The purpose of this study is to examine these hypothesis: forward CS will acquire negative drive properties, and backward CS positive reinforcing properties in conditioning by the noxious stimulus. The experiments were designed with twofold intention. The first was a supplementary examination to Barlow's result and to make clear the contradictory of many past studies. The second was to indicate that the hypothesis on the backward CS could be proved with more stable form by using instrumental conditioning method. In Exp. I, Ss were conditioned by inescapable procedure, in Exp. II, Ss were conditioned by procedure shifting from inescapable to escapable training and in Exp. III, by instrumental procedure as in general avoidance experiment. These secondary properties were tested in the new learning situation in Exp. I and the extinctive situation in Exp. II and III.
Experiment I: 313 rats were divided into four groups: Forward Group (5-sec. signal immediately followed by 10-sec. shock), Backward Group (10-sec. shock immediately followed by 5-sec. signal), US only Group (10-sec. shock only). and CS only Group (5-sec. signal only). Each group was divided into four (1, 5, 15 and 30) trial groups. An electric light stimulus served as the conditioned stimulus and electric shock as the unconditioned stimulus (5 trials a day) 22-24 hours after training (stimulus presentation in pairs) each group was further divided into two subgroups and tested. On the test stage, the bar was inserted in the same situation. CS occured for half of the Ss for, however, long the Ss pressed the bar (CS-on). For the other half of the Ss, CS illumination was presented when the Ss was put into the apparatus. For these Ss down the bar terminated the CS (CS-off). This bar press duration was measured for 12min after the first response.
Results: For the Backward Group, 1 trial group proved the hypothesis (t=2.621, df=17, p<0.05), 15 trial group showed the tendency, 30 trial group was neutral and 5 trial group negative. (See Fig. 2)
Experiment II: 60 rats were divided into three groups: Forward Group, Backward Group and Control Group (US only). On the training stage, as soon as US was presented, the bar inserted, and the bar press response terminated US. CS fixed 5-sec duration and unless the bar press response occured US terminated 10-sec duration (5 trials a day). This resulted in the Ss pressing the bar more than four trials within 5-sec from US on set for two consecutive days. They were tested in 22-24 hours after achieving it. The test stage was similar to Exp. I. This bar press duration was measured for 15min after being introduced into the apparatus.
Results: The difference between two subgroups of Forward Group during the test hypothetically showed the significant effect (t=3.265, df=19, p<0.05), that of the Backward Group had no significant effect (t=0.892, df=17, p>0.05), and in analysis of variance, the interaction (Group×Subgroup) showed the significant effect (F=7.768, df=2; 59, p<0.01). (See Fig. 6)
Experiment III: 66 rats were divided into three groups: Forward Group. Backward Group and Control Group. On the training stage, as soon as stimulus (CS or US) was presented, the bar was inserted and the bar press response terminated US. But unless the response occured, US was kept. Only in Forward Group, as soon as the bar press response occured during the presentation of CS, CS was terminated and US was not given. The others were similar to Exp. II.
Results: The difference between two subgroups of Forward Group and Backward Group during the test, the significant effect was seen (t=5.178, df=19, p<0.01; t=2.454, df=22, p<0.05). (See Fig. 10)
The results of three experiments made were that, as to secondary properties of Backward CS in
