component (n.)

A term used in GENERATIVE LINGUISTICS to refer to the main sections into which a generative GRAMMAR is organized. In Noam Chomsky’s Syntactic Structures (1957), three components are recognized: the PHRASE-STRUCTURE component (which generates a set of UNDERLYING STRINGS), the TRANSFORMATIONAL component (which acts on these strings in various OPTIONAL and OBLIGATORY ways, introducing SEMANTIC changes), and the MORPHOPHONEMIC component (which converts each syntactic string into a string of PHONOLOGICAL UNITS). In Chomsky’s Aspects of the Theory of Syntax (1965), the model is radically altered. The phrase-structure component is replaced by a base component, which generates the underlying PHRASE-MARKERS representing the DEEP STRUCTURE of SENTENCES, i.e. all semantically relevant grammatical notions. The base component contains the CATEGORIAL and LEXICAL components (or sub-components) of the grammar. Two things then happen to these markers: (a) they are semantically interpreted, using the rules of the semantic component (which has no equivalent in the Syntactic Structures model), and (b) they are converted into SURFACE structures through the transformational component (which contains largely obligatory RULES, the optional ones now being handled by choices made in the base rules). Lastly, a phonological component operates on the surface structures, providing them with a PHONETIC interpretation.

In SEMANTICS, the term refers to an irreducible FEATURE in terms of which the SENSE of LEXICAL ITEMS can be analyzed, e.g. girl can be analyzed into the components ‘human’, ‘female’, ‘child’, etc. Componential analysis is a semantic theory which developed from a technique for the analysis of kinship vocabulary devised by American anthropologists in the 1950s. It claims that all lexical items can be analyzed using a finite set of components (or ‘semantic features’), which may, it is felt, be UNIVERSAL. Certainly, several sets of lexical items exist to show the strengths of the approach (e.g. the correspondences between boy/girl, man/ woman, ram/ewe, etc., can be stated in terms of [+male] v. [−male] or [−female] v. [+female]. There are several limitations to the componential models of analysis so far suggested, such as the extent to which BINARY analyses are possible for many lexical items, the claimed universality of components, and the justification for selecting one value rather than the other for a possible component (e.g. whether the above example should be analyzed in terms of [+male] or [−female]).

‘Componential analysis’ is also found in a general sense in linguistics, especially in Europe, referring to any approach which analyses linguistic units into components, whether in PHONOLOGY, grammar or semantics. In this view, PRAGUE SCHOOL phonological analysis is componential, as are the analyses of WORD-AND-PARADIGM MORPHOLOGY.

In some approaches to PHONOLOGY (e.g. DEPENDENCY PHONOLOGY), component is used for a FEATURE represented as a single (‘unary’) element, rather than as a BINARY opposition. The term is given special status in unary component theory.