Lexicography, Terminology, and Translation: Text-based Studies in Honour of Ingrid Meyer

Chapter 2. Relevance in Dictionary Making: Sense Indicators in the Bilingual Entry1

25 INTRODUCTION Lexicographic relevance became a hot topic with the development of the text corpus: this is now, for many lexicographers, the principal source of information about their headwords. Gone are the days when you looked at a blank sheet of paper, consulted a meagre card index and a plethora of other dictionaries, and wished for more information. For the common words of the English language, the British National Corpus2 (from which our examples are taken) offers far more material than anyone can hope to handle in the context of a commercial dictionary—or, indeed, in a lifetime. Faced with screen aer screen of key word in context (KWIC) concordances, as illustrated in Figure 1, the lexicographer really is hard put, within the time constraints, to make sense of it all. Some publishing houses offer their dictionary writers a certain amount of formal training; the majority still rely on the new lexicographer learning by trial and error at the wordface. Even aer training, discovering and recording the essential facts about a word still depend more on the skill of the individual than on a systematic approach to the data. If the evidence in a corpus is to be exploited to its full potential, not merely by one individual dictionary editor but consistently by the whole editorial team, then a theoretically grounded approach to lexicographic relevance is an essential item in the lexicographer’s toolkit. It is convenient to consider the lexicographic process in two distinct phases (seeAtkins 1993 for a fuller account). In the initial stage (analysis) CHAPTER 2 RELEVANCE IN DICTIONARY MAKING: SENSE INDICATORS IN THE BILINGUAL ENTRY1 BERYL T. SUE ATKINS AND PIERRETTE BOUILLON 26 Beryl T. Sue Atkins and Pierre
e Bouillon of dictionary compiling, when lexicographers are studying the way that the word behaves in the language, they look at the evidence (corpus data, their own notes, etc.), record facts about the headword as they find them (meanings, constructions, collocates, participation in multiword expressions, register, language variety, style, etc.), establish provisional sense distinctions, a
empt to order the facts and the exemplifying sentences according to these distinctions, and thus create a rich database entry from which may be extracted the material needed for the particular dictionary on which they are working. The greatest danger at this stage is that, if there is no theoretical basis for the analysis, the collection of facts will be patchy and inconsistent, without any means of ensuring that no important aspect of the word’s behaviour has been overlooked. When editors come to the task of formulating the actual dictionary entry (synthesis), in the absence of any theoretical underpinning there Figure 1: KWIC concordances for argue 1. The teachers and medics were arguing about who has what of my time. 2. This is a key factor arguing against the existence of such a relationship. 3. “You’ll stop arguing and do as you’re damned well told!” 4. We spent most of our time in cafés arguing and holding hands. 5. These features argue for a local origin. 6. Margaret Mead argues for a nurture perspective on behaviour. 7. There was a lot of arguing going on between Mom and Dad. 8. Dr. Wilson argues that, if ants disappeared, most of. . . . 9. Richard Dawkins has argued that it is their genes that survive. 10.This situation argues that a serious tax should be levied. 11.The popular press have argued the case. 12.The platoon commander was arguing with a gang of Christian Phalangists. Relevance in Dictionary Making 27 is no means of ensuring that their approach to these tasks is consistent from A to Z of the dictionary—a process that may cover a number of years and involve a large team of lexicographers. For every entry and subentry,themajordecisioniswhattoputinor,moretantalizingly,what to leave out; obviously, native-speaker intuition informs the selection, but here again objective facts (which normally amount to no more than frequency statistics in the current corpus) are thin on the ground. It is essential that the selection be based on a clear overview of how the word actually behaves, a good counterweight to the salient usages available to native-speaker intuition (see Hanks 2000 for a discussion of social versus cognitive salience). So much for monolingual dictionaries. In the case of bilingual dictionaries, of course, the synthesis stage also includes the “transfer” process, when the source-language items are...