The present study explores some of the factors that determine how difficult a classification will be to learn or remember. By a "classification" we mean, here, simply a grouping of a given set of stimuli into two or more mutually exclusive and exhaustive classes. The learning or memorization of a classification can be regarded as a process of associating, to each stimulus, a certain response. This response might be the verbal label arbitrarily assigned to the class containing that stimulus, or it might be the act of sorting that stimulus into the bin arbitrarily assigned to its class. The essential feature of a classification task, however, is that the same response is assigned to several different stimuli. Accordingly, we reserve the term identification task for cases in which a different response is paired with each stimulus. In either case, the word "memorization" is intended, here, to refer to those con-

Reliable acquistion of the pigeon's key‐peck response resulted from repeated unconditional (response‐independent) presentations of food after the response key was illuminated momentarily. Comparison groups showed that acquisition was dependent upon light—food pairings, in that order.

No abstract

The relation between the form of auto-shaped responses to the lighting of a key and the consummatory responses of pecking grain and (Irinking wvater was examined in pigeons.Responses on the key were analyzed by means of high-speed photography, recordings of the force of contact, and judges' ratings of response-formii based on filni and videotape recordings. The first experiment showed that food-deprived birds presented grain as a reinforcer responded on the key with a grain-pecking movement, while water-deprived birds presented water as a reinforccr responded with drinking-like movemllents. The second and third experimnents showed that the resemllblance between auto-shaped and consummatory responses does not require the dominance of the deprivational state appropriate to the reinforcer. Changing the domiiinant state of deprivation did not illinmlediately change the form of the key response, and in subjects simultaneously deprived of food and water, the form1 of response depended on the reinforcer. In the fourth and fifth experiments, subjects simultaneously deprived of food and water received one stimuluis signalling food and another signalling water in a randomii series. In miiost subjects, the response to each stimulus resembled the consummatory responsc to the particular reinforcer that was signalled by the stimulus. This result demonstrates the role of association between a stimulus and a reinforcer in producing a rcsemiiblance of the auto-shaped response to the consuimmatory response.The procedure used in auto-slhaping (Brown and Jenkins, 1968) conforms to the paradigm for classical conditioning (Pavlov, 1927 respond to the conditioned stimulus (CS) and conditioned response (CR). As in classical conditioning, the auto-shaped response has no effect upon the occurrence of the reinforcer; the presentation of grain is contingent upon illumination of the key, but is not contingent upon the animal's behavior.Standard control procedures have shown that auto-shaped pecking is not an artifact of sensitization or pseudo-conditioning (Brown and Jenkins, 1968; Brown, 1968a Brown, , 1968b. The pigeon's approach to and contact with the key dlepends upon the contingent pairing (association) of the stimulus and reinforcer. The parallel with classical conditioning suggests that the specific form of the auto-shaped behavior might also depend upon a stimulusreinforcer association. Perhaps the pigeon pecks the key because pecking is the consummatory response elicited by grain.The stiggestion receives some support from experiments witlh otlher species and other reinforcers. Gardner (1969) reported that bobwhite quail, like pigeons, peck at stimuli that have been paired witlh grain. Squier (1969) reported that auto-shaped tilapia reacted to response keys with species-specific feeding movements, such as those used ordinarily in 163

Previous experiments have shown that educated adults generally fail to show an intuitive appreciation of correlation or contingency when judging the relation between events on die basis of a serial presentation. The effect on judgment of displaying information serially or in a summary form was examined. In contrast with some previous experiments, the events to be judged were identified in a way which should strongly suggest that the operation of chance must be taken into account. The Ss judged the amount of control exerted by cloud seeding over rainfall. The events (seeding or no seeding followed by rain or no rain) were presented to one group only serially, to a second group only in an organized summary, and to a third group in both ways with the serial display preceding the summary. Only in the group which received the summary without the serial display were die judgments of a majority of Ss more consistent with an appropriate rule of judgment involving a comparison of probabilities than with one or another of several inappropriate rules involving the frequency of certain favourable events.