Concepts And Similarity

Throughout this chapter we have being making implicit use of the notion of similarity without saying much about what it is. In the treatment of prototype theory, we assumed that prototypes were formed by noting the similarity of instances to one another, by finding the attributes they have in common. In the treatment of exemplar theories we saw that instances were retrieved from memory on the basis of similarity. But, how exactly does similarity work? In this section, we outline an important model of similarity and relate recent research that questions this model.

Tversky's contrast model of similarity

One of the oldest and most successful models in cognitive psychology is Tversky's contrast model (Tversky, 1977). This model accounts for the similarity judgements made by people involving concepts described verbally or diagrammatically. Until recently, it was also the model implicitly or explicitly assumed by many concept theorists (see Smith, 1988). Since 1977, the contrast model has been developed and tested extensively by Tversky and his colleagues (Tversky, 1977; Tversky & Gati, 1978).

The model maintains that the similarity of two concepts is based on some function of the attributes shared by the concepts less the attributes that are distinctive to both:

where a and b are two concepts, s is the similarity of these two concepts, A is the set of attributes of concept a and B is the set of attributes of concept b. In this formula, A 0 B gives you the attributes that are common to the two objects, A—B gives you the attributes that are distinctive to a, and B—A the attributes that are distinctive to b (note that this is not an absolute distinctiveness, but just what is distinctive in one concept relative to the other). In general, this formula predicts that as the number of common features increases and the number of distinctive features decreases, the two objects a and b become more similar. The function f has a role in weighting certain attributes according to their salience and importance. The parameters 8, a, and p are used as multipliers to reflect the relative importance of the common and distinctive attribute-sets. For instance, when people judge the similarity of two objects they tend to weight the commonfeatures set as being more important than the distinctive-feature sets, whereas the distinctivefeature sets assume more importance in judgements of difference.

The effects of these 8, a, and p parameters also appear in the asymmetries that appear in similarity judgements, where it has been found that the similarity of a to b is not equal to the similarity of b to a; s(a, b)i$(b, a). Tversky points out that in similarity statements there is a subject and a referent, we say that "a (subject) is like b (referent)". Furthermore, the choice of the referent and the subject is in part determined by the most important or salient concept; the more prominent concept being the referent. We say that "North Korea is like China", when China is the more prominent concept in the pair. When we reverse the roles of the concept the similarity of the two concepts

Healing The Inner Child

Healing The Inner Child

Get All The Support And Guidance You Need To Be A Success At Changing Your Life. This Book Is One Of The Most Valuable Resources In The World When It Comes To What You Need To Know About Spiritual Emotional Freedom.

Get My Free Ebook


Post a comment