Similarity, identification, and categorization: Comment on Ashby and Lee (1991).

F. G. Ashby and W. W. Lee (see record 1991-23345-001) tested various models derived from the general recognition theory (GRT; Ashby and N. A. Perrin, 1988; Ashby and J. T. Townsend, 1986) on their ability to predict and interrelate similarity, categorization, and identification performance. This commentary (1) argues that contrary to Ashby and Lee's suggestion, the likelihood-based GRT cannot generally predict categorization from identification without incorporating selective attention, (2) argues that the categorization rule in the likelihood-based GRT is extremely close in spirit to R. M. Nosofsky's (1986) exemplar-based similarity model, (3) reports new model-based analyses that call into question Ashby and Lee's interpretation of their identification-confusion data, (4) raises questions about the identification and similarity models tested by Ashby and Lee, and (5) criticizes Ashby and Lee's methods of fitting and evaluating the various models. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Selective attention and the formation of linear decision boundaries: Reply to Maddox and Ashby (1998).

In this reply to W. T. Maddox and F. G. Ashby's (see record 1997-38560-018) commentary, the author argues that (a) Maddox and Ashby's current stance represents a marked departure from their previously published claims about the unimportance of selective attention in categorization, (b) they are inconsistent with their own work when they criticize S. C. McKinley and R. M. Nosofsky's (see record 83-21759) tests of the linear-boundary models, (c) their arguments about modeling averaged data have no bearing on the central conclusions reached by McKinley and Nosofsky, and (d) they make incorrect assertions regarding the application and predictions of the exemplar model. Finally, the author defends the theoretical progress that has been made in recent years with the exemplar model and argues instead that it is the decision-bound theory of Ashby and Maddox that is in need of greater constraints. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Confusable and discriminable stimuli: Comment on Nosofsky (1986) and Shepard (1986).

R. M. Nosofsky (see record 1986-16297-001) modeled identification and categorization performance with highly similar stimuli by using a model involving a "Gaussian" relationship between similarity and Euclidean distance. Nosofsky found that this model fitted his data better than a model involving similarity as an exponential decay function of city-block distance. Nosofsky's stimuli evoked independent perceptual dimensions. By using a multivariate model of similarity, in which perceptual variation is included, this article provides a possible reconciliation of the kind of result found by Nosofsky (1986) and Shepard's (1986) theory concerning the relationship between similarity and perceptual distance. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Learning and attention in multidimensional identification and categorization: Separating low-level perceptual processes and high-level decisional processes.

Four observers completed identification and categorization tasks. Learning and attention processes were examined by applying general recognition theory (F. G. Ashby & J. T. Townsend, 1986), which separates perceptual, decisional, and attentional processes. Learning led to decision regions that became more nearly optimal. Learning had little effect on perceptual processes in identification and decisional integration categorization tasks but affected perceptual processes in decisional selective attention categorization tasks, leading to perceptual selective attention. These findings suggest that (a) identification and categorization invoke decision strategies that are localized in the striatum (F. G. Ashby, E. M. Waldron, W. W. Lee, & A. Berkman, 2001) and (b) perceptual and decisional attention systems exist and are mediated by distinct brain structures (M. I. Posner & S. E. Petersen, 1990). (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Gaussian general recognition theory and perceptual independence.

Recently, F. G. Ashby and J. T. Townsend (1986) introduced a theory of perceptual identification, the General Recognition Theory (GRT), that rigorously defined previously obscure notions of perceptual independence of stimulus dimensions. Also, they established many theoretical relationships between these different types of independence whereby experimenters could obtain support for (or falsification of) one type or another. Ashby and Townsend's Theorem 4 stated several assumptions, including optimality, are sufficient for a strong form of perceptual independence. It is shown here by counterexample that this theorem is false. A review of the basic concepts of GRT is provided, with a discussion of the aforementioned limitation and its implications. An amended version of Theorem 4 is offered. Simulation results illustrate optimistic prospects for testing for independence in lieu of the counterexample. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Attention, similarity, and the identification–categorization relationship.

