The English articles (THE, A, AN) are normally described in terms of the grammar of the language. This is only natural, since they are extremely frequent, fit into certain well-defined syntactic slots, and usually help to communicate only very broad aspects of textual meaning. However, as John Sinclair has pointed out (1999, pp.160-161), the articles are also found as components of many lexico-phraseological units, and in such cases a normal grammatical description may not be of relevance. An example he gives is the presence of A in the phrase 'come to a head', where ‘A has little more status than that of a letter of the alphabet’ (p.161). Sinclair also makes the observation that, ‘I do not know of an estimate of the proportion of instances of A, for example, that are not a realisation of the choice of article but of the realisation of part of a multi-word expression.’ (p.161). The present paper addresses the questions raised by Sinclair, and does so with reference to both the definite and the indefinite article. It focuses, in particular, on the spoken language, and presents the results of analyses of random samples of the articles in the spoken component of the British National Corpus (hereafter BNC-spkn). According to the data in Leech et al (2001, p.144), THE is the most frequent word in BNC-spkn and A is the sixth most frequent (a rank position which remains unaltered when the frequencies of A and AN are combined). Using the BNCweb interface, and specifying that the relevant word forms should be ‘articles’, the total numbers of tokens are: an 19,049; a 200,004; the 409,060. Since the numbers are very high, the samples investigated also contained a reasonably large number of tokens (500). The relative samples corresponded to the following proportions of tokens in BNC-spkn: an 2.62%, a 0.25%, the 0.12%. The latter two are very low percentages, and for this reason, three separate samples of each were investigated, in order to see the extent to which the samples differed. Analysis of article usage was carried out in the first instance by reading right-sorted concordance lines. Whenever doubts arose, larger contexts were retrieved from the corpus. Various reference works were also consulted, including Berry (1993), Francis et al (1998), and various corpus-based dictionaries and grammars. The data presented includes: description of the various types of lexico-phraseological unit found; the proportions of the samples judged to involve the different lexico-phraseological phenomena identified; the problems encountered when deciding whether or not phraseology is an important factor in specific instances of article usage; and the number of tokens in each sample which were in some way irrelevant, for example because they involved speaker repetition of the article, or the non-completion of a noun phrase.
The lexico-phraseology of THE and A/AN in spoken English: a corpus-based study
COFFEY, STEPHEN JAMES
2015-01-01
Abstract
The English articles (THE, A, AN) are normally described in terms of the grammar of the language. This is only natural, since they are extremely frequent, fit into certain well-defined syntactic slots, and usually help to communicate only very broad aspects of textual meaning. However, as John Sinclair has pointed out (1999, pp.160-161), the articles are also found as components of many lexico-phraseological units, and in such cases a normal grammatical description may not be of relevance. An example he gives is the presence of A in the phrase 'come to a head', where ‘A has little more status than that of a letter of the alphabet’ (p.161). Sinclair also makes the observation that, ‘I do not know of an estimate of the proportion of instances of A, for example, that are not a realisation of the choice of article but of the realisation of part of a multi-word expression.’ (p.161). The present paper addresses the questions raised by Sinclair, and does so with reference to both the definite and the indefinite article. It focuses, in particular, on the spoken language, and presents the results of analyses of random samples of the articles in the spoken component of the British National Corpus (hereafter BNC-spkn). According to the data in Leech et al (2001, p.144), THE is the most frequent word in BNC-spkn and A is the sixth most frequent (a rank position which remains unaltered when the frequencies of A and AN are combined). Using the BNCweb interface, and specifying that the relevant word forms should be ‘articles’, the total numbers of tokens are: an 19,049; a 200,004; the 409,060. Since the numbers are very high, the samples investigated also contained a reasonably large number of tokens (500). The relative samples corresponded to the following proportions of tokens in BNC-spkn: an 2.62%, a 0.25%, the 0.12%. The latter two are very low percentages, and for this reason, three separate samples of each were investigated, in order to see the extent to which the samples differed. Analysis of article usage was carried out in the first instance by reading right-sorted concordance lines. Whenever doubts arose, larger contexts were retrieved from the corpus. Various reference works were also consulted, including Berry (1993), Francis et al (1998), and various corpus-based dictionaries and grammars. The data presented includes: description of the various types of lexico-phraseological unit found; the proportions of the samples judged to involve the different lexico-phraseological phenomena identified; the problems encountered when deciding whether or not phraseology is an important factor in specific instances of article usage; and the number of tokens in each sample which were in some way irrelevant, for example because they involved speaker repetition of the article, or the non-completion of a noun phrase.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
