tv.um.es

Corpus linguistics has exerted a profound effect upon lexicography for the last decades, since the art of making dictionaries has benefited from the methods and tools provided by the former. For instance, the use of corpora for compiling dictionaries allows to know which words are more common in a language and how many meanings they can carry. In addition, corpus texts can be checked to discover the syntagmatic relations between words, that is, which words usually accompany other words. Among these syntagmatic relations, several types can be distinguished, e.g. idioms, phrasal verbs, collocations. The present study primarily focuses on the last type. This phenomenon can be defined as “a relation of affinity which holds between words in a language, and which is revealed by the typical co-occurrence of words” (Seretan, 2011, p. 10). Collocations belong to the phraseology of words, and phraseological items are said to take precedence over single words (Granger & Paquot, 2008, p. 29; Paquot, 2015, p. 461) for the simple reason that, sometimes, the meaning of a given word can only be unveiled by considering the words around it, such as delexical verbs. Apart from that, a good command of collocation is an important indicator of native-like proficiency and fluency (Bartsch, 2004, p. 20), so it is hardly surprising that dictionaries targeted at second language learners incorporate collocational relationships into their entries. However, according to Xiao (2015), “collocation information in dictionaries, if any, has traditionally been provided in the form of illustrative examples” (p. 115). In other words, dictionaries have not clearly specified which collocations words are part of, but these word combinations need to be inferred from the example sentences taken from a corpus.
Therefore, the aim of this paper is to test this hypothesis and examine how collocations are represented in the entries for the noun bottle in the ‘Big Five’, that is, “the five major electronic monolingual learners’ dictionaries of English” (Paquot, 2015, p. 470), namely, Collins English Dictionary (CED), Longman Dictionary of Contemporary English (LDOCE), Oxford Advanced Learner’s Dictionary (OALD), Cambridge Dictionary of English (CDE) and Macmillan English Dictionary (MED). What these dictionaries have in common is that they are all corpus-based, which means that lexicographers have used corpora in order to get precise information about words. Nevertheless, since they are based on different corpora, it is expected that the senses of the word under study and its syntagmatic attractions differ from one dictionary to another. In fact, preliminary results reveal that the OALD, the CDE and the MED adopt a more systematic approach to collocational representation, since they contain more explicit references to collocations than the CED and the LDOCE, where they need to be inferred from the corpus sentences. Additionally, our study demonstrates that not only are corpora used to collect data on the distinct senses of words in specific contexts as well as their phraseology, but they can also provide information about the frequency of use of the words in a language.