UC Santa Cruz

Abstract

Alexis D. Mourenza

(Dis)Appearing Minds: Methodological Assumptions and Epistemological Biases in Animal Behavior and Cognition Research

A debate in the nonhuman animal cognition literature exists between those providing ‘associative’ accounts and those providing more ‘mentalistic’ explanations of the cognitive mechanisms underlying nonhuman animal behavior. The former maintain that all nonhuman animal behavior can be explained as the result of conditioned associations or innate reflexes and the latter propose that some nonhuman animals make use of higher levels of mental representation. One area of inquiry that has received significant attention asks whether humans are unique in their ability to form representations of another’s mental states, that is, whether all other animals rely on strictly associative mechanisms that enable a kind of “behavior-reading” rather than “mind-reading.” I will propose that the naturalistic tasks put to the subjects allows the formulation of seemingly contradictory conclusions to be drawn from the experimental results. By examining only species-typical traits in a naturalistic setting, researchers are confronted with the possibility that the observed behaviors result from trained associations acquired in the animal subject’s individual learning history or stimulus-bound relations acquired in the evolutionary history of the species, rather than resulting from the subject’s ability to consciously reason about it’s environment (including the other individuals that are part of it).

Richard Samuels and Stephen Stich have proposed that reasoning in humans is underwritten by two distinct neuronal systems, one that is uniquely human and is constituted by a number of higher-order, domain-specific mental modules and another that is shared with other animals and is constituted by a set of cognitively simpler, domain-general problem-solving skills, like association. Ronald Schusterman and colleagues’ demonstration of the eventual acquisition of an equivalence concept by their sea lion subject, Rio, in a nonnaturalistic experimental paradigm will be provided as a case-study for the attribution of a dual-processing cognitive system to a nonhuman animal.

Association is purported to be able to account for learning, the structure of mental states, and the way certain thoughts relate to other thoughts. But Eric Mandelbaum has shown that accepting an associative account of one of these mental processes does not entail that we must also accept associative accounts of the others. The primate theory of mind literature will be reexamined in light of a more thorough understanding of the varieties of associationist theses and the conflation of them. I will propose that the burden of proof should be lifted from those who seek to explain numerous diverse complex behaviors in a nonhuman animal species by reference to higher-order cognitive mechanisms when doing so accounts for the larger body of data, rather than advising them to explain a single experimental result in isolation by postulating a complex web of conditioned associations for which no evidence is available to validate the assumption that such associational learning has taken place. Further, ecologically-invalid, nonnaturalistic experimental paradigms eliminate such assumptions from consideration and thereby provide the strongest evidence of higher-order cognitive mechanisms operating in a nonhuman animal subject.