simplicity (n.)

A measure proposed by GENERATIVE LINGUISTIC theory which would automatically assign factors to competing linguistic analyses that would determine which of them was the most satisfactory; also called a simplicity metric. Simplicity is here defined quantitatively, in terms of the number of constructs (symbols, RULES, etc.) used in formulating an analysis; this is also often referred to as an ECONOMY measure. Perhaps the most widespread criterion is the number of FEATURES required in order to state a PHONOLOGICAL GENERALIZATION, and much thought has been given by generative phonologists to ways in which such generalizations can be more economically stated, using various kinds of notational abbreviation, e.g. ALPHA NOTATION. But the measure relies on a notion of simplicity which still requires much theoretical and methodological elucidation. It is proving extremely difficult to evaluate simultaneously the many variables entering into an analysis, especially the closer that analysis gets to the language as a whole. A simplification made in one part of the analysis may lead to unexpected consequences, in terms of great complexity (or COST) elsewhere. There is also the regular possibility that adult NATIVE-SPEAKERS, and children learning a language, will not always prefer the simpler of the two solutions; and little progress has been made in relating simplicity to other aspects of intuitive evaluation, such as NATURALNESS. Simplicity, then, and its FORMALIZATION, remain a controversial topic.