Panel 102

: CONCEPT HIERARCHIES

• People use hierarchies to represent relationships of class inclusion between categories; that is, to include one category within another (e.g., the category of chair within the category for furniture).

• Human conceptual hierarchies have three levels; a superordinate level (e.g., weapons, furniture), a basic level (e.g., guns, chair), and a subordinate level of specific concepts (e.g., hand-guns, rifles, kitchen chairs, armchairs).

• The basic level is the level at which concepts have the most "distinctive attributes" and it is the most cognitively economic; it is the level at which a concept's attributes are not shared with other concepts at that level.

• Categories at the basic level are critical to many cognitive activities; for example, they contain concepts that can be interacted with using similar motor movements, they have the same general shape, and they may be associated with a mental image that represents the whole category.

• The position of the basic level can change as a function of individual differences in expertise and cultural differences.

(like brown), to distinguish it from other subordinate concepts (e.g., canary, defined as feathered, animate, two-legged, small, yellow). This means that a specific concept will tend to have more attributes in common with its immediate superordinate than with a more distant superordinate. For example, sparrow should have more attributes in common with its immediate superordinate, bird, than with its more distant superordinate animal. Several computational models of this type of theory have been proposed (Collins & Quillian, 1969, 1970; Quillian, 1966; see also Chapter 1). The details of the Collins and Quillian model are shown in Panel 10.2 (see also Figure 10.2).

Evidence for the defining-attribute view

Several early studies seemed to support this theory and its particular instantiation as a semantic network model. The early work by Bruner et al. (1956) using artificial categories clearly assumes this theory. Similarly, Collins and Quillian used sentence-verification tasks to find support for their model of the theory. In these tasks, subjects were asked to say whether simple sentences of two forms were true or false. First, they were asked whether "an INSTANCE was a member of a SUPERORDINATE" (e.g., "Is a canary an animal?" or "Is a canary a fish?"). Second, subjects were asked whether "an INSTANCE had a certain ATTRIBUTE" (e.g., "Can a canary fly?", "Does a canary have skin?"). In both of these cases Collins and Quillian's predictions were confirmed. In the INSTANCE-SUPERORDINATE sentence it was found that the greater the distance between the subject and predicate of the sentence in the hierarchy, the longer it took to verify the sentence. And in the INSTANCE-ATTRIBUTE case the place of the attribute in the hierarchy relative to the instance, predicted the time taken to verify the sentence. However, some of their other predictions were not supported: Reaction times to questions that were false (e.g., a canary is a stone) were very fast even when many links should have been traversed to answer the question.

Canary (

Canary (

is pink ii edible swims upstream to lay egg*

is pink ii edible swims upstream to lay egg*

Business Correspondence

Business Correspondence

24 chapters on preparing to write the letter and finding the proper viewpoint how to open the letter, present the proposition convincingly, make an effective close how to acquire a forceful style and inject originality how to adapt selling appeal to different prospects and get orders by letter proved principles and practical schemes illustrated by extracts from 217 actual letter.

Get My Free Ebook


Post a comment