last changed in 2002

Concepts are 'ideas' - not a very well defined notion, but one which most cognitive psychologists are willing to entertain. (We shall define the notion 'idea' more precisely below so as to use it as a technical term.) It is a unit of thought (cognition) which has some kind of 'content', so that it corresponds to some category of experiences; Happiness, Dog and Three are all concepts, and in language the same is true of the word happiness and the categories Noun and Subject. In WG, a concept is simply a node in a (symbolic) network whose 'content' is defined by its links to other concepts; in contrast with some other theories the concept has no internal structure - it is not a 'box' or 'frame' of information. This means that all concepts are equally simple, even though some concepts are much more richly defined than others by links to other concepts.

Some concepts are stored permanently - in other words, they are held in memory - whereas others are created temporarily in order to represent bits of current experience. The aim of processing is to match each temporary concept with some permanent one, so that the temporary one isa the permanent one - i.e. the temporary one is classified as an example of the permanent one. Classification is the first step towards understanding. The match between the two need not be perfect, but should be the best available, as defined by the Best Fit Principle. Mismatches are compatible with the isa relationship because of default inheritance, which allows any characteristics of the 'model' concept to be overridden by those of the instance. For example, grean (peas) can be interpreted as an example of GREEN in spite of the misspelling, just as a three-legged dog can be 'interpreted' as a dog. This means that some examples are 'better' than others, and some examples are harder than others to classify. These are two so-called 'prototype effects' which are taken as evidence that categories (general concepts) are prototypes rather than 'classical categories' - i.e. categories with clear boundaries and definitions.

Given the WG view of concepts, it is reasonable to assume that most of them are learned rather than innate.

One particularly interesting fact about concepts is that some of them are concepts of concepts - 'second-order concepts'. For example, I can distinguish between London and my concept of London:

(1) London is big.

(2) I've recently revised my concept of London.

(Of course even London is just a concept, but in my mind it stands for the real thing, so I attach to it all the characteristics that I think the real thing has; I don't think of it as a concept, unlike 'my concept of London'.)

It is useful to have a general term for concepts of concepts, so we shall call them ideas. In the grand ontology of (my) knowledge, London isa Place, but Concept-of-London isa Idea. One particularly important kind of idea is a proposition.







This is an article from the Encyclopedia of Word Grammar and English grammar. If you refer to it, please give the url as "".