The effects of verbal performance descriptions on nonverbal operant responding.

The effect of reinforced verbal descriptions on key-pressing rate was studied in the context of reinforcement for pressing on schedules opposed to the verbal description and schedules delivering a constant or randomly chosen point value regardless of pressing rate (nondifferential schedules). Undergraduates' key presses produced points exchangeable for lottery tickets on alternating schedules. Subjects experienced one of four manipulations. In Experiment 1, after schedule control had been demonstrated using a "medium rate" schedule for each of two response keys, subjects were awarded maximum points for choosing one of five verbal descriptions of "the best way to earn points" for each key. Subjects experienced either (a) maximum points for verbal descriptions of "press very fast" for one key and "press very slowly" for the other, with the schedule gradually moved from medium to oppose this description, or (b) maximum schedule points for a very fast rate on one key and very slow rate on the other, with the maximum points for verbal descriptions gradually moved to oppose the schedule. Key-pressing rates conformed to the active schedule, not to the verbal performance description. In Experiment 2 subjects received maximum points for verbal descriptions of "press very fast" for one key and "press very slowly" for the other while the same nondifferential schedule was operative for both keys. Correspondence of pressing rate to verbal description was either complete, transient, or absent. The precise discriminative control of the schedules employed may account for less verbal response-rate control in the present versus past research. Possible differences between computer- and experimenter-generated verbal behavior are discussed.