Collocations in context

Abstract

The idea that text in a particular field of discourse is organized into lexical patterns, which can be visualized as networks of words that collocate with each other, was originally proposed by Phillips (1983). This idea has important theoretical implications for our understanding of the relationship between the lexis and the text and (ultimately) between the text and the discourse community/ the mind of the speaker. Although the approaches to date have offered different possibilities for constructing collocation networks, we argue that they have not yet successfully operationalized some of the desired features of such networks. In this study, we revisit the concept of collocation networks and introduce GraphColl, a new tool developed by the authors that builds collocation networks from user-defined corpora. In a case study using data from McEnery's (2006a) study of the Society for the Reformation of Manners Corpus (SRMC), we demonstrate that collocation networks provide important insights into meaning relationships in language.

The Ecuadorian Physiotherapist in National Printed Media: A Pilot Study of the Image of a Health Practitioner

“…For the purposes of analysis, all function words and personal names were removed from the list of collocates. In terms of the collocation parameter notation proposed by Brezina, McEnery, and Wattam (2015), these settings might be written as '3a+6a-MI(3)+LL(p<0.05), L5-R5, C1-NC3; function words and personal names removed'. The collocates were then categorised according to the most relevant WordNet domains (Bentivogli, Forner, Magnini, & Pianta, 2004), with the Spanish WordNet accessed through the multilingual lexical online database MultiWordNet 1.4.2 (Cendejas, Barceló, Sidorov, Gelbukh, & Chanona-Hernandez, 2012).…”

“…Recently, several text analytic tools has been developed providing convenient ways to simplify and reduce textual data dimensionality with the use of collocation networks (Brezina et al, 2015), cluster analysis, and multidimensional scaling (Cleophas & Zwinderman, 2015, pp. 3-46).…”

A Pattern for Concept Identification from English Translated Quran

“…The multi terms also can be identified by identifying terms collocation. The collocation will determine a sequence of terms that frequently appear more together than would be expected by chance collocation [20]. For example the noun Fire can be a collocation term when it occurs frequently with the term raging.…”

“…It is considered as a phraseological perspective and from a frequency-based perspective [23]. Mostly, three criteria is used to identify collocations which is the distance, frequency and exclusivity [20]. The frequency identification is a simple method to determine the collocation of terms.…”

Biting off More Than They Can Chew? The Impact of Pedagogical Application of Corpus on Vocabulary Ability of Intermediate-Level ESL Learners in Mainland China: A Quasi-Experimental Study

“…The new-GSL is a robust comparison of four language corpora (LOB, BNC, BE06, and EnTenTen12) of a total size of over 12 billion running words, compiled by Brezina and Gablasova (Brezina et al, 2015). The teaching materials used in the experimental group are selected and edited by manipulating the new-GSL.…”