Confirmation of Ecological and Evolutionary Models

This paper concerns hypothesis testing and confirmation in evolutionary and ecological theory. I outline specific criteria used in evaluating evidence for theories, and demonstrate the use of each criterion through examples from various branches of evolutionary biology and ecology. The philosophical discussions in Roughgarden (1983), Strong (1983), Simberloff (1983), and Quinn and Dunham (1983), which focus on a Popperian approach to theory testing and acceptance, present some important issues in the testing of evolutionary and ecological explanations. I find that imprecision of criteria of testing and confirmation is the weakest point in these discussions. As an alternative to a Popperian approach (as defended by, e.g., Simberloff 1983), and to other approaches commonly cited by biologists (e.g., J. Platt’s “strong inference”), I suggest a new description of confirmation that includes a detailed classification of the ways in which a theory may be confirmed (see, e.g., Oster and Wilson 1978 for an endorsement of Platt’s 1964 paper). Roughgarden (1983) proposes that one establishes an empirical fact in science “by building a convincing case for that fact.” What counts as a convincing case depends, according to Roughgarden, on “common sense and experience” (pp. 583–584). As Strong (1983) rightly points out, the appeal to common sense is problematic; common sense may not be “common” to all scientists concerned, and it says nothing about testing. Strong’s solution to problems of confirmation and testing is not much of an improvement, though: “our