۱- Introduction

۲- General method

۳- Experiment 1

۴- Experiment 2

۵- Experiment 3

۶- Experiment 4

۷- General discussion

References

Abstract

Skinner’s (1948) ‘Superstition’ in the Pigeon paper proposed that accidental response-reward contingencies, via adventitious reinforcement, could operantly condition the behaviors of pigeons under fixed-time (response-independent) schedules of food delivery. Skinner likened the behavior of pigeons under these fixed-time schedules to the superstitious behavior of humans and proposed that both response patterns were the result of contiguous pairings of rewards following some response. Alternative explanations of superstitious behavior have included Staddon and Simmelhag’s (1971) stimulus substitution account and Timberlake and Lucas’s (1985) elicited species-typical appetitive behavior account. Under both these alternative explanations of superstitious behavior, observations of pigeons under fixed-time schedules revealed a lack of idiosyncratic responding, which is a critical element in Skinner’s explanation of superstitious behavior via adventitious reinforcement. The following study implemented 4 fixed-time schedule experiments to further study superstition. In Experiment 1, male and female pigeons were compared, which provided support for the disparity in response patterns observed in previous studies. Experiments 2–۴ examined the behavior of roller pigeons, ring-necked doves, and bantam chickens. In all the above studies, a lack of idiosyncratic responding and emergence of species-typical foraging behavior was observed. The results provide additional evidence that the ‘superstitious’ behavior that emerges in pigeons and other organisms under response-independent food schedules is the result of elicited species-typical food getting behaviors, and that these behaviors emerge as a result of frequent food deliveries in environments that support such foraging repertoires.

Introduction

Skinner (1948) applied the term “superstition” to stereotyped, idiosyncratic behaviors of pigeons that emerged when a wall hopper filled with grain was briefly presented for 2–۴ s on a Fixed-Time 15 s (FT-15 s) schedule. According to Skinner’s informal account, after less than an hour of exposure to such a schedule, six out of eight pigeons developed a dominant, idiosyncratic response during the inter-food interval. These responses included circling, pendulum movements of the neck and head, and head tossing. Skinner (1948) labeled these responses superstitious because they appeared in the absence of a programmed response contingency between the behavior and the reward. He compared them to the behavior of a bowler applying “body English” after they released the ball as if trying to guide the ball into the pins from a distance, and to people engaged in rituals that have been related to success at card games in the past. Three explanations have been used to account for the superstitious behavior of pigeons. Skinner (1948) argued that each contiguous pairing between a behavior and a reward increased the future probability of that response, thereby increasing the likelihood that the future presentations of the hopper would follow or overlap that response again. In other words, Skinner posited the existence of a feedback effect whereby an “accidental” reward contingency (i.e., adventitious reinforcement) increased the strength of any response it followed, and a feedforward loop whereby the increase in the reinforced response increased the likelihood it would be followed by food again. Although, several other investigators have reported similar behaviors in pigeons produced by FT schedules (Eldridge et al., 1988; Justice and Looney, 1990; Neuringer, 1970), none have directly tested Skinner’s model of how superstitious behavior arose as a function of the adventitious reward proximity (i.e., contiguity) and the resultant increase in the likelihood of more adventitiously reinforced responding. In most studies of superstition in pigeons, response-dependent fixed- or variable-interval (FI; VI) schedules are first introduced, and only one response topography is examined: key pecking.