Central tendencies, extreme points, and prototype enhancement effects in ill-defined perceptual categorization

Area of Science:

• Cognitive Psychology
• Perception
• Machine Learning

Background:

• Exemplar-based models of categorization often assume prototypes represent psychological central tendencies.
• Empirical findings of extreme prototype enhancement effects challenge this assumption.
• Understanding category representation is crucial for cognitive modeling.

Purpose of the Study:

• To investigate the psychological representation of category prototypes in perceptual classification.
• To test whether prototypes function as central tendencies or extreme points relative to contrast categories.
• To evaluate the implications for exemplar-based models of categorization.

Main Methods:

• Three perceptual classification experiments using ill-defined category structures.
• Participants learned categories from distorted dot patterns and made similarity judgments.
• Multidimensional scaling (MDS) analysis of similarity data.

Main Results:

• Prototypes were classified more accurately than other category instances (extreme prototype enhancement).
• MDS revealed prototypes as psychological extreme points, not central tendencies.
• This pattern held for novel symmetric patterns and grid patterns.

Conclusions:

• Category prototypes can be represented as psychological extreme points, challenging the central tendency assumption.
• Exemplar-based models, like the Generalized Context Model, can accommodate these findings when incorporating MDS solutions.
• The study refines our understanding of how humans represent and categorize information.