Blocking in category learning.

Abstract

Many theories of category learning assume that learning is driven by a need to minimize classification error. When there is no classification error, therefore, learning of individual features should be negligible. We tested this hypothesis by conducting three category learning experiments adapted from an associative learning blocking paradigm. Contrary to an error-driven account of learning, participants learned a wide range of information when they learned about categories, and blocking effects were difficult to obtain. Conversely, when participants learned to predict an outcome in a task with the same formal structure and materials, blocking effects were robust and followed the predictions of error-driven learning. We discuss our findings in relation to models of category learning and the usefulness of category knowledge in the environment.

Prior knowledge enhances the category dimensionality effect

“…In this respect, category learning is different from classical conditioning or other forms of cue learning in which people do not have the belief that most properties should be learned. Indeed, recent work from our lab has shown that an analogue of the classical conditioning blocking effect is not found when people believe they are learning categories but is found when the same problem is construed as predicting the computer's response (Bott, Hoffman, & Murphy, 2007). KRES's success suggests an alternative account of the effect of dimensionality, based to do with an aspect of KRES's learning rule.…”

Feature inference learning and eyetracking

“…For example, if a category seems realistic, or useful, subjects may then be motivated to learn about it. Conceivably, our use of visual features (vs. the semantic features used by Bott et al, 2007) and opaque (and meaningless) category labels ''A" and ''B" (vs. ''Mobbles" or ''Streaths") meant that our stimuli were not sufficiently ''category like" for our participants to believe there was anything interesting to learn about them above and beyond the experimental task. In addition, our experimental instructions may have emphasized the inference task so strongly that we inadvertently undermined any motivation to learn about the categories that might have otherwise existed.…”

The costs of supervised classification: The effect of learning task on conceptual flexibility.

“…Moreover, supervised classification can yield more learning than other procedures. For example, Bott, Hoffman, and Murphy (2007) found that learners acquired more information when predicting an outcome they knew to be a category label compared with predicting a meaningless outcome (a low or high tone). This occurred despite the use of a "blocking" paradigm in which both groups of subjects were first trained on a single cue that predicted the outcome perfectly (also see Hoffman & Murphy, 2006).…”