Componential analysis

Abstract: It is reasonable to make a distinction between 'lexicology' and 'lexicography', the former dealing with theoretical issues in the analysis of the lexicon, and the latter dealing with the production of the descriptive instrument commonly known as the 'dictionary'. Lexicology has been of less help to the practical lexicographer than might be expected, for many misleading notions cling to the word meaning and make it unsuitable for precise use as a technical term. Especially misleading is the notion that a word is an empty form into which a segment of meaning is poured. The popular view is that a word is a receptacle for a 'thought'. Most linguists avoid such a naive outlook and accept 'meaning' as a suffusing characteristic of all language phenomena. By definition, language is differentiated from noise by its symbolizing function. For the lexicographer, the crucial question about meaning is how to slice it up into manageable units, usually called SENSES. But how are the senses arrived at? In the dichotomy between the 'God's truth' approach and the 'hocus pocus' approach, nowhere in linguistics is 'hocus pocus' more evident. Fortunately, a pragmatic technique comes to the lexicographer's rescue. In accord with the OED tradition, he collects large bodies of documented quotations, studies them for similarities and differences, and puts them into various piles. For a large-scale dictionary, the differences can be emphasized, but for a smaller scale the similarities will cause the number of senses to be reduced. As Martin Joos has said: \"The proper stopping-point in this procedure seems to be a matter of taste, and we congratulate the worker when he stops just where we would have stopped.\" The discriminations are arbitrary, particularly