Technical Report #163

Two Reports on Category Learning in Baboons and Humans

Category Learning in Baboons and Humans, Talk presented at the 88th Annual Meeting of the Southern Society for Philosophy and Psychology.

Kruschke, J. K., Fagot, J. & Vauclair, J.

Abstract

Kruschke (1996, JEP:LMC, 22, 3-26) argued that the "inverse base-rate effect" in human category learning is caused by human learner's ability rapidly to shift attention to distinctive features of novel categories. We investigated whether baboons would exhibit similar behavior in a comparable category learning design. The baboons did not show categorization behavior comparable to humans, which we interpret as indicating that the baboons did not shift attention as humans did. The ability to shift attention away from previously learned features facilitates preservation of prior learning and rapid acquisition of new categories, and such attentional abilities may be a hallmark of human cognition.

Assessment of the Inverse Base-Rate Effect in Humans (Homo Sapiens) and Baboons (Papio papio).

Fagot, J., Depy, D., Vauclair, J. & Kruschke, J. K.

Abstract

Medin and Edelson (1988) demonstrated that, under certain experimental conditions, people may select a rare category when tested with conflicting cues whose actual diagnosticity is equal for rare and common categories. Here, this inverse base-rate effect was investigated with people and baboons. Using a video-formatted conditional matching-to-sample task implying joystick manipulation, subjects sequentially learned two two-feature categories which shared a common feature. People encoded the two features of the initially learned category, and predominantly the distinctive feature of the subsequently learned category. By contrast, although both features of the first category were initially encoded, the baboons ultimately retained only the distinctive features of each category. It is proposed that the effects of base rates on categorization depend on the ability to propositionally represent relations between the features defining the categories.