ABSTRACT

INTRODUCTION In Chapters 5 to 7 we have argued that the defeasibility of everyday reasoning creates problems for most cognitive scientific approaches and, in particular, for theories in the psychology of reasoning. In this chapter we turn to two of our responses to counter-arguments to our position. The first section of this chapter (Mental Models and Defeasibility) is based on our response (Chater & Oaksford, 1993) to a paper by Alan Garnham (1993) directly arguing against our position presented in Chapters 4 and 5 from the perspective of mental models theory. The second section (Mental Logics and Defeasibility) is drawn from Oaksford and Chater (1995b) and responds to arguments often made by mental logicians (e.g. Politzer & Braine, 1991; Rips, 1994) to the effect that the human reasoning may not be defeasible because default rules are always false. They therefore claim that the apparent defeasibility of human inference is not to be explained by a theory of inference, but by pragmatic or performance factors. We argue that neither of these defences against the problems that we raised for logicist cognitive science is successful.