ABSTRACT
Adaptationist approaches in evolutionary ecology often take it for granted that natural selection maximizes fitness. Consider, for example, the following quotations from standard textbooks:
The majority of analyses of life history evolution considered in this book are predicated on two assumptions: (1) natural selection maximizes some measure of fitness, and (2) there exist trade-offs that limit the set of possible [character] combinations.
(Roff 1992: 393)The second assumption critical to behavioral ecology is that the behavior studied is adaptive, that is, that natural selection maximizes fitness within the constraints that may be acting on the animal.
(Dodson et al. 1998: 204)Individuals should be designed by natural selection to maximize their fitness. This idea can be used as a basis to formulate optimality models.
(Davies et al. 2012: 81) Yet there is a long history of skepticism about this idea in population genetics. As A.W.F. Edwards puts it:[A] naive description of evolution [by natural selection] as a process that tends to increase fitness is misleading in general, and hill-climbing metaphors are too crude to encompass the complexities of Mendelian segregation and other biological phenomena.
(Edwards 2007: 341) 50Is there any way to reconcile the adaptationist’s image of natural selection as an engine of optimality with the more complex image of its dynamics we get from population genetics? This has long been an important strand in the controversy surrounding adaptationism. 1 Yet debate here has been hampered by a tendency to conflate various different ways of thinking about maximization and what it entails. In this chapter I distinguish, at a deliberately coarse grain of analysis, four varieties of maximization principle. 2 I then discuss the logical relations between these varieties, arguing that, although they may seem similar at face value, none entails any of the others. I then turn briefly to the status of each variety, arguing that, while each type of maximization principle faces serious problems, the problems are subtly different for each type.