Latent Extinction under Irrelevant Drive

Abstract

Within the last decade the concept of anticipatory goal response has come to carry an increasing explanatory load in S-R interpretations of learning. One of the most pervasive assumptions which has been made about the r, mechanism is that r,s which are elicited by hunger and associated with eating are to some extent incompatible with those elicited by thirsr. This assumption is implicit in Spence's interpretation of irrelevant incentive learning (1951, pp. 279-280). It is made explicit by Kendler and Levine (1953) who that the anticipatory eating responses of dry food are, in some manner, incompatible with the resultant condition of water deprivation (probably conflict with anticipatory drinking responses). More recently, Ritchie has argued that the assumption is necessary for an 7, interpretation of double drive discriminations (Ritchie & Owen, 1959) and has postulated that all fractional reactions are incompatible with each (Ritchie, 1959). On the other hand, Deutch (1956) has questioned the accuracy of Hullian derivations of latent learning, insight, and reasoning which rely on the r, mechanism and has indicated that the assumption of ro competition will not save the derivations. The present srudy tested a theorem which may be derived from this 7, compecition hypothesis when it is combined with the theoretical properties assigned to ro in the latent extinction sirnation. In brief, it is suggested that, when Ss are trained to obtain food under hunger drive and then are given direct placements into the now unbaited goal box, 7, eating (r,e) is being elicited as a free operant and is being extinguished to the goal box cues. The reduction in the strength of r7, is assumed to lower reaction potential by reducing the value of incentive motivation (K) . This argument is supported by the fact that variables whicb theoretically increase the emission of r0 during the latent extinction placements result in a faster extinction of the running response (Pliskoff, 1955; Moltz & Maddi, 1956). The theoretical implication is that, if the simultaneous presence of hunger and thirst results in compecition of the r, associated with eating and drinking, and if the latent extinction effect is dependent on the rate of emission of drive relevant ~~s in the goal box, then preextinction exposures under conditions of hunger should result in greater latent extinction than under both hunger and thirst. The following theorem may be derived: if hungry Ss are trained to make a choice response to obtain dry food, and given pre-extinction exposures to the unbaited goal box under drive con-