Sharing and Cooperation in Pigeons General Discussion Table Of Contents History Title Page List of Figures, Tables Introduction Purpose, Method Results and Discussion General Discussion References Abstract

General Discussion

This study attempted to demonstrate sharing or cooperation in pigeons. The definitions of sharing and cooperation both require an increase In correspondence of reinforcers and/or responses between subjects across sessions. Sharing is characterized by the responding of only one subject which would produce reinforcement for itself in a trial. Cooperation is characterized by only one subject engaging in responding which would result in reinforcement for the other subject in a trial (Hake and Vukelich, 1972). The form of sharing or cooperation attempted In this study required that both subjects of a pair respond on each trial for either subject to receive reinforcement and is therefore a mixture of sharing and cooperation, as has been the case for previous cooperation studies (e.g., Daniel, 1942; Daniel, 1943; Taylor and Espamer, 1971). The term sharing will be used to refer to both sharing and cooperation procedures hereafter, unless the distinction between cooperation and sharing is relevant.

The current free response-role distribution procedure is also a mixture of what Hake and Vukelich (1972) refer to as response-sharing and response-exchange procedures. The requirement that subjects respond to distribute response roles qualifies the procedure as a response-exchange procedure (exchange of takes or gives); the requirement that both subjects respond during the matching-to-sample interval qualifies the procedure as a response-sharing procedure.

The definition of sharing requires the possibility of deviation from an equitable distribution of responding, and/or the possibility of deviation from an equitable distribution of reinforcement frequency or density between subjects (Hake and Vukelich, 1972). If an equitable distribution of reinforcement or an equitable distribution of response/reinforcement ratios is imposed on subjects, deviation from equity may not occur, increases in the correspondence of reinforcements and/or responses per reinforcement may not be shown, and sharing or cooperation cannot be demonstrated (Hake and Vukelich, 1972; Hake, Olvera, and Bell, 1975)

The free response-role distribution procedure in this study allowed subjects to determine the distribution of reinforcements between subjects. Subject control of reinforcement distribution could have, and did, result in an inequitable distribution of reinforcements between subjects. Subject control of reinforcement distribution could have, but did not, result in an increased correspondence in the number of reinforcements received by subjects across sessions. The free role-distribution procedure qualifies as a sharing procedure; however, no sharing effect was shown. Rather, the study ended with both subjects in each pair engaging in competitive responding, a type of behavior that has been used as a baseline in studying sharing with humans (Hake, Vukelich, and Olvera, 1975; Hake and Schmid, 1981). Competition has been defined as a method of responding in which subjects engage in "take" responding which results in a reinforcer for only the subject to first complete the response requirement (Gasparotto, 1983; Hake, Olvera, and Bell, 1975; Hake, Vukelich, and Olvera, 1975).

The behavior labeled cooperation in other animal studies may have been under the control of competitive contingencies also (Daniel, 1942; Daniel, 1943; Taylor and Espamer, 1971). The studies by Daniel produced alternate "take' responding in the form of subjects responding in such a manner as to produce shocks for the other subject and/or themselves. The shocks produced an alternation of subjects at the food crock (Marcucella and Owens, 1975). While subjects did alternately feed, and the alternation was under the control of social stimuli, the alternate feeding was not under the control of the food-reinforced cooperation procedure. Similar problems occurred in the study by Taylor and Espamer (1971), as discussed previously. Boren (1966) demonstrated share-like responding under the control of the food reinforcement contingency, for a time, but did not demonstrate social control of the alternations. The lack of, or possible lack of, social control of alternately responding and feeding in Boren (1966) may in part account for the failure of the share-like responding to maintain.

