Academic Vocabulary in Learner Writing

Corpus and Discourse Series editors: Wolfgang Teubert, University of Birmingham, and Michaela Mahlberg, University of Liverpool. ˇ Editorial Board: Paul Baker (Lancaster), Frantisek C ermák (Prague), Susan Conrad (Portland), Geoffrey Leech (Lancaster), Dominique Maingueneau (Paris XII), Christian Mair (Freiburg), Alan Partington (Bologna), Elena TogniniBonelli (Siena and TWC), Ruth Wodak(Lancaster), Feng Zhiwei (Beijing). Corpus linguistics provides the methodology to extract meaning from texts. Taking as its starting point the fact that language is not a mirror of reality but lets us share what we know, believe and think about reality, it focuses on language as a social phenomenon, and makes visible the attitudes and beliefs expressed by the members of a discourse community. Consisting of both spoken and written language, discourse always has historical, social, functional, and regional dimensions. Discourse can be monolingual or multilingual, interconnected by translations. Discourse is where language and social studies meet. The Corpus and Discourse series consists of two strands. The first, Research in Corpus and Discourse, features innovative contributions to various aspects of corpus linguistics and a wide range of applications, from language technology via the teaching of a second language to a history of mentalities. The second strand, Studies in Corpus and Discourse, is comprised of key texts bridging the gap between social studies and linguistics. Although equally academically rigorous, this strand will be aimed at a wider audience of academics and postgraduate students working in both disciplines. Research in Corpus and Discourse Conversation in Context A Corpus-driven Approach With a preface by Michael McCarthy Christoph Rühlemann Corpus-Based Approaches to English Language Teaching Edited by Mari Carmen Campoy, Begona Bellés-Fortuno and Ma Lluïsa Gea-Valor Corpus Linguistics and World Englishes An Analysis of Xhosa English Vivian de Klerk Evaluation and Stance in War News A Linguistic Analysis of American, British and Italian television news reporting of the 2003 Iraqi war Edited by Louann Haarman and Linda Lombardo

Evaluation in Media Discourse Analysis of a Newspaper Corpus Monika Bednarek Historical Corpus Stylistics Media, Technology and Change Patrick Studer Idioms and Collocations Corpus-based Linguistic and Lexicographic Studies Edited by Christiane Fellbaum Meaningful Texts The Extraction of Semantic Information from Monolingual and Multilingual Corpora Edited by Geoff Barnbrook, Pernilla Danielsson and Michaela Mahlberg Rethinking Idiomaticity A Usage-based Approach Stefanie Wulff Working with Spanish Corpora Edited by Giovanni Parodi Studies in Corpus and Discourse Corpus Linguistics and The Study of Literature Stylistics In Jane Austen’s Novels Bettina Starcke English Collocation Studies The OSTI Report John Sinclair, Susan Jones and Robert Daley Edited by Ramesh Krishnamurthy With an introduction by Wolfgang Teubert Text, Discourse, and Corpora. Theory and Analysis Michael Hoey, Michaela Mahlberg, Michael Stubbs and Wolfgang Teubert With an introduction by John Sinclair

This page intentionally left blank

Academic Vocabulary in Learner Writing From Extraction to Analysis Magali Paquot .

Continuum International Publishing Group The Tower Building 80 Maiden Lane 11 York Road Suite 704. including photocopying. Typeset by Newgen Imaging Systems Pvt Ltd. recording. without prior permission in writing from the publishers. or any information storage or retrieval system. No part of this publication may be reproduced or transmitted in any form or by any means. New York London SE1 7NX NY 10038 www. India Printed and bound in Great Britain by the MPG Books Group . Chennai. British Library Cataloguing-in-Publication Data A catalogue record for this book is available from the British Library. ISBN: 978-1-4411-3036-5 (hardcover) Library of Congress Cataloging-in-Publication Data A catalog record for this book is available from the Library of Congress. electronic or © Magali Paquot 2010 All rights reserved.


