Academic journal article Psychonomic Bulletin & Review

A Case for Restricted-Domain Relational Learning

Academic journal article Psychonomic Bulletin & Review

A Case for Restricted-Domain Relational Learning

Article excerpt

Monkeys and pigeons learned a same/different task with pairs that were selected from a training set of eight picture stimuli. They showed no novel-stimulus transfer and hence no abstract-concept learning. They were also tested with novel pairs of the eight training pictures (i.e., combinations that had not been used in training) and with inverted pictures of the training pairs. If the subjects had learned the task item-specifically (e.g., if-then or configural learning), they should have failed these tests, but they performed well with novel combinations of training pictures and inverted pictures, suggesting that they learned the task relationally (i.e., on the basis of the relationship between the two pictures that were presented in each trial). This somewhat paradoxical conclusion of relational learning in the absence of abstract-concept learning is contrary to most theories of abstract-concept learning. The implications of this conclusion are discussed in the context of restricted-domain relational learning.

Children initially process the visual domain in terms of specific item features and later transition to relationships between or among different visual stimuli (see, e.g., Doumas, Hummel, & Sandhofer, 2008; Gentner, 2003; Gentner & Rattermann, 1991). Relational learning begins as global similarity (i.e., identity), but it can later extend to shared attributes (see, e.g., Siegler & Alibali, 2004; Smith, 1984, 1989). The transfer of relational learning to novel stimuli has been the hallmark of abstract-concept learning, and it is based on an overarching relational rule that is applied to pairs of novel stimuli.

Failures of novel-item relational transfer have been taken not only as evidence against abstract-concept learning, but also as evidence for item-specific learning. This assumption of item-specific learning when concept learning fails has a long history in determining which species have the cognitive capability to learn abstract concepts (see, e.g., Carter & Werner, 1978; Premack, 1978, 1983a, 1983b; Thomas, 1980). Carter and Werner (1978) and Premack (1978) elaborated on Skinner's (1950) suggestion that pigeons might learn specific stimulus-response chains, or if-then associations, for each sample and correct comparison choice, and on Lashley's (1938) suggestion that animals might learn configural patterns of stimulus combinations (i.e., gestalt configurations). Why these hypotheses remained untested for so many decades is a bit of a mystery. Equally puzzling is why these untested hypotheses were accepted as fact (fait accompli) for so many decades.

Our first indication that something might be amiss in assuming item-specific learning when novel-stimulus transfer fails came from studies in which pigeons, rhesus monkeys, and capuchin monkeys learned a same/different (S/D) task with 8 picture items but showed no novelstimulus transfer (Katz & Wright, 2006; Katz, Wright, & Bachevalier, 2002; Wright & Katz, 2006; Wright, Rivera, Katz, & Bachevalier, 2003). If these subjects had learned the task item-specifically, then in a series of training-set expansions (16, 32, 64, 128, etc.), they should at some point have switched from item-specific learning to relational learning and shown a sudden increase in novelstimulus transfer. There was no sudden increase or discontinuity in the novel-stimulus transfer function, however; novel-stimulus transfer increased at a steady (linear) rate with training-set doublings.

On the one hand, these abstract-concept learning results were encouraging: Abstract-concept learning was found across pigeons and two different monkey species. On the other hand, these results were puzzling. If these species did not learn the S/D task item-specifically when they failed the test for abstract-concept learning, what did they learn?

The purpose of the present research was to train other groups of pigeons and monkeys in a similar eight-item S/D task and then test them for item-specific learning when they failed to show abstract-concept learning. …

Search by... Author
Show... All Results Primary Sources Peer-reviewed


An unknown error has occurred. Please click the button below to reload the page. If the problem persists, please try again in a little while.