The Evolution
3
The Evolution of Semantic Memory and Spreading Activation
The purpose of this paper is to demonstrate that it is possible
to deduce the structure of memory by mathematically analyzing the
environment in which it evolved.
The theory arrived at employs three very restricted forms of
intersection search, which in turn are used to explain various facili-
tative effects (e.g. Brown & Block., 1980; Meyer & Schvaneveldt, 1971;
Tulving & Gold, 1963), some production experiments (Freedman & Loftus,
1971; Loftus 1973; Loftus & Loftus, 1974), the nature of polsemy
(Anderson & Ortony, 1975), riddles, and the effects of word order on
meaning and paired-associate learning. The theory is similar in outlook
to the spreading-activation and intersection theories of Collins and
Loftus (1975), and Quillian (1967, 1968, 1969), but it contrasts with
the above in that the intersections employed are of three distinct
types (deriving from the use of just two types of links), none of which
are ever-widening searches, and all of which operate under restrictions
imposed by evolutionary considerations.
For an interesting review of the subject in general from the
point of view of two authors who argue persuasively that there is a
need to integrate the information-processing approach with restrictions
derived from evolutionary principles, and who present some valuable
preliminary ideas as to what these principles should be, consult
Lachman and Lachman (1979).