Acknowledgements List of abbreviations List of figures List of tables Introduction Part I: Academic vocabulary Chapter 1 What is academic vocabulary? 1.1. Academic vocabulary vs. core vocabulary and technical terms 1.1.1. Core vocabulary 1.1.2. Academic vocabulary 1.1.3. Technical terms 1.1.4. Fuzzy vocabulary categories 1.2. Academic vocabulary and sub-technical vocabulary 1.3. Vocabulary and the organization of academic texts 1.4. Is there an ‘academic vocabulary’? 1.5. Summary and conclusion Chapter 2 A data-driven approach to the selection of academic vocabulary 2.1. Corpora of academic writing 2.2. Corpus annotation 2.2.1. Issues in annotating corpora 2.2.2. The software 2.3. Automatic extraction of potential academic words 2.3.1. Keyness 2.3.2. Range 2.3.3. Evenness of distribution 2.3.4. Broadening the scope of well-represented semantic categories 2.4. The Academic Keyword List 2.5. Summary and conclusion

xi xiii xv xvii 1

9 9 10 11 13 13 17 22 25 27

29 31 34 34 36 44 46 48 50 53 55 61


Contents Part II: Learners’ use of academic vocabulary

Chapter 3 Investigating learner language 3.1. The International Corpus of Learner English 3.2. Contrastive Interlanguage Analysis 3.3. A comparison of learner vs. expert writing 3.4. Summary and conclusion Chapter 4 Rhetorical functions in expert academic writing 4.1. The Academic Keyword List and rhetorical functions 4.2. The function of exemplication 4.2.1. Using prepositions, adverbs and adverbial phrases to exemplify 4.2.2. Using nouns and verbs to exemplify 4.2.3. Discussion 4.3. The phraseology of rhetorical functions in expert academic writing 4.4. Summary and conclusion Chapter 5 Academic vocabulary in the International Corpus of Learner English 5.1. A bird’s-eye view of exemplification in learner writing 5.2. Academic vocabulary and general interlanguage features 5.2.1. Limited lexical repertoire 5.2.2. Lack of register awareness 5.2.3. The phraseology of academic vocabulary in learner writing 5.2.4. Semantic misuse 5.2.5. Chains of connective devices 5.2.6. Sentence position 5.3. Transfer-related effects on French learners’ use of academic vocabulary 5.4. Summary and conclusion Part III: Pedagogical implications and conclusions Chapter 6 Pedagogical implications 6.1. Teaching-induced factors 6.2. The role of the first language in EFL learning and teaching 6.3. The role of learner corpora in EAP materials design

67 67 70 72 78 81 81 88 90 95 106 108 122

125 125 142 142 150 154 168 174 177 181 192

201 201 203 206

Contents Chapter 7 General Conclusion 7.1. Academic vocabulary: a chimera? 7.2. Learner corpora, interlanguage and second language acquisition 7.3. Avenues for future research Appendix 1: Expressing cause and effect Appendix 2: Comparing and contrasting Notes References Author index Subject index


211 211 215 216 219 226 235 240 257 261

This page intentionally left blank


There are several people without whom this book would never have been written. First and foremost, I want to express my deepest and most sincere gratitude to my PhD supervisor, Professor Sylviane Granger, for her infectious enthusiasm, her intellectual perceptiveness and her unfailing expert guidance. I am greatly indebted to you, Sylviane, for giving me the opportunity to join the renowned Centre for English Corpus Linguistics seven years ago now! I have been lucky enough to undertake research in an environment where writing a PhD also means collaborating with many fellow researchers on up-and-coming projects, attending thoughtprovoking conferences, organizing seminars, conferences and summer schools, as well as lecturing and offering guidance to undergraduate students. I am also very grateful to my colleagues and friends at the Centre for English Corpus Linguistics - Céline, Claire, Fanny, Gaëtanelle, Jennifer, Marie-Aude, Suzanne and Sylvie – for making the Centre for English Corpus Linguistics such an inspiring and intellectually stimulating research centre. I also wish to thank them for their moral and intellectual support and for all the entertaining lunchtimes we spent together talking about everyday life . . . and work. I am indebted to a great number of colleagues not only for supplying me with corpora, corpus-handling tools and references, but also for providing helpful comments on earlier versions and stimulating ideas for my research. I would like to thank Yves Bestgen, Liesbet Degand, Jean Heiderscheidt, Sebastian Hoffmann, Scott Jarvis, Jean-René Klein, Fanny Meunier, Hilary Nesi, John Osborne and JoAnne Neff van Aertselaer. I am also grateful to an anonymous reviewer for recommendations on the first draft of the text. I gratefully acknowledge the support of both the Communauté française de Belgique, which funded my doctoral dissertation out of which this book has grown, and the Belgian National Fund for Scientific Research (F.N.R.S).

