INTRODUCTION

Dennett's (1971, 1987) ideas about the intentional stance have elicited different reactions from philosophers and scientists. Philosophers have focused on Dennett's alleged anti-realism about the mental while scientists have focused on his three-way distinction among zero-order, first-order, and second-order intentionality. For most philosophers, anti-realism (at least concerning the mental states of human beings) is a no-no; for most cognitive scientists, the three-way distinction is useful. This contrasting reaction is typical of a larger pattern: philosophers are more inclined to cite work with which they disagree while scientists are more inclined to cite work on which they wish to build.

Dennett (1987, 1991b) has denied that he is an anti-realist, but that has not stopped philosophers from continuing to affix a scarlet letter “A” to his work. Nor has the specter of anti-realism stopped cognitive ethologists from using the three-way distinction to articulate a central methodological principle, which they call “the principle of conservatism.” According to this principle, hypotheses that explain an organism's behavior by attributing lower-order intentionality are preferable to hypotheses that explain the behavior by attributing higher-order intentionality (see, for example, Cheney and Seyfarth [1990, 2007]).

My subject here is the principle of conservatism. What, exactly, does it say and what is its justification? Cognitive scientists often regard the principle as an instance of a more general methodological maxim, namely the principle of parsimony, a.k.a. Ockham's razor.