Toward a Lexicon-based Lexicography

Toward a Lexicon-based Lexicography Rufus H. Gouws 1. Introduction To give a valid account of the complete lexical stock of a language , the selection of lexical items included in a dictionary should represent the typological variety of that language. Because words constitute the largest component of the lexicon, the tradition has been established to treat a large collection of headwords. This approach has led to word-based dictionaries, i.e., dictionaries displaying a word-bias by narrowly construing lemmas as words, even though words are not the only elements in the lexicon. Multiword units and morphemes also function as lexical items. To ensure a sound treatment ofall these lexical items the traditional word-based lexicography should be replaced by a broader lexicon-based approach that offers a more comprehensive reflection of the lexicon by listing and treating multilexical and sublexical lemmas. All the lexical items of a language are regarded as potential lemmas; however, not all words qualify as lexical items. Words that occur only as components of multiword lexical items should not themselves be regarded as lexical items. Before selecting the lemmas for a dictionary the lexicographer must specify the linguistic criteria according to which lexical items are identified. These criteria will determine the choice of lemmas. This paper focuses on the way in which sublexical and multilexical lemmas may be managed in lexicon-based dictionaries. Emphasis will be on the treatment of multiword units and their lemmatization, but a brief discussion will also be given of the types of items to be included as sublexical lemmas and the formal presentation of them. 76Rufus Gouws Key Terms»¦Lexical item: any element ofthe lexicon—(1) word, e.g., table, spoon; (2) sublexical lexical item, e.g., dis-; (3) multiword lexical item, e.g., home ground.»¦Lemma: lexical item treated as head of an entry: words as lexical lemmas; sublexical lexical items as sublexical lemmas, elements larger than words as multiword lemmas. Lexical lemmas are words included as lemmas. wSublexical lemmas: affix, e.g., dis- in discontinue, -ing in watering ; stem, e.g., water in watering, mat in doormat; combining form, e.g., -naut in astronaut. wMultilexical lemmas are single lexical lemmas consisting of more than one word: (1) multiword units; (2) loanwords, e.g., bonafide; idioms, e.g., die appeltjie val nie ver van die boom nie 'like father, like son'; (3) discontinuous items, e.g.,ge-...-te in gedermte 'entrails'. wSublexical items: words or stems that form a part of a lexical item but cannot function as lexical items, e.g., plaas 'place' in the multiword lexical item m plaas van 'instead of; hetsy...hetsy 'either ... or' (hetsy may not occur independently). 2. Sublexical lemmas The lexicon contains a substantial number of items smaller than words—including affixes, combining forms, and stems. Many of these items are productive in word-forming processes and constitute a part of the active lexicon of the standard language. Dictionaries have to give an account of these items but their limited occurrence as components of complexes and compounds must be recorded. The word-based approach impedes the unprejudiced representation of these other lexical items. This disadvantage applies to multiword lexical items in a more comprehensive way than to elements smaller than words because there is a more direct resemblance in form between a word and a subword item than between a word and a multiword item. Lexicon-based Lexicography77 Consequently dictionaries generally display a more satisfactory treatment of sublexical lemmas than of multilexical lemmas. 2.1Affixes and combining forms In English dictionaries the lemmatization of lexical items smaller than words primarily focuses on affixes and combining forms. Affixes constitute an important component of the lexicon and the user of word-based dictionaries is accustomed to a sound treatment of this category of lexical items. Affixes as sublexical lemmas, therefore, will not be discussed further in this paper. Combining forms have a frequent and productive occurrence in English. According to Bauer (1983, 213) these elements usually function as affixes but appear to be distinct from affixes in some cases. Combining forms deriving from Latin and Greek (e.g., electro-, hydro-, -naut, -phile) are restricted to morphologically complex structures . They cannot function as independent words...