Preliminary Suggestions on the Use of Linguistically Defined Units as Variables in Verbal Conditioning Experiments

摘要

Psychological studies in the general area of verbal conditioning have tended to follow this approximate model: a particular type or segment of verbal activity (response) is isolated and its rate of occurrence is controlled or patterned in the test situation by applying a given type or 'schedule' of reinforcement to the subject or subjects. Thirty-one such studies reviewed by Krasner are compared to the more recent experiments of Bachrach and others. The series of experiments in verbal conditioning accomplished under support of the Group Psychology Branch of the Office of Naval Research by Bachrach and various co-workers differ from most of those described by Krasner in that the experimental situation is more rigidly controlled, with monitoring and reinforcing done mechanically (particularly in more recent experiments). As a correlate to this, the responses selected for reinforcement tend to be objectively measurable or countable physical events--amplitude of speech, duration of pause, duration of speaking, and rate of speech (by counting syllables--or amplitude peaks, which may coincide fairly closely with syllable nuclei).