Abstract
This article analyzes a type of Problem Solving Behavior in which (a) the responses were chained by direct contingencies, (b) a new response was included (one of the behavior chained was, then, new), and (c) all individual responses were maintained under the same motivational condition as that which controlled in the problem solving test. The first experiment systematically replicated Epstein's 1985 study by separately training a pigeon to push a box towards a green spot, after non-directional pushing, to climb and stand on the box, and to peck a facsimile of a banana in the bird's reach. Differences with regard to Epstein's study were: no extinction of flying or jumping with the banana out of the bird's reach, using a naive pigeon. In the test (with the banana out of reach and the box on the opposite side of the chamber) the pigeon stretched repeatedly towards the banana, pushed towards it, stopped under it, climbed, and pecked it several times. Experiment 2 evaluated how the different responses contributed to problem solving. A test was conducted after the training of each behavior. Four pigeons participated; two of the four pigeons were trained to push a box towards a target and two were trained to push the box away from a target in and attempt to isolate the stimuli controlling directional pushing in the test. In each test, only responses acquired previously, either directly or by generalization, appeared during each test. Pigeons trained to push towards or away from the spot, pushed, respectively, towards or away from the banana in the test, except for one pigeon which never pushed in any direction. No flying or jumping appeared. The test conducted after all the responses had been trained (Final Test) showed sequences with all the possible responses, but the order in which they appeared formed unsuccessful or "almost" successful chains ("almost" because the pigeon stood on the box with its back to the banana, but pecking did not occur: the banana was not in the bird's visual field). Some factors likely to explain these results are discussed as a function of: (a) the comparative rates of reinforcement among the possible responses in the test, which probably accounts for the order of the chained responses (and hence, of the successful or unsuccessful chains), (b) the schedule of reinforcement for each behavior, which probably gave rise to successful connected chains (one response quickly followed by another) or successful but disconnected chains (unnecessary repetitions of the responses constituting the solution to the problem). Along with evidence from other studies these results led to a consideration of different types of successful problem solving.