And last. Magali Paquot Louvain-la-Neuve November. 2009 . but not least. Arnaud: thank you for making it all worthwhile.xii Acknowledgements On a more personal note. I would like to express my deepest thanks to my parents and friends for everything they have done to help me while I was working on this book.

1953) International Corpus of Learner English (Granger et al. 2002) International Corpus of Learner English (version 2) (Granger et al. Université catholique de Louvain Contrastive Interlanguage Analysis Constituent Likelihood Automatic Word-tagging system Corpus de Dissertations Françaises English for academic purposes English as a foreign language English as a second language English for specific purposes General Service List (West. 2000) British Academic Written English (BAWE) Pilot Corpus British National Corpus Baby BNC Academic Corpus British National Corpus – academic sub-corpus British National Corpus – academic sub-corpus (discipline: humanities and arts) British National Corpus – spoken sub-corpus Computer-assisted language learning Centre for English Corpus Linguistics.List of abbreviations AKL AWL BAWE BNC B-BNC BNC-AC BNC-AC-HUM BNC-SP CALL CECL CIA CLAWS CODIF EAP EFL ESL ESP GSL ICLE ICLEv2 IL L1 L2 LDOCE4 Academic Keyword List (my own list) Academic Word List (Coxhead... 2009) interlanguage First language Foreign language Longman Dictionary of Contemporary English (4th edition) .

Lancaster University WordSmith Tools (version 4) LOCNESS LogL MC MED2 MLD NS NNS pmw POS SLA UCREL WST4 .xiv List of abbreviations Louvain Corpus of Native Speaker Essays Log-likelihood statistical test Micro-Concord Corpus Collection B Macmillan English Dictionary for Advanced Learners (second edition) Monolingual learners’ dictionary Native speaker Non-native speaker Per million words Part-of-speech Second language acquisition University Centre for Computer Corpus Research on Language.

3: BNCweb Collocations option Figure 4.6: The phraseology of rhetorical functions in academic prose Figure 5.5: Distribution of the noun ‘solution’ Figure 3.List of figures Figure 1.3: Distribution of the words example and law in the 15 sub-corpora Figure 2.2: WordSmith Tools – WordList option Figure 2.1: The relationship between academic and sub-technical vocabulary Figure 2.4: WordSmith Tools Detailed Consistency Analysis Figure 2.1: ICLE task and learner variables (Granger et al.1: Exemplification in the BNC-AC-HUM Figure 4..2: Contrastive Interlanguage Analysis (Granger 1996a) Figure 3.1: Exemplifiers in the ICLE and the BNC-AC-HUM Figure 5.1: A three-layered sieve to extract potential academic words Figure 2.2: The distribution of the adverb ‘notably’ across genres Figure 4.3: The use of the adverb ‘notably’ in different genres 21 45 49 50 51 53 68 70 77 89 93 94 95 103 121 127 131 131 .3: The distribution of ‘by way of illustration’ across genres Figure 4. 2002: 13) Figure 3.2: The use of the prepositions ‘like’ and ‘such as’ in different genres Figure 5.5: The distribution of the verbs ‘illustrate’ and ‘exemplify’ across genres Figure 4.4: The distribution of ‘to name but a few’ across genres Figure 4.

