Discrimination learning and pre-delay reinforcement in 'delayed response.'

Abstract

For references see (a, page 104, footnote 2). Crannell (5) has recently performed an experiment with rats with a set-up employing four alternative pathways, which were to be run successively in whatever order the rat selected, no second run of any one path being permitted. Delayed response tests were instituted by interposing a delay between the third and fourth runs, a correct post-delay response being the choice of the (fourth and remaining) pathway not previously chosen. Each run was rewarded with food. Successful delays up to 4 hours were observed, even though the animals did not maintain any stereotyped sequence of pre-delay runs in repeated tests. This progressive elimination of three paths constitutes the series of training trials which determines the rat's response after delay. The criticisms of Maier's experiment, by those denying it delayed response status, would undoubtedly be applied to this one.* This no-food situation is not to be confused with the 'latent learning' maze situation, in which the animal has never before associated the correct maze path with food reward. There is no 'latent learning' here, in the accepted sense.

Delayed Matching in the Pigeon

“…There are reports of still other cases in which, due to the length of the delay or the variety of intervening activity, the presence of covert chaining does not seem probable. It may be, as Cowles (1941) and others have suggested, that the behavioral processes involved in such cases are essentially the same as those that occur when a discrimination is learned and retained. That is, a stimulus gains discriminative control during its presentation in the pre-delay period ("one-trial learning"), and retains this status upon its subsequent presentation in the same way that an ordinary discrimination is learned and then retained from day to day or from year to year.…”

NON‐SPATIAL DELAYED ALTERNATION BY THE PIGEON1

“…An alternative to a response chaining explanation of delayed response performance derives from considering the performance as simply one fuy m of discrimination learning. By this view (e.g., Cowles, 1941;Harlow, 1951), the delayed response problem consists of two phases: the learning phase when the cue to be retained is presented, and the retention test when the delayed response choice is made. The reason that response accuracy decreases as a function of delay, then is that the learning during the presentation of the cue is incomplete due to its being presented for only one trial and to the great interference from previous learning.…”

Effects of practice on the delayed response of retardates.

“…On the level of theory, the practice variable might be assumed to play a pertinent role in interpretations of DR as discrimination learning. Such theorists as Cowles (1941), Nissen, Riesen, and Nowlis (1938), and Spaet and Harlow (1943) have attempted to reduce DR in whole or part to processes involved in discrimination learning. The last authors have, for example, said that delayed reaction is among other things "... a series of discrimination reversals learned in a single trial to a signified reward" (p. 430).…”