Preference for locus of punishment in a response sequence

Food-deprived pigeons pecked a key under a schedule in which grain was made available after the seventieth peck. In each sequence of 70 responses, either the first, middle, or final response was followed by electric shock. Before the first response of each sequence, each response on a second key changed the color of the food key and the schedule of shock that was correlated with the food key color. Each pigeon preferred a schedule of shock, in that each of the three shock schedules did not occur equally often. The preferred shock schedule and the strength of the preference varied among the pigeons. The overall rate of responding by a pigeon under a given shock schedule was directly related to the pigeon's relative preference for that schedule, except when shock after the first response in the sequence was the most preferred schedule.     

Reversal of preference under progressive-ratio schedules by punishment

A progressive-ratio reinforcement schedule, in which successive reinforcements required an additional 50 responses, was programmed on one key. A response on a second key reset the progressive-ratio schedule to the first step. Before punishment, all pigeons consistently reset the schedule after reinforcement on the first step, thereby minimizing the number of responses required for reinforcement. Punishment was a brief electric shock contingent upon each response on the reset key. The first effect of punishment was to change the frequency of extra responses on the reset key. Under higher intensities of punishment, the pigeons completed the advanced steps of the progressive-ratio schedule before resetting to the first step. Completions of advanced steps were accompanied by decreases in the overall rate of responding and the rate of reinforcement. When the punishment contingency was removed, the major features of pre-punishment performance were recovered.     

Fractional punishment of fixed-ratio performance

Key-peck responses of pigeons under a fixed-rate 60 (Exp. I) or fixed-ratio 99 (Exp. II) schedule of positive reinforcement were punished by response-dependent electric shock during a segment of the ratio. The punishing stimulus was scheduled in one of three locations: the first third of the ratio, the middle third, or the final third. At high shock levels, the different loci of punishment differentially affected the typical fixed-ratio performance pattern. Post-reinforcement pauses were lengthened by all punishment conditions but to a greater degree when the responses in initial third of the ratio were punished. Disruption of responses before the punished segment of the ratio was a conspicuous feature of the performances when the middle or final third of the ratio was punished. Two of the punishment conditions produced similar effects on both fixed-ratio baselines but punishing the final third of the ratio suppressed the punished responses of the ratio only with the fixed-ratio 99 schedule. General effects of all punishment conditions included consistent intra-session recoveries of partially suppressed performances, the rapid recovery of the FR performances after the punishment dependency was removed after complete suppression, and the facilitation of overall and/or local response rates of most subjects by low-intensity shock.     

Self-imposed timeouts under increasing response requirements

Self-imposed timeouts by pigeons working under a progressive-ratio food schedule were studied under different conditions. The main findings were (1) continued production of timeouts over an extended series of sessions, (2) more frequent responding on the key with the timeout consequence than on a key having no consequence, (3) an inverse relationship between number of timeouts and level of body weight, (4) production of timeouts when the timeout duration was brief, lengthy, or controlled by the pigeon, and (5) dependence of self-imposed timeouts on variables controlling responding under the progressive-ratio schedule. Under all experimental conditions, with the exception of performances at the high body weight, timeouts were more frequent during the longer progressive-ratio steps and usually were localized in the post-reinforcement pause or the early part of the step. The timeout behavior could be interpreted as either an escape from aversive stimuli generated by the progressive-ratio schedule or as a response reinforced by the consequent stimulus change.     

Response-produced timeouts under a progressive-ratio schedule with a punished reset option

Three behavioral options were available to food-deprived pigeons: (1) pecking one key resulted in food reinforcement according to a 50-response progressive-ratio schedule, (2) pecking a second key reset the progressive-ratio schedule to the initial progressive-ratio step, and (3) pecking a third key produced a 3-min timeout period. Pecks on the reset key were shocked. Under low and intermediate shock intensities, timeouts were not produced; under high shock levels, timeouts were produced regularly. Timeouts occurred during the initial period of a progressive-ratio step and were more frequent during the longer steps of the progressive-ratio schedule. Response-produced timeouts under these experimental conditions could be interpreted either as an escape from aversive behavioral options or as a low-probability behavior emerging when the food reinforcement schedule exerted weaker control.     