A unified quantitative approach to modeling Ss' identification and categorization of multidimensional perceptual stimuli is proposed and tested. Two Ss identified and categorized the same set of perceptually confusable stimuli varying on separable dimensions. The identification data were modeled using R. N. Shepard's (see record 1959-05134-001) multidimensional scaling-choice framework, which was then extended to model the Ss' categorization performance. The categorization model, which generalizes the context theory of classification developed by D. L. Medin and M. M. Schaffer (see record 1979-12633-001), assumes that Ss store category exemplars in memory. Classification decisions are based on the similarity of stimuli to the stored exemplars. It is assumed that the same multidimensional perceptual representation underlies performance in both the identification and categorization paradigms. However, because of the influence of selective attention, similarity relationships change systematically across the 2 paradigms. Findings provide some support for the hypothesis that Ss distribute attention among component dimensions so as to optimize categorization performance and that Ss may have augmented their category representations with inferred exemplars. Results demonstrate that excellent predictions of categorization performance can be made given knowledge of performance in an identification paradigm. (51 ref) (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

The role of similarity in artificial grammar learning.

The authors examine the role of similarity in artificial grammar learning (AGL; A. S. Reber, 1989). A standard finite-state language was used to create stimuli that were arrangements of embedded geometric shapes (Experiment 1), connected lines (Experiment 2), and sequences of shapes (Experiment 3). Main effects for well-known predictors from the literature (grammaticality, associative global and anchor chunk strength, novel global and anchor chunk strength, length of items, and edit distance) were observed, thus replicating previous work. However, the authors extend previous research by using a widely known similarity-based exemplar model of categorization (the generalized context model; R. M. Nosofsky, 1989) to fit grammaticality judgments, by nested regression analyses. The results suggest that any explanation of AGL that is based on the existing theories is incomplete without a similarity process as well. Also, the results provide a foundation for further interpreting AGL in the wider context of categorization research. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Exemplar representation without generalization? Comment on Smith and Minda's (2000) "Thirty categorization results in search of a model."

J. D. Smith and J. P. Minda (see record 1999-15928-001) conducted a meta-analysis of 30 data sets reported in the classification literature that involved use of the "5–4" category structure introduced by D. L. Medin and M. M. Schaffer (1978). The meta-analysis was aimed at investigating exemplar and elaborated prototype models of categorization. In this commentary, the author argues that the meta-analysis is misleading because it includes many data sets from experimental designs that are inappropriate for distinguishing the models. Often, the designs involved manipulations in which the actual 5–4 structure was not, in reality, tested, voiding the predictions of the models. The commentary also clarifies various aspects of the workings of the exemplar-based context model. Finally, concerns are raised that the all-or-none exemplar processes that form part of Smith and Minda's (see record 1999-15928-001) elaborated prototype models are implausible and lacking in generality. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Attention and learning processes in the identification and categorization of integral stimuli.

Studied the relationship between identification and categorization learning of integral-dimension stimuli within the framework of an exemplar-based generalization model. The model was used to predict learning in 6 different categorization conditions on the basis of data obtained in a single identification learning condition. 34 undergraduates were tested in the identification learning condition and 191 undergraduates were tested in the categorization learning conditions. An assumption in the model is that because of selective attention to component dimensions, similarity relations may change in systematic ways across different experimental contexts. Results indicate that selective attention may play a critical role in determining the identification–categorization relationship for integral stimuli. Similarity among exemplars decreased as a function of identification learning. It is noted that alternative classification models, including prototype, multiple-prototype, average distance, and value-on-dimensions models, were unable to account for the results. (50 ref) (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Base-rate effects in multidimensional perceptual categorization

The optimality of human performance when category base rates differ was investigated in 2 multidimensional perceptual categorization tasks. All participants were sensitive to differences in base rate, even during their 1st experimental session. Nearly half of the participants learned the optimal decision bound by their final experimental session. Little evidence for conservative cutoff placement was found (i.e., an underestimation of category base-rate differences). In fact, participants who did not learn the optimal decision bound tended to use a decision bound that overestimated the base-rate difference. Across all conditions participants showed a clear shift toward the optimal decision bound with experience. These data suggest that experienced participants are highly sensitive to differences in category base rate. The model-based analyses suggest that the decision-bound model of categorization (Ashby, 1992a; Ashby & Maddox, 1993; Maddox & Ashby, 1993) provides a powerful tool for investigating the limits of human categorization performance.     

