Verbal-nonverbal interactions during a computerized adaptation of Catania et al.'s (1982) experimental task: The effects of embedding an aversive contingency in the nonverbal component

A computerized adaptation of the task designed by Catania et al. (1982) to study verbal-nonverbal interactions was developed. A random-interval random-ratio multiple schedule of reinforcement (MSR) operated on buttons appearing on the screen, which participants clicked to earn points exchangeable for money. Automatic shaping of verbal reports entailed multiple-schedule interruptions, during which participants completed blanks with guesses regarding multiple-schedule performance and earned points depending on the quality of their guesses. The study tested the adapted task and extended the notion that discriminative schedule control plays a role in verbal-nonverbal interactions (Torgrud & Holborn, 1990) by scheduling a response cost (RC) for high rates during the random-interval component. Sixteen participants were exposed to four conditions: MSR only; MSR + shaping; MSR + RC; and MSR + RC + shaping. Differential button-pressing rates were typically established during RC conditions and systematically preceded successful shaping of guesses. Differential responding ceased after removal of the RC, despite the fact that congruent guessing was previously established. This finding indicated poor verbal governance, and does not support the notion that low discriminative control promotes verbal control. RC produced random-ratio response rates that matched the low rates that were characteristic of random-interval performance. It is suggested that verbal-nonverbal interactions may be responsible for this effect.