Verbal Control and the Descriptive Completeness of Rules in Concept-Identification Tasks

Abstract

<p>When Ss on a card-sorting task were required to make a rapid sorting response prior to stating their rule on each trial, actual frequencies of correct classification were found to be significantly higher than the frequencies predicted from the Ss' trial-by-trial rules. These disparities were observed even though virtually every placement was consistent with the rule given on the same trial. No disparities were found when Ss stated their rule prior to placing each card.</p> <p>The observed disparities indicate that the stated rules were insufficient to describe all of the stimulus cues used in determining the placements made. They also suggest that verbal rules do not necessarily control above-chance sorting performance unless the experimental conditions encourage verbal control over responding.</p>