Selective attention and the formation of linear decision boundaries: Comment on McKinley and Nosofsky (1996).

S. C. McKinley and R. M. Nosofsky (see record 83-21756) compared a linear decision-bound model with the generalized context model (GCM) in their ability to account for categorization data from experiments that used integral- or separable-dimension stimuli and required selective attention or attention to both dimensions. McKinley and Nosofsky found support for the GCM and concluded that decision-bound theory needs to incorporate assumptions about selective attention. In this commentary it is argued that (a) unlike the GCM, decision-bound theory provides a framework for independently investigating perceptual and decisional forms of selective attention; (b) the effect of stimulus integrality on the form of the optimal decision bound is misinterpreted; (c) averaged data is biased against decision-bound theory and toward the GCM; (d) many a priori predictions of the GCM are violated empirically; and (e) exemplar theory has lost much of its initial theoretical structure. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Prototype and Exemplar Accounts of Category Learning and Attentional Allocation: A Reassessment.

In a recent article. J. P. Minda and J. D. Smith (2002; see record 2002-00620-002) argued that an exemplar model provided worse quantitative fits than an alternative prototype model to individual subject data from the classic D. L. Medin and M. M. Schaffer (1978) 5/4 categorization paradigm. In addition, they argued that the exemplar model achieved its fits by making untenable assumptions regarding how observers distribute their attention. In this article, we demonstrate that when the models are equated in terms of their response-rule flexibility, the exemplar model provides a substantially better account of the categorization data than does a prototype or mixed model. In addition, we point to shortcomings in the attention-allocation analyses conducted by J. P. Minda and J. D. Smith (2002). When these shortcomings are corrected, we find no evidence that challenges the attention-allocation assumptions of the exemplar model. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Separating perceptual and decisional attention processes in the identification and categorization of integral-dimension stimuli.

Four observers performed matching, identification, and categorization with stimuli that varied along the integral dimensions: brightness and saturation. General recognition theory (F. G. Ashby & J. T. Townsend, 1986) was applied to quantify the separate influences of perceptual and decisional processes within and across tasks, with a focus on separating perceptual from decisional attention processes. Good accounts of the identification data were obtained from perceptual matching representation. This perceptual representation provided a good account of the categorization data, except when decisional selective attention to 1 stimulus dimension was required. Decisional selective attention reduced the attended-dimension perceptual variance relative to the unattended-dimension perceptual variance, with a larger reduction resulting when brightness, as opposed to saturation was attended. Implications for color vision research are discussed. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Comparing Categorization Models.

Four experiments are presented that competitively test rule- and exemplar-based models of human categorization behavior. Participants classified stimuli that varied on a unidimensional axis into 2 categories. The stimuli did not consistently belong to a category; instead, they were probabilistically assigned. By manipulating these assignment probabilities, it was possible to produce stimuli for which exemplar- and rule-based explanations made qualitatively different predictions. F. G. Ashby and J. T. Townsend's (1986) rule-based general recognition theory provided a better account of the data than R. M. Nosofsky's (1986) exemplar-based generalized context model in conditions in which the to-be-classified stimuli were relatively confusable. However, generalized context model provided a better account when the stimuli were relatively few and distinct. These findings are consistent with multiple process accounts of categorization and demonstrate that stimulus confusion is a determining factor as 10 which process mediates categorization. (PsycINFO Database Record (c) 2011 APA, all rights reserved)     

Perceptual separability, decisional separability, and the identification-speeded classification relationship

Three observers participated in several sessions of identification and speeded classification. Predictions from F. G. Ashby and W. T. Maddox's (1994a) response time (RT) theory of speeded classification were tested. A key theoretical concept is decisional separability (DS), that is, that the decision about the level of 1 component does not depend on the perceived value of the other component. The theory predicted that DS would hold in the filtering task but would be violated in the redundancy task, resulting in the redundancy gains. To summarize the results, (a) DS held in the filtering conditions, but not in the redundancy conditions; (b) redundancy gains occurred; (c) despite the redundancy gains, strong evidence was obtained that the stimulus components are perceptually separable; (d) 2 new models that each assumed RT decreases with the distance between the percept and the decision bound provided good accounts of the RT distributions and accuracy rates; and (e) the shift from identification to speeded classification influenced both perceptual and decisional processes.     

