Common label pretraining: Interference and/or facilitation?

Abstract

Several tests of the hypotheses of acquired equivalence and distinctiveness of cues (Dollard & Miller, 1950) have employed common-label (CL) and/or different-label (DL) treatments during the pretraining or predifferentiation (PD) phase of the investigations (Robinson, 1952;Ellis, Bessemer, Divine, & Trafton, 1962;Katz, 1963;Ellis & Muller, 1964;del Castillo & Ellis, 1968). Equivalence or CL training has required that Ss learn common labels to clusters of two or more stimuli within a test list. Distinctiveness or DL training has required that Ss learn a different label to each stimulus. The majority of these studies have compared the performances of CL and DL groups on recognition or simple discrimination tasks with the performance of a control which has merely observed (0) the critical stimuli during PD training. Under limited conditions of stimulus complexity and response relevance, DL training has facilitated test task performance, while CL training has resulted in performance inferior to that of the 0 control. These data have been interpreted as consonant with the supposition that CL training increases the overall generalization tendencies among the stimuli of the test list, while DL training enhances their discriminability through the acquisition of identical and distinctive verbal response-produced cues, respectively.Recently, del Castillo & Ellis (1968) tested the acquired equivalence (AE) and acquired distinctiveness (AD) hypotheses in the setting of a transfer task which required Ss to make new discriminative motor responses to the stimuli of the verbal PD task. Using four visual stimuli which differed along the single dimension of light intensity, the authors found no evidence for AE, although the AD hypothesis was supported by the superior performance of DL-treatment groups compared with 0 controls. One factor which may have contributed to the failure of evidence for AE (and to the strong evidence for AD within the same design) is the limited appropriateness of the 0 control in the AB-AC stimulus differentiation paradigm. While the 0 treatment does familiarize Ss with the stimuli of the transfer task, and may as well allow Ss to form systematic implicit associations to these stimuli (Ellis & Homan, 1968), it does not give Ss the opportunity to acquire the performance set necessary to the efficient mastery of a subsequent associative task in which the overt paced production of discriminative motor responses is required. Thus, in proceeding from PD to transfer tasks, O-treatment Ss may suffer a performance decrement attributable to differences between task demands. This would at once vitiate the expected superiority of the performance of o Ss to those given the CL treatment, and maximize the superiority of DL to 0 treatments.The purpose of the present experiment was to explore Psychon-. Sci., 1968, Vol. 13 (6) within the AB-AC paradigm the proactive consequences of DL and CL training with respect to a control group given irrelevant training (lCL) and an additional control given no training at all (NL). W...

Ambivalent proaction following common label pretraining

“…RDL training resulted in fewer errors on the test task than either RCL or ICL training; RCL training resulted in fewer errors than did ICL training. These results confirm earlier evidence (Marshall, 1968b) that the number of verbal training responses and relevancy of training are significant sources of transfer in the AB-AC pre differentiation paradigm.…”

“…yet in the latter study, an RCL treatment group performed as well as the RO control. In another study (Marshall, 1968b), the consequences of Psychon. Sci., 1970, Vol.…”