Acquired Distinctiveness and Equivalence in Human Discrimination Learning: Evidence for an Attentional Process
Bonardi, Charlotte, Graham, Steven, Hall, Geoffrey, Mitchell, Chris, Psychonomic Bulletin & Review
In a first stage of training, participants learned to associate four visual cues (two different colors and two different shapes) with verbal labels. For Group S, one label was applied to both colors and another to both shapes; for Group D, one label was applied to one color and one shape, and the other label to the other cues. When subsequently required to learn a task in which a given motor response was required to one of the colors and one of the shapes, and a different response to the other color and the other shape, Group D learned more readily than Group S. The task was designed so that the associations formed during the first stage of training could not generate differential transfer to the second stage. The results are consistent, however, with the proposal that training in which similar cues are followed by different outcomes will engage a learning process that boosts the attention paid to features that distinguish these cues.
In a recent report, Hall, Mitchell, Graham, and Lavis (2003) made use of experimental designs and theoretical concepts derived from studies of associative learning in animals to further the analysis of acquired equivalence and distinctiveness effects in human discrimination learning. In their basic experimental design (based on one used with pigeons by Bonardi, Rey, Richmond, & Hall, 1993; see also Kaiser, Sherburne, Steirn, & Zentall, 1997), human participants received initial training in which four different geometrical figures (A, B, C, and D) were used to signal two different outcomes. Presentations of A and B were both followed by, for example, the presentation of the nonsense syllable wug; presentations of C and D were both followed by the nonsense syllable zif. In the next stage, the participants were required to learn a discrimination. In the consistent condition, they had to make one motor response (e.g., to press a key on the left of a keyboard) to presentations of A and of B, and a different motor response (to press a key on the right) to C and to D; in the inconsistent condition, one response was required to A and C and the other response to B and D.
The discrimination was acquired more readily in the consistent than in the inconsistent condition. That is, performance was superior when the task required the participants to make the same response to cues that had shared a common outcome in the first stage of training, and different responses to cues that had been trained initially with different outcomes. The observation that training in which two cues are associated with a common event can enhance generalization between them has been called the acquired equivalence effect; the observation that discrimination between two similar cues will be facilitated by prior training in which each has been associated with a different outcome has been called the acquired distinctiveness effect (see Hall, 1991, for a review).
Hall et al. (2003) considered two possible explanations for their results, one based on an extension of standard associative learning principles and the other on learned changes in attention. The associative account assumes that in the first stage of training each of the four cues will become associated with its outcome, so that on subsequent presentations A and B will both tend to activate the representation of wug, and C and D the representation of zif. These outcome representations will thus be activated during the discrimination stage of the procedure. When the subject learns to make a particular response to a given cue (e.g., to respond "left" to A), the associate of A will be activated and will also become a cue for performing that response. Since B also activates this associate, a tendency to make the same response will be elicited immediately when B is presented, which will facilitate performance for participants in the consistent condition. For participants in the inconsistent condition (who are required to respond "right" to B), this tendency will need to be overcome and will detract from efficient discrimination performance. …