The response-role distribution procedure of this study did produce alternate "take" responding under the control of both 1) the responding of the other subject, and 2) the food-reinforcement contingency. Control of take responding was demonstrated by differential rates and/or latencies of take responding on trials on which both subjects engaged in take responding as compared to those trials on which only one of the subjects of a pair engaged in take responding. Control of take responding by the food-reinforcement contingency was demonstrated by the differential take-response rates and latencies between those conditions in which food-reinforcement was available and those conditions in which food-reinforcement was not available in the follower response role. Control of take responding by the food reinforcement contingency was also indicated by differences in the proportion of trials with competitive-take responding between those conditions in which food reinforcement was available and those conditions in which food reinforcement was not available in the follower response role.

Even though social control of food-reinforced take responding was demonstrated, the alternating take-responding was competitive. The take-responding was competitive as measured by the increase in competitive take attempts per session by both subjects during no-foll-food conditions, and as measured by increases in take-response rates and/or decreases in take-response latencies during no-foll-food conditions as compared to toll-food conditions. That is, when food was available in only one response role subjects tended to take the food reinforced response role away from each other rather than to alternately "let" each other take the food reinforced response role (cf. Hake, Olvera, and Bell, 1975; Hake and Schmid, 1981). For sharing to have been demonstrated in the free role-distribution phase of this study, subjects would have had to come under the control of the differences in response requirements between competitive responding and share responding, as is indicated by previous studies with humans (Hake and Vukelich, 1973; Hake, Olvera, and Bell, 1975; Hake and Schmid, 1981). As stated by Hake, Olvera, and Bell, competition requires more responding (i.e., competition results in higher effective response/reinforcer ratios).

The pigeons in this study competed on roughly 35-45% of trials per session during no-foll-food conditions, which means that 55-65% of the reinforcements were distributed non competitively. Subjects experienced both the competitive and sharing response requirements. However, two factors may have played a major role in the failure to establish a discrimination between the competitive and sharing contingencies.

One factor may have been the relative brevity of the subject-controlled no-foll-food response role distribution conditions. Subjects were in the no-foll-food conditions for a total of 25 sessions (pair 1) and 13 sessions (pair 2). (in contrast, subjects required 16 to 54 sessions to show socially controlled match-key discrimination. The socially controlled match-key discrimination was established; the competition/sharing discrimination was not established. Is the match-key discrimination in some sense inherently "easier", establishabled in fewer trials, than the competition/sharing discrimination? Perhaps running the subject-controlled no-foll-food response role distribution conditions for over a longer period would, by itself, have resulted in a sharing effect.

A second factor in the failure of the subjects to discriminate between competitive and sharing contingencies may have been the size of the take-response itself (Hake, Olvera, and Bell, 1975; Olvera and Hake, 1976; Hake and Schmid, 1981). That is, the differences in effective ratios between the competitive and sharing contingencies may have been too slight to be either discriminable or effective in producing sharing. The response requirement on the take-key was between 5- 15 responses at the start of a session. This response requirement could go down to 1 for a subject who failed to take the leader role for several trials or could increase by one response for each trial on which a subject took the leader role. While the exact maximum take-response requirement that occurred is not known, given the observation of subjects during sessions, the maximum ratio encountered by a subject is not likely to have been over 30 responses. If the take response requirement had been raised, a shift from competition to sharing might have occurred. The previously cited human sharing and cooperation studies support the view that the take-response requirement may have been the major factor in the failure to produce share responding in the relatively brief no-foll-food conditions. Animal studies on the aversive aspects of large fixed ratios lend further credence to the critical nature of the size of the take-response (Appel, 1963; Dardano, 1973; Thompson, 1964; Thompson, 1965).

While this study did fail to demonstrate sharing, a procedure which could be used to generate and study sharing in non-humans has been developed. The free role-distribution procedure (a) provides a means for allowing subjects to determine the distribution of reinforcements, (b) allows deviations from equitable distributions of reinforcements between subjects, but also allows for an equitable distribution, (c) allows methods of distributing reinforcement other than by sharing (i.e., competition), (d) provides a means of demonstrating social control of the distributive (take) responding, and (e) provides a means of demonstrating control of the distributive responding by the food reinforcement contingency, all of which are critical to the demonstration of sharing or cooperation.