Academic journal article Psychonomic Bulletin & Review

Memory for Objects in Canonical and Noncanonical Viewpoints

Academic journal article Psychonomic Bulletin & Review

Memory for Objects in Canonical and Noncanonical Viewpoints

Article excerpt

This article investigates how the perspective from which we see an object affects memory. Object identification can be affected by the orientation of the object. Palmer, Rosch, and Chase (1981) coined the term canonical to describe perspectives in which identification performance is best. We present two experiments that tested the effects of object perspective on memory. Our results revealed a double dissociation between task (recognition and recall) and type of object perspective. In recognition, items studied in the noncanonical viewpoint produced higher proportions of "old" responses than did items studied in the canonical viewpoint, whereas new objects presented from a noncanonical viewpoint produced fewer "old" responses than did new objects presented from the canonical viewpoint. In free recall, conversely, objects studied from the noncanonical viewpoint produced lower recall rates than did objects studied from the canonical viewpoint. These results, which reveal a pattern similar to word frequency effects, support the psychological reality of canonical viewpoints and the frequency-of-exposure-based accounts of canonical viewpoint effects.

The present article is concerned with a simple question: Does the viewpoint from which we see a 3-D object affect our memory for that object? The consideration of the role of viewpoint in object recognition has an involved history. It has been shown repeatedly that viewpoint affects identification performance. Palmer, Rosch, and Chase (1981) found the following viewpoint effects: (1) Viewpoint affected ratings of goodness of photographs of familiar objects; (2) viewpoint affected the formation and description of a mental image of a familiar object; and (3) viewpoint affected response times and error rates in naming objects. Moreover, the viewpoint that maximized object-identification performance was often considered by participants to be the best; it was also the easiest for participants to imagine and describe. Palmer et al. coined the term canonical to describe this seemingly best view. Additional evidence for the existence of canonical viewpoints was provided by Blanz, Tarr, and Bülthoff (1999), Jolicoeur (1985), and Verfaillie and Boutsen (1995).

The role of viewpoint in object recognition has been debated vigorously. Several theorists have cited viewpoint effects as evidence that viewpoint-specific representations underlie object recognition (Edelman & Bülthoff, 1992; Tarr & Bülthoff, 1995). Others have posited that objects are stored in a 3-D representation without specific, privileged viewpoints (Biederman, 1987). Canonical viewpoint effects do occur; hence, for the latter theories to hold, the effects must reflect decision variables rather than result from matches or mismatches to underlying representations (see, e.g., Rouder, Ratcliff, & McKoon, 2000).

One explanation of canonical viewpoints that is consistent with both viewpoint-specific and viewpoint-irrelevant theories is that they represent a frequency effect. Tarr and Pinker (1989), for example, suggested that canonical viewpoints for novel objects form only after repeated viewings at a specific orientation. Karnath, Ferber, and Bülthoff (2000) suggested that the neurons that underlie the recognition of an object become attuned to the features present in the most recognizable views of that object, and that this tuning evolves with experience.

The notion that frequency affects memory is not restricted to the effects on object recognition. Frequency of occurrence is one of the most studied variables in the memory for words. The effect of frequency of occurrence is surprisingly complex and varies with the memory task. In recognition memory paradigms, in which participants are asked to judge whether words at test were previously studied, there is an advantage for low-frequency words over high-frequency words (Jacoby & Dallas, 1981; Peters, 1936; Scarborough, Cortese, & Scarborough, 1977; Underwood & Freund, 1970). …

Search by... Author
Show... All Results Primary Sources Peer-reviewed

Oops!

An unknown error has occurred. Please click the button below to reload the page. If the problem persists, please try again in a little while.