Exemplars, Prototypes, and the Flexibility of Classification Models.

J. P. Minda and J. D. Smith (2001) showed that a prototype model outperforms an exemplar model, especially in larger categories or categories that contained more complex stimuli. R. M. Nosofsky and S. R. Zaki (2002) showed that an exemplar model with a response-scaling mechanism outperforms a prototype model. The authors of the current study investigated whether excessive model flexibility could explain these results. Using cross-validation, the authors demonstrated that both the prototype model and the exemplar model with a response-scaling mechanism suffered from overfilling in the linearly separable category structure. The results illustrate the need to make sure that the best-fitting model is not chasing error variance instead of variance attributed to the cognitive process it is supposed to model. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Perceptual separability, decisional separability, and the identification–speeded classification relationship.

Three observers participated in several sessions of identification and speeded classification. Predictions from F. G. Ashby and W. T. Maddox's (1994a) response time (RT) theory of speeded classification were tested. A key theoretical concept is decisional separability (DS), that is, that the decision about the level of 1 component does not depend on the perceived value of the other component. The theory predicted that DS would hold in the filtering task but would be violated in the redundancy task, resulting in redundancy gains. To summarize the results, (a) DS held in the filtering conditions, but not in the redundancy conditions; (b) redundancy gains occurred; (c) despite the redundancy gains, strong evidence was obtained that the stimulus components are perceptually separable; (d) 2 new models that each assumed RT decreases with the distance between the percept and the decision bound provided good accounts of the RT distributions and accuracy rates; and (e) the shift from identification to speeded classification influenced both perceptual and decisional processes. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Response-time tests of logical-rule models of categorization.

A recent resurgence in logical-rule theories of categorization has motivated the development of a class of models that predict not only choice probabilities but also categorization response times (RTs; Fifi?, Little, & Nosofsky, 2010). The new models combine mental-architecture and random-walk approaches within an integrated framework and predict detailed RT-distribution data at the level of individual participants and individual stimuli. To date, however, tests of the models have been limited to validation tests in which participants were provided with explicit instructions to adopt particular processing strategies for implementing the rules. In the present research, we test conditions in which categories are learned via induction over training exemplars and in which participants are free to adopt whatever classification strategy they choose. In addition, we explore how variations in stimulus formats, involving either spatially separated or overlapping dimensions, influence processing modes in rule-based classification tasks. In conditions involving spatially separated dimensions, strong evidence is obtained for application of logical-rule strategies operating in a serial-self-terminating processing mode. In conditions involving spatially overlapping dimensions, preliminary evidence is obtained that a mixture of serial and parallel processing underlies the application of rule-based classification strategies. The logical-rule models fare considerably better than major extant alternative models in accounting for the categorization RTs. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Relations among categorization, induction, recognition, and similarity: Comment on Sloutsky and Fisher (2004).

V. M. Sloutsky and A. V. Fisher (see record 2004-14313-002) reported 5 experiments documenting relations among categorization, induction, recognition, and similarity in children as well as adults and proposed a new model of induction, SINC (similarity, induction, categorization). Those authors concluded that induction depends on perceptual similarity rather than conceptual knowledge. Despite the useful contributions of this work, there are some important limitations. The experimental designs examined a limited range of phenomena that are not the most revealing about the use of nonperceptual information. The main results involved a simple triad task, for which the SINC model's predictions are equivalent to the predictions of previous models of inductive reasoning. It is also unclear whether the SINC model can account for the observed relations between similarity and recognition. Implications for future work on induction and related cognitive activities are discussed. (PsycINFO Database Record (c) 2010 APA, all rights reserved)