Фрагменты когнитивной психологии. Бабушкин А.П. - 15 стр.

                                             15
the real world and that concepts are mental description of them. We believe that this
distinction is misleading because concepts need not have real-world counterparts,
because the set of potential real-world categories is indefinitely large, and because
people may impose rather than discover structure in the world.
        The most active ground for research and theorizing about concepts is involved
with determining the structure of concepts. There are several theories concerning
conceptual structure:
The classical view. The classical view has had a profound historical impact on
psychology and linguistics. On this view, concepts are structured around defining
features-features that are singly necessary and jointly sufficient to define the concept.
The category “even number” has a classical definition provided by the feature “evenly
divisible by two”; this definition acts as a fixed criterion for evaluating whether an
object belongs to the concept. According to the pure form of the classical view, all
concepts are like “even number” in being organized around defining features.
The prototype view. The prototype view developed out of dissatisfaction with the
classical view. Its central claim is that concepts are organized around a prototype or best
example that summarizes the most common or typical features among a concept’s
instances. The similarity between an item and a concept’s prototype determines
whether, and how well, the item belongs in the concept.
The exempler view. Unlike the earlier views, the exemplar view does not create a
summary description to represent concepts. Concepts are represented by the individual
instances (or examplars) that constitute it. To decide whether an item belongs to a
concept, the item is compared with the concept’s exemplars. If the item is similar to the
exemplars then it will be placed in the concept.
        Concepts serve multiple functions. They are:
(1) Classification. The classification function involves determining that a particular
     instance is an instance of a concept or that one particular concept is a subset of
     another (for example, robbins are birds).
(2) Understanding and explanation. Classification allows intelligent systems to form
     their experience into meaningful chunks and to construct an interpretation of it. Old
     knowledge is used to analyze the current situation. If a person recognizes an animal
     as a “rattle-snake”, for example, he or she will interpret their situation as dangerous.
(3) Prediction. In the example of the rattlesnake just described one may also access
     knowledge that might allow one to make predictions concerning the future. One
     might, for instance, predict that sudden movements might cause the rattlesnake to
     strike and on the basis of this prediction one might slowly back away from the
     snake to avoid being bitten.
(4) Reasoning. Concepts and conceptual structure support reasoning. One does not need
     to store every fact and possibility if one can derive knowledge from the information
     that is stored. From the knowledge that all animals breather, that birds are animals,
     and that flamingos are birds, one may reason that flamingos breathe. Concepts
     support not only logical inferences but also plausible reasoning. From “doctor” one
     might infer “large salary, long and irregular work hours”.