Document Type



In matching-to-sample tasks, symmetry is defined by the interchangeability of samples and their corresponding comparison stimuli. The purpose of the present study was to identify sources of both intra- and intersubject variability in the formation of symmetry. In Experiment 1, samples and correct comparison stimuli were selected according to three levels of presumed sequential dependence: (a) letters in alphabetical order (high sequential dependence), (b) letters in alphabetical order, but skipping every other letter (moderate sequential dependence), and (c) Greek letters (no sequential dependence). The subjects were assigned to three groups, starting with a specific level of sequential dependence, and proceeding with ascending or descending levels of sequentiality. The results showed that both stimulus material and task order influenced the probability of responding in accordance with symmetry versus responding in accordance with the sequential relation during the tests. Classes of reported collateral responses correlating with either of the emergent relations were derived from postexperimental interviews. Subjects' reports of potentially bidirectional verbal classes such as visualizing/naming or repeated naming of sample-comparison combinations correlated with symmetry. Reports of sequentially related responding as in naming letters in alphabetical order correlated with test responding in accordance with the sequential relation. In a second experiment, the subjects were specifically instructed to engage in responses pertaining to the derived classes in successive blocks of repeated exposures to the same matching-to-sample task. The results of Experiment 2 showed that when the subjects complied with instructions to engage in bidirectional verbal sequences, they also responded in accordance with symmetry. When subjects complied with instructions to engage in unidirectional verbal sequences, they responded in accordance with sequence during the tests.