The effects of drive and discrimination training on stimulus generalization.

Abstract

The effects of drive and discrimination training on stimulus generalization were studied in the pigeon. Three groups of 18 Ss each, maintained at 60%, 70%, and 80% of ad lib. weight, respectively, were trained to respond to a key illuminated by a light of 550 m[mu]. After 5 days of VI training, Ss were tested for generalization to stimuli ranging from 500 m[mu] to 600 m[mu]. Then the three groups were subdivided into three discrimination problem groups, 550 mu positive for each, but with 590 m[mu], 570 m[mu], and 558 m[mu] negative, respectively. During discrimination training, responding to S+ was VI reinforced, responding to S[long dash] never reinforced. During the course of discrimination training short generalization tests were given periodically to all Ss. After the criterion was met, all Ss were subjected to a final generalization test. Conclusions were (a) minutes to criterion varies inversely with the (S+, S[long dash]) difference; a beneficial effect of drive is suggested but is not conclusive (.05<P.10). (b) Discrimination training produces a general steepening of <the PDG, a lowering of the gradient in the region of S[long dash], and a shift of the central tendency from the region of S+ in the direction away from S[long dash], amount of shift varying inversely with the (S+, S[long dash]) difference. This finding is replicated for all three levels of drive, (c) Generalization gradients obtained during the course of discrimination training reveal that the mean of the gradient shifts in a negatively accelerated manner as a function of amount of discrimination training. The latter seems to be the vehicle through which the effect of the (S+, S[long dash]) difference on the location of the PDG is mediated. Displacement varies with amount of discrimination training, independent of any direct effect of the (S+, S-) difference.