Diagnosticity and prototypicality in category learning: a comparison of inference learning and classification learning

Category knowledge allows for both the determination of category membership and an understanding of what the members of a category are like. Diagnostic information is used to determine category membership; prototypical information reflects the most likely features given category membership. Two experiments examined 2 means of category learning, classification and inference learning, in terms of sensitivity to diagnostic and prototypical information. Classification learners were highly sensitive to diagnostic features but not sensitive to nondiagnostic, but prototypical, features. Inference learners were less sensitive to the diagnostic features than were classification learners and were also sensitive to the nondiagnostic, prototypical, features. Discussion focuses on aspects of the 2 learning tasks that might lead to this differential sensitivity and the implications for learning real-world categories.     

Sorting out categories: Incremental learning of category structure

Two experiments examine how inferences might promote unsupervised and incremental category learning. Many categories have members related through overall similarity (e.g., a family resemblance structure) rather than by a defining feature. However, when people are asked to sort category members in a category construction task, they often do so by partitioning on a single feature. Starting from an earlier result showing that pairwise inferences increase family resemblance sorting (Lassaline & Murphy, 1996), we examine how these inferences lead to learning the family resemblance structure. Results show that the category structure is learned incrementally. The pairwise inferences influence participants’ weightings of feature pairs that were specifically asked about, which in turn affects their sorting. The sorting then allows further learning of the categorical structure. Thus, the inferences do not directly lead learners to the family resemblance structure, but they do provide a foundation to build on as the participants make additional judgments.     

From children's hands to adults' ears: gesture's role in the learning process

Children can express thoughts in gesture that they do not express in speech--they produce gesture-speech mismatches. Moreover, children who produce mismatches on a given task are particularly ready to learn that task. Gesture, then, is a tool that researchers can use to predict who will profit from instruction. But is gesture also useful to adults who must decide how to instruct a particular child? We asked 8 adults to instruct 38 third- and fourth-grade children individually in a math problem. We found that the adults offered more variable instruction to children who produced mismatches than to children who produced no mismatches--more different types of instructional strategies and more instructions that contained two different strategies, one in speech and the other in gesture. The children thus appeared to be shaping their own learning environments just by moving their hands. Gesture not only reflects a child's understanding but can play a role in eliciting input that could shape that understanding. As such, it may be part of the mechanism of cognitive change.     

Overconfidence effects in category learning: a comparison of connectionist and exemplar memory models

Exemplar and connectionist models were compared on their ability to predict overconfidence effects in category learning data. In the standard task, participants learned to classify hypothetical patients with particular symptom patterns into disease categories and reported confidence judgments in the form of probabilities. The connectionist model asserts that classifications and confidence are based on the strength of learned associations between symptoms and diseases. The exemplar retrieval model (ERM) proposes that people learn by storing examples and that their judgments are often based on the first example they happen to retrieve. Experiments 1 and 2 established that overconfidence increases when the classification step of the process is bypassed. Experiments 2 and 3 showed that a direct instruction to retrieve many exemplars reduces overconfidence. Only the ERM predicted the major qualitative phenomena exhibited in these experiments.     

Learning categories at different hierarchical levels: a comparison of category learning models

Three formal models of category learning, the rational model (Anderson, 1990), the configural-cue model (Gluck & Bower, 1988a), and ALCOVE (Kruschke, 1992), were evaluated on their ability to account for differential learning of hierarchically structured categories. An experiment using a theoretically challenging category structure developed by Lassaline, Wisniewski, and Medin (1992) is reported. Subjects learned one of two different category structures. For one structure, diagnostic information was present along a single dimension (1-D). For the other structure, diagnostic information was distributed across four dimensions (4-D). Subjects learned these categories at a general or at a specific level of abstraction. For the 1-D structure, specific-level categories were learned more rapidly than general-level categories. For the 4-D structure, the opposite result was observed. These results proved highly diagnostic for evaluating the models—although ALCOVE provided a good account of the observed results, the rational model and the configural-cue model did not.     

Ten-month-olds' selective use of visual dimensions in category learning

There is now general consensus that infants can use several different visual properties as the basis for categorization. Nonetheless, little is known about when and whether infants can be guided by contextual information to select the relevant properties from amongst those available to them. We show here that by 10 months of age infants can be biased, through observational learning, to use one or the other of two object properties for classification. Two groups of infants watched an actress classifying objects by either shape (the Shape group) or surface pattern (the Pattern group). When subsequently presented with two test trials which contradicted either one or the other of the classification rules, infants in the two groups looked longer to the classification event that was incompatible with the rule that group had been familiarized to. These results are discussed with reference to the development of selective feature processing in infancy.     

Base-rate effects in category learning: a comparison of parallel network and memory storage-retrieval models

Exemplar-memory and adaptive network models were compared in application to category learning data, with special attention to base rate effects on learning and transfer performance. Subjects classified symptom charts of hypothetical patients into disease categories, with informative feedback on learning trials and with the feedback either given or withheld on test trials that followed each fourth of the learning series. The network model proved notably accurate and uniformly superior to the exemplar model in accounting for the detailed course of learning; both the parallel, interactive aspect of the network model and its particular learning algorithm contribute to this superiority. During learning, subjects' performance reflected both category base rates and feature (symptom) probabilities in a nearly optimal manner, a result predicted by both models, though more accurately by the network model. However, under some test conditions, the data showed substantial base-rate neglect, in agreement with Gluck and Bower (1988b).     

Manufacturing a cartesian closed category with exactly two objects out of a C-monoid

A construction is described of a cartesian closed category A with exactly two elements out of a C-monoid such that can be recovered from A without reference to the construction.The first author was partially supported by the Dutch government through the SPIN project PRISMA; the second author was partially supported by the EEC through Esprit project 415.     