2007b: IW13) 132 140 145 153 161 165 187 191 195 202 208 209 210 . learner writing and speech (based on Gilquin and Paquot.5: The treatment of ‘namely’ on websites devoted to English connectors Figure 5.12: Features of novice writing . 2007b: IW5) Figure 6. ‘that is’ and ‘that is to say’ (Gilquin et al.’.9: Collocational overlap Figure 5. 2007b: IW9) Figure 6.6: The use of ‘despite’ and ‘in spite of’ in different genres Figure 5..10: A possible rationale for the use of ‘according to me’ in French learners’ interlanguage Figure 5.e.2: Comparing and contrasting: using nouns such as ‘resemblance’ and ‘similarity’ (Gilquin et al.3: Reformulation: Explaining and defining: using ‘i. native-speaker and EFL novices’ writing and native speech (per million words of running text) Figure 6..7: The frequency of speech-like lexical items in expert academic writing..1: Connectives: contrast and concession ( Jordan 1999:136) Figure 6.8: Phraseological cascades with ‘in conclusion’ and learner-specific equivalent sequences Figure 5.4: Distribution of the adverbials ‘for example’ and ‘for instance’ across genres in the BNC Figure 5.11: A possible rationale for the use of ‘let us in French learners’ interlanguage Figure 5.Frequency in expert academic writing.xvi List of figures Figure 5. 2008) Figure 5.4: Expressing cause and effect: ‘Be careful’ note on ‘so’ (Gilquin et al.

16: Table 2.14: Table 2.13: Table 2.2: Table 1.7: Table 2.5: Table 2.1: Table 2.12: Table 2.2: Table 2.18: Composition of the Academic Corpus (Coxhead 2000: 220) Chung and Nation’s (2003: 105) rating scale for finding technical terms.6: Table 2.4: Table 2.10: Table 2.9: Table 2.11: Table 2.3: Table 2.8: Table 2.1: Table 1.List of tables Table 1. as applied to the field of anatomy Word families in the AWL The corpora of professional academic writing The re-categorization of data from the professional corpus into knowledge domains The corpora of student academic writing Examples of essay topics in the BAWE pilot corpus An example of CLAWS vertical output CLAWS horizontal output [lemma + POS] CLAWS horizontal output [lemma + simplified POS tags] Simplification of CLAWS POS-tags CLAWS tagging of the complex preposition ‘in terms of’ Semantic fields of the UCREL Semantic Analysis System USAS vertical output USAS horizontal output The fiction corpus Number of keywords Automatic semantic analysis of potential academic words Distribution of grammatical categories in the Academic Keyword List The Academic Keyword List The distribution of AKL words in the GSL and the AWL 12 14 17 31 32 33 34 39 40 40 41 41 42 43 44 47 47 54 55 56 60 .17: Table 2.3: Table 2.15: Table 2.

4: Table 4.13c: Table 4.10: Table 4.xviii List of tables Breakdown of ICLE essays BNC Index – Breakdown of written BNC genres (Lee 2001) Ways of expressing exemplification found in the BNC-AC-HUM The use of ‘for example’ and ‘for instance’ in the BNC-AC-HUM The use of ‘example’ and ‘for example’ in the BNC-AC-HUM Significant verb co-occurrents of the noun ‘example’ in the BNC-AC-HUM Adjective co-occurrents of the noun ‘example’ in the BNC-AC-HUM The use of the lemma ‘illustrate’ in the BNC-AC-HUM The use of the lemma ‘exemplify’ in the BNC-AC-HUM The use of imperatives in academic writing (based on Siepmann. 2005: 119) Ways of expressing a concession in the BNC-AC-HUM Ways of reformulating.9: Table 4.3: Table 4.14b: Table 4.13e: Table 4.2: Table 4.5: Table 4.12: Table 4.14a: Table 4.13a: Table 4.13f: Table 4.1: Table 4.7: Table 4. paraphrasing and clarifying in the BNC-AC-HUM Ways of expressing cause and effect in the BNC-AC-HUM Ways of comparing and contrasting found in the BNC-AC-HUM Co-occurrents of nouns expressing cause or effect in the BNC-AC-HUM reason implication effect outcome result consequence Co-occurrents of verbs expressing possibility and certainty in the BNC-AC-HUM suggest prove appear tend 69 74 Table 3.2: Table 4.14: Table 4.13: Table 4.14c: Table 4.11: Table 4.1: Table 3.6: Table 4.14d: 89 91 95 96 100 103 105 107 109 109 110 112 115 115 115 116 116 117 117 119 119 120 120 120 .8: Table 4.13d: Table 4.13b: Table 4.

