Discrimination reversal following learning without “errors”

Abstract

Pigeons were trained on a discrimination using the "errorless" procedure devised by Terrace and then given a discrimination reversal. Their extinction to the previously positive cue during reversal was similar to a control group given continuous reinforcement during original training. Most of the Ss in the errorless group, however, failed to learn the reversal because they did not respond to the previously Ilegative cue during five days of reversal training.

Effects of Error and Errorless Discrimination Acquisition on Reversal Learnng

“…First, the children who acquired the discrimination with few or no errors perseverated in their responding to the original S+ on the first trial after reversal; children acquiring the discrimination with a relatively large number of responses to S-showed little perseveration. This tendency of subjects acquiring a discrimination errorlessly to perseverate on the initial S+ trial after reinforcement contingencies are either removed or reversed is also consistent with previous reports (e.g., Marsh and Johnson, 1968;Wilkie and Raimer, 1974). Second, the children who acquired the initial discrimination with errors emitted a substantially larger number of errors after the first reversal trial than did the children who learned errorlessly.…”

“…To date, only one investigation has attempted to study the effects of abruptly reversing the roles of S+ and S-after the initial discrimination is acquired errorlessly. Marsh and Johnson (1968) were unsuccessful in obtaining reversal learning with pigeons after a successive color discrimination was acquired errorlessly. Only one of five pigeons responded even to a small extent to the new S+ (previously S-).…”

“…Marsh and Johnson (1968) were unsuccessful in obtaining reversal learning with pigeons after a successive color discrimination was acquired errorlessly. Only one of five pigeons responded even to a small extent to the new S+ (previously S-).…”

INHIBITORY STIMULUS CONTROL FOLLOWING ERRORLESS DISCRIMINATION LEARNING1

“…Using an errorless autoshaping procedure similar to that employed by Wessells (1973), Wilkie and Ramer (1974) found that an errorless S-was more resistant to autoshaping than an errorful S-. Marsh and Johnson (1968) were the first investigators to suggest that an errorless S-is an inhibitory stimulus. First they established an errorless successive discrimination.…”

Errorless learning in newborn chicks

“…Several aspects of previous investigations suggest the possibility that errorless learning cannot occur unless an organism has a rather extensive preexperimental history. First, all previous publications on errorless learning have involved either adult nonhumans (e.g., Marsh & Johnson, 1968;Schusterman, Kellogg, & Rice, 1965;Terrace 1963Terrace , 1966Terrace , 1969 or human subjects at least 4 years of age (e.g., Sidman & Stoddard, 1967;Storm & Robinson, 1973;Terrace, 1974). Second, Warren, Brookshire, Ball and Reynolds (1960) conditioned a position and brightness discrimination III chicks ranging from 3 to 63 days old.…”