The ghost condition: imitation versus emulation in young children's observational learning

Although observational learning by children may occur through imitating a modeler's actions, it can also occur through learning about an object's dynamic affordances--a process that M. Tomasello (1996) calls "emulation." The relative contributions of imitation and emulation within observational learning were examined in a study with 14- to 26-month-old children. The effectiveness of a "ghost" condition, in which the effective operation of the means apparatus was seen to occur without human agency, was compared with that of a standard modeling procedure in which the child saw an experimenter demonstrate the means action. The ghost condition was as likely to encourage observational learning as was the modeling condition; indeed, performance in the ghost condition was significantly better. The role of emulation in the development of observational learning is discussed in the context of a possible form of goal directedness without agency.     

SUSTAIN: a network model of category learning

SUSTAIN (Supervised and Unsupervised STratified Adaptive Incremental Network) is a model of how humans learn categories from examples. SUSTAIN initially assumes a simple category structure. If simple solutions prove inadequate and SUSTAIN is confronted with a surprising event (e.g., it is told that a bat is a mammal instead of a bird), SUSTAIN recruits an additional cluster to represent the surprising event. Newly recruited clusters are available to explain future events and can themselves evolve into prototypes-attractors-rules. SUSTAIN's discovery of category substructure is affected not only by the structure of the world but by the nature of the learning task and the learner's goals. SUSTAIN successfully extends category learning models to studies of inference learning, unsupervised learning, category construction, and contexts in which identification learning is faster than classification learning.     

Young children's learning of size comparison strategies: effect of origami exercises

The authors examined the process of change in size comparison strategies among preschool-age children. Twelve 5- and 6-year-old children were provided with origami exercises and size comparison tasks for 5 days consecutively, and another twelve 5- and 6-year-old children received size comparison tasks only. Children's strategies for the size comparison tasks were analyzed in terms of placement and adjustment. In contrast with the children in the control condition, the children who were provided with exercises in origami increased their use of superimposition and adjustment strategies by 2 dimensions across the 5 days. The authors argue that origami exercises helped the children connect their implicit understanding of relative size with the use of more effective strategies for size comparison.     

Context and cue cards in young children's testimony: a comparison of brief narrative elaboration and context reinstatement

In a comparison of interview procedures that aimed to assist retrieval strategies in children's event recall, 72 children aged 4-6 years were questioned about a witnessed event. Context reinstatement and brief narrative elaboration (an abbreviated version of K. J. Saywitz and L. Snyder's [1996] procedure) produced similar levels of correct recall, and both elicited more correct recall than did a control condition, with no increase in errors. Combining these procedures did not further improve performance. The superiority of narrative elaboration over the control group was evident in free recall and did not depend on explicit prompting with cue cards. These findings suggest that incorporating brief narrative elaboration training in investigative interviews with children may, like context reinstatement, be a valuable mnemonic aid.     

Developmental change in young children's use of haptic information in a visual task: the role of hand movements

Preschoolers who explore objects haptically often fail to recognize those objects in subsequent visual tests. This suggests that children may represent qualitatively different information in vision and haptics and/or that children’s haptic perception may be poor. In this study, 72 children (2½-5 years of age) and 20 adults explored unfamiliar objects either haptically or visually and then chose a visual match from among three test objects, each matching the exemplar on one perceptual dimension. All age groups chose shape-based matches after visual exploration. Both 5-year-olds and adults also chose shape-based matches after haptic exploration, but younger children did not match consistently in this condition. Certain hand movements performed by children during haptic exploration reliably predicted shape-based matches but occurred at very low frequencies. Thus, younger children’s difficulties with haptic-to-visual information transfer appeared to stem from their failure to use their hands to obtain reliable haptic information about objects.     

Semisupervised category learning: The impact of feedback in learning the information-integration task

In a standard supervised classification paradigm, stimuli are presented sequentially, participants make a classification, and feedback follows immediately. In this article, we use a semisupervised classification paradigm, in which feedback is given after a prespecified percentage of trials only. In Experiment 1, feedback was given in 100%, 0%, 25%, and 50% of the trials. Previous research reported by Ashby, Queller, and Berretty (1999) indicated that in an information-integration task, perfect accuracy was obtained supervised (100%) but not unsupervised (0%). Our results show that in both the 100% and 50% conditions, participants were able to achieve maximum accuracy. However, in the 0% and the 25% conditions, participants failed to learn. To discover the influence of the no-feedback trials on the learning process, the 50% condition was replicated in Experiment 2, substituting unrelated filler trials for the no-feedback trials. The results indicated that accuracy rates were similar, suggesting no impact of the no-feedback trials on the learning process. The possibility of ever learning in a 25% setting was also researched in Experiment 2. Using twice as many trials, the results showed that all but 2 participants succeeded, suggesting that only the total number of feedback trials is important. The impact of the semisupervised learning results for ALCOVE, COVIS, and SPEED models is discussed.     

The genesis and use of exemplar vs. prototype knowledge in abstract category learning

Accurate classification of new exemplars in an abstraction paradigm may be due to their similarity to old exemplars rather than to abstract category (or prototype) knowledge. In the present study, subjects received initial training on a two-category problem before being transferred to a task in which half of the exemplar-response pairs had their responses reversed while the remaining half of the pairs were unchanged. When transfer occurred with no delay and involved old exemplars, more errors occurred for changed than for unchanged pairs. This result implies the use of exemplar-specific rather than abstract category knowledge. However, when transfer was delayed by 24 or 72 h, errors occurred equally often for changed and unchanged pairs. This result suggests that exemplar-specific knowledge is no longer used. Since subjects were still able to accurately classify exemplars prior to the transfer task at these delays, some form of abstract category knowledge is implicated.