International Journal of Comparative Psychology

When analyzing animal behavior, it is important to consider the influence of learning principles. The create response of bottlenose dolphins, elicited by a discriminative stimulus, or an SD (visual cue presented to an animal by a trainer), has been described as an elective, often novel response based on arbitrary preferences of individual animals. The goal of this study was to identify the potential influence of reinforcement theory, response class, and primacy and recency on the create responses of bottlenose dolphins. Three, male subjects with an established mastery of the create paradigm, identified in this study as a non-specific, non-repeat contingency, were assessed over the course of two months while under stimulus control ( pre-assessment ), followed by evaluations of the create response ( create assessment ) using a double-blind sampling model. During the pre- and create assessments, each response was quantified regarding response class, frequency of request, and reinforcement type, frequency, and magnitude. When presented with the create SD, the dolphins elected to produce behaviors predominantly associated with the more recent training context (create assessment) versus behaviors associated with training that occurred months prior (pre-assessment), which may demonstrate the effects of primacy versus recency. Additionally, the create trials were associated with reinforcement on a high frequency and magnitude, fixed, low ratio schedule, and the subjects most often performed the behaviors associated with the greatest magnitude of primary reinforcement, which highlights the influence of reinforcement and the law of effects. Lastly, two subjects never responded with high energy behaviors in the create contingency, and one subject performed significantly more low and medium energy responses when compared to high energy behaviors, capturing the effects of a response class characterized by intensity under a fixed ratio reinforcement schedule. Thus, the create response was not represented by arbitrary elective preferences but rather, partially driven by the learning theories examined.