Acceso usuarios
, Natalia Judith Laso (secret.) , Magali Paquot (voc.) The present dissertation is a corpus-based investigation of the frequency, structure and functions of lexical bundles in published scientific writing in English, whose main objective is the creation of an inventory of the most frequent and pedagogically useful lexical bundles in scientific prose, one that can be utilized in a variety of teaching applications. In this study, three- to six-word lexical bundles were extracted from a 1.3 million word sample from the Health Science Corpus, a collection of published articles in biology and biochemistry. This initial list was filtered and enhanced through the application of the Mutual Information (MI) statistic and of a set of exclusion criteria established to satisfy the pedagogical objectives of the study. Following the SciE-Lex investigation (Verdaguer et al., 2009) the remaining lexical bundles were grouped together using like keywords. The present study additionally used the concept of prototypical bundle, which is based on Sinclair’s (2004) notion of canonical units of meaning, to tackle the semantic and structural connections between similar bundles. The structural and functional characteristics of the lexical bundles were explored through careful concordance analysis, which made it possible to categorize the bundles using modified versions of Biber et al.’s (1999) structural framework and Hyland’s (2008a) functional taxonomy. These quantitative and qualitative analyses reveal how native expert writers employ recurrent word strings in the construction of a coherent, well-structured and convincing scientific text that conforms with the conventions of the genre. They bring to light the different functions that lexical bundles perform in scientific discourse, and how these functions enable writers to address their research concerns, achieve their communication goals and elicit the desired reaction from their target audience. They also show the typical structural realizations of these bundle functions, as well as important aspects of usage that non-native writers need to be aware of to be able to incorporate these expressions in their own writing. The study also compares the results obtained from the corpus of published scientific articles to the lexical bundles found in a smaller corpus of biomedical research articles written by native Spanish-speaking scientists, who are all non-native users of English. In accordance with the methodology proposed by Cortes (2004), the lexical bundles identified in the HSC were treated as target bundles and subsequently searched for and analyzed in the corpus of non-native writing. This comparison uncovered non-native writers’ overuse of certain bundles, a tendency that results in unnecessary repetitiveness and lack of variation, as well as their restricted use of participant- oriented bundles, which points to their limited awareness of the usage and importance of this particular function. The dissertation also discusses the pedagogical implications of its final product, a practical list of lexical bundles in scientific English for use in teaching applications, and how it addresses the six major challenges that hinder the successful introduction of lexical bundles in EAP classrooms and teaching materials, as identified by Byrd and Coxhead (2010).
