The challenges of mixing associational learning theory with information-based decision-making theory

Behavioral ecologists frequently incorporate associational learning (AL) concepts into studies of choice behavior. Within behavioral ecology, AL is often considered a mechanism for information gathering. AL also provides alternative explanations of behavioral phenomena up to the level of motivational organization over the lifetime. AL assumes that all inputs to the learning system interact through a multistep process with feedbacks to control behavior and that cues are characterized by contingencies, whereas behavioral ecology assumes that learning inputs independently control responses, are in conflict, and convey information. Integrating the 2 perspectives is not straightforward and can lead to conflicting predictions or loss of predictive power. I examine 2 sets of case studies. First, I look at parallel research programs on mating in quail. Second, I consider how AL concepts have been integrated into foraging studies of nectarivores. The papers on quail mating demonstrate that to a large degree, the 2 approaches explain similar behaviors in compatible ways. The nectarivore papers show how the theories diverge, with AL predicting challenging results. Future studies should examine how much individuals select between sources of information and how much they respond to combinations of and interactions between cues within the process described by AL, using experimental designs that allow explicit cross-paradigm comparisons through the use of identical measurements of response.