Control of concurrent avoidance and appetitive behaviors by an indicator of shock proximity

Two monkeys were exposed to concurrent schedules in which every fifth response on one lever had the dual consequences of delivering food and stepping a 20-step light counter upward; responses on a second lever stepped the light downward. By responding appropriately on each lever, a ration of food could be obtained without exposure to a brief shock that occurred when the counter reached the twentieth step. Behavioral patterns of light-counter control were analyzed when the monkeys were handicapped by several types of discontinuities in the 20-step counter: (1) different sets of four consecutive lights inoperative, (2) the upper or lower 10 steps inoperative, and (3) all lights inoperative. Generally, the monkeys maintained the light in the upper five steps of the 20-step light counter if the lights immediately preceding the shock position were functional. However, when these counter positions were inoperative, subjects maintained the counter around the highest operative position. With all positions of the light counter inoperative, no systematic pattern of responding on the food and avoidance levers occurred.     

An aversive stimulus as a correlated block counter in FR performance

After the performances of two pigeons had stabilized on an FR 100 grain reinforcement schedule, response-contingent shock was introduced in the form of a block counter. When shock intensity increased in successive thirds of the ratio, duration of post-reinforcement pauses increased and local rate in an advanced part of the ratio declined. With decrements of shock intensity in successive thirds of the ratio, running rate was not consistently affected but pauses after reinforcements were longer than those under the increasing counter. The results are interpreted in terms of the interaction among some controlling variables operating in the behavioral chain.     

Selective punishment of fixed-ratio performance

Key-pecking of pigeons, maintained by an FR 50 grain reinforcement schedule, was punished by shocking the first, middle, or last response of the ratio. Under high shock levels, the three punishment conditions produced differential effects on the behavior. Punishment of the first response of the ratio resulted in consistent and extended post-reinforcement pausing and frequent extended breaks after the initial response(s) of the ratio. Punishment of the 25th response disrupted responding in the first half of the ratio with little effect on the last half of the ratio. Punishment of the final response resulted in breaks and local rate changes in various parts of the ratio. Durations of pauses after reinforcement were more variable when the 25th and 50th responses were punished relative to those when the first response was punished.     

Variable location of punishment in a response sequence

Key pecking of pigeons under a fixed-ratio 100, grain reinforcer schedule was followed by electric shock occurring once in each sequence of 100 responses with the shocked response varying irregularly in successive sequences. Under this shock schedule, a localized suppression of responding in a response sequence was not correlated with the probability of shock at different points in the sequence. High shock levels increased the duration of post-reinforcement pauses and suppressed responding during the first half of the response sequence. This suppression often persisted after the shocked response when shock occurred early in the sequence. The shock schedule did not produce a consistent suppressing effect on responding during the last half of the response sequence.     

Modification of observing behavior

The effects on monitoring performance of a limited hold (LH) terminated by a bell or by electric shock and of periodic rest periods were investigated. The sensitivity of basal skin conductance to different monitoring conditions was also studied. Humans monitored a visual display in several 2-hr sessions under the Holland procedure. In each session, signals were programmed by one of the following three multiple schedules: a variable interval of 6 min (VI 6) and a VI 6 with a limited hold (LH) terminated by a bell; a VI 6 and a VI 6 LH terminated by electric shock; a VI 6 LH terminated by shock and a time-out (TO). The two components of each schedule alternated at 15-min intervals. The average observing response rate of eight subjects was lowest during the VI 6 periods, higher in periods when the LH was terminated by a bell, and highest when the LH was terminated by shock. Periodic TO periods did not appreciably increase the response rate under the VI 6 LH shock component. A differential level of palmar skin conductance under the two components of a schedule was present only under the mult VI 6 LH shock-TO.