Response-intensity generalization in a monkey

Abstract

Behavior variability is one of the important aspects of human behavior. In order to establish the scientific behavior theory of the higher organisms, we must study the nature and characteristics of behavior variability. The writer believes that the best results of this investigation will be obtained by experiments using a monkey as a subject, because the behavior of a monkey is complex next to man, and is more easily observed than man under controlled experimental condition.
The purpose of the present experiments is to investigate the response-intensity generalization phenomenon as one aspect of behavior variability.
The apparatus used was a pulling response box. The main parts consisted of a pulling response ring, a food release mechanism, and a recording device for response intensity.
The subject was one female monkey. The twenty-day conditioning taials (40 trials per day) were given. In the first half or ten days of experiment (Exp. 1), the reinforcements were given when the response ring was pulled to the extent of 1.5 kilograms or more, but were never given for the pull of less than 1.5 kilograms. In the last ten days (Exp. 2) the upper restriction of 2 kilograms, was imposed on ' the conditioned pulling responses previously set up under a 1.5-kilogram lower limit. Therefore, in Exp. 2, all responses above the 2-kilogram limit were eleminated.
The following results were obtained :
1) The dap-to-day percentages of correct responses (reinforced responses) showed an increasing tendency, and more correct responses were found in Exp. 1 than in Exp. 2 (Fig. 1).
2) The mean reaction latencies decreased as a function af the experimental day (Fig. 2). And it was found that the mean reaction latencies calculated by trials on the last 2.5 days had similar decreasing tendencies in both experiments as the day-to-day process'(Fig. 3).
3) The response intensity distribution on the last 100 trials showed that a few nonreinforced responses below the lower limit still occurred in Exp. 1 (Fig. 4). And in Exp. 2, they were found outside of both upper and lower restrictions (Fig. 5). These non-reinforced responses were considered as the generalized respgnses on responseintensity dimension.
4) The range of non-reinforced responses below the lower limit was broader in Exp. 1 than in Exp. 2.
Conclusion : The writer submits that these experiments confirm the possibility of instrumental conditioning study on behavior variability in a monkey. And he discusses the validity of Hull's explanation of response-intensity generalization, and suggests that there are many problems left unsolved.