Association for Behavior Analysis International

Social behavior is a topic of enormous scientific importance that spans disciplines from neuroscience to anthropology. And while the topic has received much empirical and theoretical attention, it has been virtually neglected in behavior analysis. The goal of this symposium is to bring together several lines of research on social behavior in non-human animals. Zentalls presentation will focus on varieties of imitation, including the proper control conditions needed to separate flexible and deferred imitation from simpler forms of imitation. Velasco and colleagues will present data on cooperative behavior with pigeons, comparing behavior under individual contingencies that under interlocking social contingencies. Serrano and colleagues will present data on a cooperation task, in which rats choose between working individually and working collectively. Hackenberg & Tan will present data with rats in a mutual cooperation task requiring collective responses. Barrett will discuss the presentations from a comparative-evolutionary perspective, situating the work within the field of comparative cognition, and identifying promising areas for future research.

In a prisoner’s dilemma game (PDG), two subjects are paired together and given the opportunity to “cooperate” or “defect”. The immediate consequences of each choice always favor defecting over cooperating (at least, for the individual that is defecting). However, if both players defect, the consequences are worse (for both) than if both had cooperated. The probability of reciprocation in a PDG determines the extent to which the long-term contingencies favor cooperation over defection. With perfect reciprocation, cooperation in the current trial should be reinforced by the other player’s cooperation in the next trial and defection in the current trial should be punished by the other player’s defection in the next trial. As such, establishing coordination (a high level of reciprocation) might be an important key in establishing cooperation. In the present experiment, four pigeons were paired together to participate in an iterated coordination game. Each pigeon chose between cooperating on an FR 23 and defecting on an FR 15. Each cooperation response reduced the cooperation ratio requirement for the other pigeon. Under these conditions, all 4 pigeons eventually learned to cooperate despite the higher local rate of reinforcement for defection.

This study aims to produce an experimental analogue of cooperation in pigeons. An operant conditioning box split in two chambers by a transparent wall was used. Each chamber contained a feeder and two vertically aligned keys near the wall so that subjects could visualize each other while responding on the keys. Each key on one chamber was aligned to a correspondent key on the other. On each trial, all keys were simultaneously illuminated with red or green lights. Two pigeons were simultaneously placed in the box (one in each chamber). Then, they were trained individually to produce 3-seconds of food by pecking one key per trial in a 10-s fixed interval, no matter the color presented on the keys (individual contingency). After performances have stabilized, a cooperation contingency was superimposed to the individual contingency: pigeons could produce additional 4-seconds of food if they coordinately responded at same-height keys on the red lights and at different-height keys on the green lights. This condition lasted until pigeons' interlocked behaviors have produced the additional food systematically throughout a session. Results showed the selection of the target cooperative behaviors.

The aim of the experiment was to assess choice between alternatives that required individual or conjoint responses as a function of the alternatives' value. Five male rats worked in a chamber divided into two sections. In each section, there were a lateral and a central rail that both sections shared. Subjects responded by moving a ball from one extreme of the rail to the other. Each session comprised four forced and six choice trials. We assumed that the amount and delay of reinforcement determined the value of each alternative. We manipulated the amount as the number of food pellets and measured the delays. As the joint response depended on the subjects, their behavior determined the delays. Subjects worked alone in Phase 1 and by pairs in Phase 2. In both phases, the subjects chose the alternative with the higher value. Delays depended on the subjects responses in both phases but varied more in Phase 2 than in Phase 1 because of the conjoint task, which reduced the value of the amount in 1 unit (M=1.05, SD=0.06). When the relation of pellets was 4:1 for the central and lateral rails the conjoint task had the greatest value.

Cagemate rat pairs were studied in a mutual cooperation paradigm, in which alternating lever presses within 0.5 s of each other produced food for both rats. The effects of this mutual reinforcement contingency were assessed on a within-pair basis across a series of conditions. Mutual responding (alternating responses less than 0.5 s apart) was higher in conditions in which it was reinforced than in extinction conditions, despite equivalent overall food rates. Later conditions were designed to assess some effects of social familiarity on mutual behavior by assigning each rat to a new partner. Such reassignment had only temporary effects on mutual responding; despite matching pairs on the basis of dissimilar response rates, mutual responding was quickly established among the new partners. Together, the results identify some conditions that affect mutual cooperation among rats, and illustrate a promising lab model of social behavior in a controlled setting.