Pragmatism and Behaviorism: "Hypotheses"

Below is a quote from psychologist E. C. Tolman. It shows the clear influence of Pragmatist thinking on Radical Behaviorism. (Personally, I think it could be done better without reference to something "within the organism", but that is a minor point.)

------------

(2) But let us turn, now, to a brief consideration of the second main subtype of cognitive behavior-readiness, what I called hypotheses; or intentions, expectations and attainments as to relations. Suppose a rat be run in a successive discrimination box. Such a box is an apparatus in which the animal has to choose one of two doors at each of four successive choice points. One of the two doors at each such point is lighted and one is dark. The lighted door may be either the one on the left or the one on the right in chance order. Thus at each such point the animal has the possibility of responding either on the basis of light-darkness or on that of right-leftness. Suppose, now further, that it be arranged by the experimenter that the correct choices shall in a day's series of 10 trials, or 40 choices in all, fall an equal number of times to the left and an equal number of times to the right, and suppose it also be arranged that the correct door be an equal number of times a dark door and an equal number of times a lighted door. Under these conditions it was found by Krechevsky, whose experiments it is I am reporting, that the rat will pick up one systematic way of behaving after another. In the first two or three days he may pick up, say, the propensity of choosing always the right hand doors. But then he will shift sooner or later to some new propensity, to that say, of choosing only the left hand doors; and then still later to that of choosing alternate right and left doors; or he may shift to choosing all the lighted doors, irrespective of side, or all the dark doors, or to choosing alternately light and dark; and so on. Each such systematic propensity will be adopted for a time and then dropped in favor of some other. And, following Krechevsky, we may now define each such intervening condition (or "I") in the organism, behind any one such systematic way of behaving, as an hypothesis. An hypothesis, behavioristically, in other words, is to be defined as nothing more nor less than a condition in the organism which, while it lasts, produces just such a systematic selectivity in behavior. Further, it appears that such an hypothesis or selectivity is equivalent to an intention or assertion of a specific relation as obtaining in the environment. In the above case these assertions are to the effect that it is such and such types of door which lead on and such and such other types which are closed. The rats assert-hypothesize-that it is the right hand doors, or the left hand doors, or alternate right and left doors or dark doors, or whatever, which, as such, lead on. And when any one such assertion proves incorrect, an animal sooner or later drops it for a new one.

In the experiment as thus far described, the problem given to the animals was actually insoluble. The correct doors were, that is, determined by chance. And no hypotheses--none of the systematic selectivities in the behaviors of the animals-could prove successful. This meant that during the entire duration of the experiment the rats kept shifting from one hypothesis to another. In other experiments, however, the situation was different. Thus in one case it was arranged that after a rat had once adopted some hypothesis with a given degree of consistency the experimenter then made that hypothesis correct. Under these conditions the animals persisted in their now correct hypothesis throughout the entire remainder of the investigation. Or, again in still another set-up, a certain hypothesis was made correct from the very beginning. In such a case the rats might begin with various wrong hypotheses. But they always ended sooner or later with the single correct one. So much for rats, let us turn now to human beings….