In a recent study by Tolman, Ritichie, and Kalish (1) on spatial learning, an experiment was designed to test experimentally a revision of the senior author's "theory of expectancy." Animals were given four periods of preliminary training on apparatus presented in diagram in figure 1. The test trial was given on apparatus essentially the same as that presented in diagram in figure 2. (See discussion of apparatus in present article.)The study by Tolman, Ritchie, and Kalish involves five assumptions relevant to our work:1. The animals that chose path 6 (36 per cent) learned the location of the food box, and those that did not choose path 6 (64 per cent) did not learn the location of the food box.2. The animals that chose path 6 acquired an expectation of food at L. "x expects food at location L."3. Tolman, Ritchie, and Kalish assume that the 64 per cent which did not choose path 6 would have chosen path 6 had they been given training sufficient to determine the location of the food box. We quote:We believe that the reason that the remaining rats failed to take the short-cut was that they had not had enough training and thus had not yet learned the location of the food. With a few more days training we should have expected that the remaining rats would have chosen the short-cut (pp. 21-22).4. The animals that chose path 6, the short-cut, inferred that it, of the 18 available pathways, was the pathway to location L.5. They assume that the light during the training period became a cue of location L and that it was essential in the attainment of their results. We quote:We believe that such choices can only be made when there are distinctive stimuli in the environment which enable the rat to judge its own location relative to other places in the environment (p. 20). . . . We should say that they were running towards the location of the former goal, and that this location was indicated by the position of the light (p. 21).Since the expression "x expects food at location L" is defined (p. 15) without reference to such distinguishable cues as the light employed in the training and in the test trial, we were led to wonder, among other things, whether the results reported demonstrated that (1) there were some x's that expected food at location L, or (2) whether the results obtained were not accountable for in terms of the

In a recent study by Gentry, Brown, and Kaplan (1) on spatial learning, results were obtained which were at variance with results found by Tolman, Ritchie, and Kalish (2). Both studies gave the animals four periods of preliminary training on the same type of apparatus and used the same type of apparatus on the test trial. Tolman, Ritchie, and Kalish found that 36 per cent of their animals chose the short-cut path from eighteen novel available pathways when the tunnel path (CD) was blocked on the test trial. The tunnel path was used by I/he animals during preliminary training. They assert that the animals "were running toward the location of the former goal, and that this location was indicated by the position of the light (p. 21)." Gentry, Brown, and Kaplan found, in a repetition of the work by Tohnan, Ritchie, and Kalish, while using four periods of training, that their animals had a tendency (37.5 per cent) to choose pathways adjacent to the blocked alley, and that none chose the short-cut path to the former location of the food box.Tolman, Ritchie, and Kalish assume that the 64 per cent of their animals which did not choose the short-cut on the test trial had not been given sufficient training to determine the location of the food box. We quote:We believe thai the reason that the remaining rats failed to take the shortcut was that they had not had enough training and thus had not yet learned the location of the food. With a few more days training we should have expected that the remaining rats would have chosen the short-cut (pp. 21-22).Gentry, Brown, and Kaplan failed to get a tendency to choose the short-cut path on the test trial following nine periods of training, rather than four, as was the case in the study by Tolman, Ritchie, and Kalish. Gentry, Brown and Kaplan found that 30 per cent of their animals chose paths 9 and 10, paths adjacent to the blocked alley, on the test trial following nine periods of training.Gentry, Brown and Kaplan explained their results in part in terms of an artifact of the experimental apparatus, i.e., the presence or absence of blocked alley CD on the test trial. It appears to us that the explanation by Tohnan, Ritchie, and Kalish of their results involves the assumption that their rats possessed a capacity to discriminate between the eighteen novel available pathways, without distinguishable features at the choice points, with which they were confronted in the test trial, and to determine (infer) without error that No. 6 pathway was a short-cut, i.e., the pathway to location L, Gentry, Brown, and Kaplan "found no evidence that any of the rats . . . used had such a capacity and much evidence that none of them did (p. 321).

No abstract

No abstract

No abstract