The synthesis of complex categories from perceptual and equivalence classes: effects of training and testing parameters

Abstract

Most real-world categories are complex since each contains some exemplars that resemble each other and others that are perceptually disparate. One example would be the many sounds of a persons’ voice when speaking and whispering, pictures of that person taken from different vantage points and in different degrees of focus, and the name of that person as written by different hands, and in different fonts. When all of these exemplars occasion the mutual selection of each other, and a function acquired by one class member generalizes to all other class members, all of the exemplars are acting as members of a complex category. A complex category can be formed by the merger of perceptual, fuzzy, and relational classes with equivalence classes. This presentation will consider how training and testing variables influence the likelihood of establishing a variety of complex categories such as linked perceptual classes, and fully elaborated generalized equivalence classes. These data provide a basis for understanding how complex categories are induced in natural settings.