Perceptual Fluency Affects Categorization Decisions - eScholarship

Perceptual Fluency Affects Categorization Decisions Sarah J. Miles (smiles25@uwo.ca) and John Paul Minda (jpminda@uwo.ca) Department of Psychology The University of Western Ontario London, ON N6A 5C2 Abstract Learning in the prototype distortion task is thought to involve perceptual learning in which category members experience an enhanced visual response (Ashby & Maddox, 2005). This re- sponse likely leads to more efficient processing, which in turn may result in a feeling of perceptual fluency for category mem- bers. We examined the perceptual fluency hypothesis by ma- nipulating fluency independently from category typicality. We predicted that when perceptual fluency was induced using sub- liminal priming, this fluency would be misattributed to cate- gory membership and would affect categorization decisions. In a prototype distortion task, participants were more likely to judge non-members as category members when they were made perceptually fluent with a matching subliminal prime. This result suggests that perceptual fluency can be reflective of category membership and may be used as a cue during some categorization decisions. In addition, the results provide con- verging evidence that some types of categorization are based on perceptual learning. Keywords: Category Learning; Prototype Learning; Percep- tual Fluency; Subliminal Priming The prototype distortion task was first introduced by Pos- ner and Keele (1968) as a method of studying how category representations are abstracted and stored. Although varia- tions of the task have been used, in the general form of the task participants are exposed to a series of dot patterns that are distortions of a common prototype and form a category. Next, participants judge whether a series of new dot patterns are also category members. The pattern of responses given by participants is thought to reflect the nature of the cate- gory representation used to make categorization judgments. More recently, the task has been used to investigate the role of perceptual learning in categorization (Casale & Ashby, 2008; Coutinho, Couchman, Redford, & Smith, 2010). FMRI studies involving the prototype distortion task have shown that perceptual learning may be important for abstract- ing visual prototypes. In three studies (Aizenstein et al., 2000; P. J. Reber, Stark, & Squire, 1998a, 1998b) visual ar- eas in the occipital cortex showed decreased activity in re- sponse to category members relative to non-members. The authors suggested that this decrease in activation might re- flect easier or faster processing of category members, similar to the type of processing fluency found in repetition prim- ing studies. More specifically, a group of visual cortical cells may learn to respond strongly to the prototype, less to non- prototypical category members and even less to items that are not in the category. Across the learning period, percep- tual learning causes the cells’ sensitivity and magnitude of response to increase (Ashby & Maddox, 2005). Because the increased response is only elicited for category members, its presence can be used as a cue to category membership. The perceptual representation system is thought to be an implicit memory system that supports the improved process- ing of previously seen stimuli, as described above (Schacter, 1990; Tulving & Schacter, 1990). The perceptual represen- tation system is also particularly sensitive to the similarity among stimuli. It can generalize across similar stimuli but not dissimilar stimuli, a process that is important for cate- gorization (Cooper, Schacter, Ballesteros, & Moore, 1992). Consequently, it has been proposed that the perceptual repre- sentation system could support the type of perceptual learning that is thought to play a role in abstracting visual prototypes (Casale & Ashby, 2008). In a study where participants were trained on exemplars that were either high or low distortions of the prototype, performance was best after training with the low distortion items. These results illustrate that this type of prototype abstraction is dependent on visual similarity and may be mediated by the perceptual representation system. Another study investigating the processes underlying pro- totype learning has come to a slightly different conclusion (Coutinho et al., 2010). In this study, participants were trained on the prototype distortion task either with stimuli that were all the same size during training and test or with stim- uli whose size varied during training and test. Performance was comparable in both versions of the task. The authors concluded that low-level perceptual learning is not the only mechanism for prototype learning because low-level percep- tual learning would have been disrupted by variations in size. Therefore, while it seems that some sort of perceptual learn- ing is important for prototype learning, it is not certain that this learning is supported by the relatively low-level percep- tual representation system. Regardless of the level at which perceptual learning oc- curs, the enhanced visual responding that accompanies cate- gory members could be informative of category membership (Ashby & Maddox, 2005). Perceptual fluency is the feeling of ease or difficulty associated with a mental task (Alter & Oppenheimer, 2009; Oppenheimer & Frank, 2008). Because category members experience an enhanced visual response, this could contribute to a feeling of perceptual fluency for category members but not for non-members. This feeling of fluency may be the cue that is used during judgments of cat- egory membership, especially in tasks such as the prototype distortion task. While perceptual fluency is generally a reliable cue about the state of the environment, it can be independently manip- ulated (Oppenheimer, 2008). For example, subliminal prim- ing, figure ground contrast, stimulus duration and stimulus repetition all affect perceptual fluency. Since perceptual flu-

Récupéré en direct depuis OpenAlex et désinversé. Les résumés ne sont pas conservés dans cette base de données : les index inversés représentent 8,6 Go des 9,3 Go de texte de la base, et le serveur dispose de 13 Go libres.