This sounds mostly right, and it’s concordant with research in evolutionary psychology over the last 35 years.
Previous applications of evolutionary theory to behavior (before the 1980s) did often model animals and humans as if they were trying to maximize inclusive genetic fitness (IGF) -- but this was usually considered a heuristic over-simplification of how animals actually make decisions (in foraging, mating, predator avoidance, parental investment, etc). This IGF-maximizing model was often useful in behavioral ecology, evolutionary game theory, optimal foraging theory, and sociobiology—but nobody took it very seriously as a functional description of how animals make decisions, and only the most naive researchers assumed that animal brains include little utility functions that directly represent IGF.
The key innovation in evolutionary psychology, in the late 1980s, as pioneered by Leda Cosmides and John Tooby, was to explicitly reject IGF-maximization as a description of animal/human decision-making, and to replace it with the notion of domain-specific psychological adaptations that evolve to handle particular aspects of fitness. In other words, we’re adaptation-executors, not fitness-maximizers. The psychological adaptations typically track ‘fitness affordances’ (e.g. food, mates, territories, offspring, kin) that are statistically associated with fitness, rather than trying to track fitness itself—much as AIs might show instrumental convergence onto generally useful instrumental goals.
This ev psych perspective makes it much easier to explain cases of ‘evolutionary mismatch’, where the modern environment doesn’t match the ancestral environment, so our adaptations might not work very well any more.
This sounds mostly right, and it’s concordant with research in evolutionary psychology over the last 35 years.
Previous applications of evolutionary theory to behavior (before the 1980s) did often model animals and humans as if they were trying to maximize inclusive genetic fitness (IGF) -- but this was usually considered a heuristic over-simplification of how animals actually make decisions (in foraging, mating, predator avoidance, parental investment, etc). This IGF-maximizing model was often useful in behavioral ecology, evolutionary game theory, optimal foraging theory, and sociobiology—but nobody took it very seriously as a functional description of how animals make decisions, and only the most naive researchers assumed that animal brains include little utility functions that directly represent IGF.
The key innovation in evolutionary psychology, in the late 1980s, as pioneered by Leda Cosmides and John Tooby, was to explicitly reject IGF-maximization as a description of animal/human decision-making, and to replace it with the notion of domain-specific psychological adaptations that evolve to handle particular aspects of fitness. In other words, we’re adaptation-executors, not fitness-maximizers. The psychological adaptations typically track ‘fitness affordances’ (e.g. food, mates, territories, offspring, kin) that are statistically associated with fitness, rather than trying to track fitness itself—much as AIs might show instrumental convergence onto generally useful instrumental goals.
This ev psych perspective makes it much easier to explain cases of ‘evolutionary mismatch’, where the modern environment doesn’t match the ancestral environment, so our adaptations might not work very well any more.