Emergence of Conditional Discriminations after Constructed-Response Matching-to-Sample Training

Abstract

Mackay (1985) reported that subjects were able to match printed words to colors after learning to construct the color names from a pool of letters. Visual feedback from the constructed color names might have been responsible for the emergent matching to sample. In this study we prevented visual feedback during the construction procedure. Also, in matching-to-sample tests Mackaÿs subjects might simply have reached for the first letter of a comparison name, as if to begin construction, and a selection of the whole word would have been recorded. In this study, subjects constructed combinations of three arbitrary forms, with each combination composed of a different sequence of the same three forms. In the subsequent matching-to-sample test, subjects could not select a comparison on the basis of a single element because all comparisons were made up of the same elements. Even with feedback and element sequence controlled, the subjects showed nearly perfect performances in the matching-to-sample tests. These results indicated that the emergent matching-to-sample performances did not require visual feedback from the constructed stimuli and were not artifacts of the sequence of elements in the comparison stimuli.