionary theory is always multilevel; at a minimum it keeps track of properties of individuals, like their genotypes, and of the population, such as the frequency of a particular gene. Other levels may also be important. Phenotypes are derived from many genes interacting with each other and the environment. Populations may be structured, perhaps divided into social groups with limited exchanges of members. Thus, evolutionary theories are systemic, integrating every part of biology. In principle, everything that goes into causing change through time plays its proper part in the theory.
This in-principle completeness led Mayr (1982) to speak of “proximate” and “ultimate” causes in biology. Proximate causes are those that physiologists and biochemists generally treat by asking how an organism functions. These are the causes produced by individuals with attributes interacting with environments and producing effects on them. Do humans use innate cooperative propensities to solve commons problems or do they have only self-interested innate motives? Or are the causes more complex than either proposal? Ultimate causes are evolutionary. The ultimate cause of an organism’s behavior is the history of evolution that shaped the gene pool from which our samples of innate attributes are drawn. Evolutionary analyses answer why questions. Why do human communities typically solve at least some of the commons dilemmas and other cooperation problems on a scale unknown in other apes and monkeys? Human-reared chimpanzees are capable of many human behaviors, but they nevertheless retain many chimp behaviors and cannot act as full members of a human community (Temerlin, 1975). Thus we know that humans have different innate influences on their behavior than chimpanzees, and these must have arisen in the course of the two species’ divergence from our common ancestor.
In Darwinian evolutionary theories, the ultimate sources of cooperative behavior are classically categorized into three evolutionary processes operating at different levels of organization.
Individual-level selection. Individuals and the variants they carry are obviously a locus of selection. Selection at this level favors selfish individuals who