Quasi-independence, fitness, and advantageousness

I argue that the idea of 'quasi-independence' [Lewontin, R. C. (1978). Adaptation. Scientific American, 239(3), 212-230] cannot be understood without attending to the distinction between fitness and advantageousness [Sober, E. (1993). Philosophy of biology. Boulder: Westview Press]. Natural selection increases the frequency of fitter traits, not necessarily of advantageous ones. A positive correlation between an advantageous trait and a disadvantageous one may prevent the advantageous trait from evolving. The quasi-independence criterion is aimed at specifying the conditions under which advantageous traits will evolve by natural selection in this type of situation. Contrary to what others have argued [Sterelny, K. (1992). Evolutionary explanations of human behavior. Australian Journal of Philosophy, 70(2), 156-172, and Sterelny, K., & Griffiths, P. (1999). Sex and death. Chicago: University of Chicago Press], these conditions must involve a precise quantitative measure of (a) the extent to which advantageous traits are beneficial, and (b) the degree to which they are correlated with other traits. Driscoll (2004) [Driscoll, C. (2004). Can behaviors be adaptations? Philosophy of Science, 71, 16-35] recognizes the need for such a measure, but I argue that she does not provide the correct formulation. The account of quasi-independence that I offer clarifies this point.