Michael A. Erickson

RULES AND EXEMPLARS IN CATEGORY LEARNING

Psychological theories of categorization generally focus on either rule- or exemplar-based explanations. We present 2 experiments that show evidence of both rule induction and exemplar encoding as well as a connectionist model, ATRIUM, that specifies a mechanism for combining rule- and exemplar-based representation. In 2 experiments participants learned to classify items, most of which followed a simple rule, although there were a few frequently occurring exceptions. Experiment 1 examined how people extrapolate beyond the range of training. Experiment 2 examined the effect of instance frequency on generalization. Categorization behavior was well described by the model, in which exemplar representation is used for both rule and exception processing. A key element in correctly modeling these results was capturing the interaction between the rule- and exemplar-based representations by using shifts of attention between rules and exemplars.

Perceptual match effects in direct tests of memory: The role of contextual fan

The aim of the present study was to determine whether physical attributes of a memory representation would affect explicit memory performance and, if so, what type of factors would affect the size of a perceptual match effect. Subjects studied words in different, uncommon fonts and were later asked whether the word had been studied earlier. Words could be re-presented in the original font, a font studied with another word, or a font not seen earlier. In two additional experiments, we varied the numbers of words studied in the same unusual font (1 vs. 12 words per font). Recognition memory for the words was better if the test and study fonts matched, and this effect was larger for fonts not shared with other words. Moreover, old judgments were most likely to be classified as remember responses when words were re-presented in the same font when it had not been studied with other words. Although we found a significant effect of levels of processing, this factor did not interact with whether the font matched between study and test. These results are consistent with the predictions of the source of activation confusion model of memory and suggest that perceptual information operates according to the same memory principles as conceptual information.

Rule-based extrapolation in perceptual categorization

Erickson and Kruschke (1998) provided a demonstration that in certain situations people will classify novel stimuli according to an extrapolated rule, even when the most similar training exemplar is an exception to the rule. This result challenged exemplar models. Nosofsky and Johansen (2000) have called this finding into question by offering an exemplar-based explanation for those data based on the perceptual features of the stimuli. Here, we describe the results of a new experiment that yields results similar to those found previouslywithout the questionable perceptual features:Participantswho learn to classify all the training stimuli have patterns of generalization that indicate a combination of rule and exemplar representation. ATRIUM, a hybrid rule and exemplar model (Erickson & Kruschke, 1998), is shown to account for these data much better than ALCOVE, an exemplar model (Kruschke, 1992).Moreover, four alternate exemplar explanations, including one suggested by Nosofsky and Johansen, cannot account for our new findings.

Executive attention and task switching in category learning: Evidence for stimulus-dependent representation

One class of multiple-system models of category learning posits that within a single category-learning task people can learn to utilize different systems with different category representations to classify different stimuli. This is referred to as stimulus-dependent representation (SDR). The use of SDR implies that learners switch from subtask to subtask as trials demand. Thus, the use of SDR can be assessed via slowed response times, following a representation switch. Additionally, the use of SDR requires control of executive attention to keep inactive representations from interfering with the current response. Subjects were given a category learning task composed of one- and two-dimensional substructures. Control of executive attention was measured using a working memory capacity (WMC) task. Subjects most likely to be using SDR showed greater slowing of responses following a substructure switch and a greater correlation between learning performance and WMC. These results provide support for the principle of SDR in category learning and the reliance of SDR on executive attention.

Probative Value of Absolute and Relative Judgments in Eyewitness Identification

It is well-accepted that eyewitness identification decisions based on relative judgments are less accurate than identification decisions based on absolute judgments. However, the theoretical foundation for this view has not been established. In this study relative and absolute judgments were compared through simulations of the WITNESS model (Clark, Appl Cogn Psychol 17:629–654, 2003) to address the question: Do suspect identifications based on absolute judgments have higher probative value than suspect identifications based on relative judgments? Simulations of the WITNESS model showed a consistent advantage for absolute judgments over relative judgments for suspect-matched lineups. However, simulations of same-foils lineups showed a complex interaction based on the accuracy of memory and the similarity relationships among lineup members.

Rule-based extrapolation: A continuing challenge for exemplar models

Erickson and Kruschke (1998, 2002) demonstrated that in rule-plus-exception categorization, people generalize category knowledge by extrapolating in a rule-like fashion, even when they are presented with a novel stimulus that is most similar to a known exception. Although exemplar models have been found to be deficient in explaining rule-based extrapolation, Rodrigues and Murre (2007) offered a variation of an exemplar model that was better able to account for such performance. Here, we present the results of a new rule-plus-exception experiment that yields rule-like extrapolation similar to that of previous experiments, and yet the data are not accounted for by Rodrigues and Murre’s augmented exemplar model. Further, a hybrid rule-and-exemplar model is shown to better describe the data. Thus, we maintain that rule-plus-exception categorization continues to be a challenge for exemplar-only models.