9: Table 5.10: Table 5.2: Table 5.21: 158 159 162 .3: Table 5.and underuse by EFL learners of specific devices to express comparison and contrast (based on Appendix 2) Speech-like overused lexical items per rhetorical function The frequency of ‘maybe’ in learner corpora The frequency of ‘I think’ in learner corpora Examples of overused and underused clusters with AKL words Clusters of words including AKL verbs which are over.14: 149 151 154 154 156 Table 5.6: Table 5.4: Table 5.8: Table 5.16: Table 5.17: Table 5.7: Table 5.15: Table 5.13: A comparison of exemplifiers based on the total number of running words A comparison of exemplifiers based on the total number of exemplifiers used Two methods of comparing the use of exemplifiers Significant adjective co-occurrents of the noun ‘example’ in the ICLE Adjectives co-occurrents of the noun ‘example’ in ICLE not found in the BNC Significant verb co-occurrents of the noun ‘example’ in the ICLE Verb co-occurrent types of the noun ‘example’ in ICLE not found in BNC The distribution of ‘example’ and ‘be’ in the ICLE and the BNC-AC-HUM The distribution of ‘there + BE + example’ in ICLE and the BNC-AC-HUM The distribution of AKL words in the ICLE Examples of AKL words which are overused and underused in the ICLE Two ways of comparing the use of cause and effect markers in the ICLE and the BNC The over.and underuse by EFL learners of specific devices to express cause and effect (based on Appendix 1) The over.12: Table 5.18: Table 5. by comparison with expert academic writing Examples of overused clusters in learner writing Verb co-occurrents of the noun conclusion in the ICLE xix 128 129 130 133 133 134 134 135 135 143 144 146 147 Table 5.11: Table 5.20: Table 5.List of tables Table 5.1: Table 5.and underused in learners’ writing.5: Table 5.19: Table 5.

xx List of tables Adjective co-occurrents of the noun conclusion in the ICLE The frequency of sentence-initial position of connectors in the BNC-AC-HUM and the ICLE Sentence-final position of connectors in the ICLE and the BNC-AC-HUM Jarvis’s (2000) three effects of potential L1 influence Jarvis’s (2000) unified framework applied to the ICLE-FR A comparison of the use of the English verb ‘illustrate’ and the French verb ‘illustrer’ ‘let us’ in learner texts The transfer of frequency of the first person plural imperative between French and English writing Le Robert & Collins CD-Rom (2003–2004): Essay writing Table 5.23: Table 5.24: Table 5.26: Table 5.1: 205 .27: Table 5.28: Table 5.25: Table 5.29: 167 178 181 183 184 188 189 191 Table 6.22: Table 5.

Studies in second language writing have established that learning to write second-language (L2) academic prose requires an advanced linguistic competence. without which learners simply do not have the range of lexical and grammatical skills required for academic writing (Jordan. they also need to have a productive knowledge of academic language.g. A questionnaire survey of almost 5. expressing ideas in correct English and linking sentences smoothly. Mastering the subtleties of academic prose is.Introduction That English has become the major international language for research and publication is beyond dispute. International refereed journal articles are regarded as the most important vehicle for publishing research findings and non-native academics who want to publish their work in those top journals often find their articles rejected. not only a problem for novice writers. MA dissertations.000 undergraduates showed that students from all 26 departments at the Hong Kong Polytechnic University experienced difficulties with the writing skills necessary for studying content subjects through the medium of English (Evans and Green.g. 2004. Hinkel. however. and many of these difficulties involve learning to use language in new ways’ (2006: 1). Several studies have shown that the distinctive. 2005). 2007b). nature of academic prose is problematic for many novice native-speaker writers (e. reports. highly routinized. 2002) or foreign language (e. Cortes. 2002. university students need to have good receptive command of English if they want to have access to the literature pertaining to their discipline. As a result.).g. 1997. essays. Reynolds. but poses an even greater challenge to students for whom English is a second (e. 1997. Gilquin et al. 2006). As noted by Biber. Nation and Waring. . Almost 50 per cent of the students reported that they encountered difficulties in using appropriate academic style. As a large number of them are also required to write academic texts (e.g. Hinkel. partly because of language problems. 2002). ‘students who are beginning university studies face a bewildering range of obstacles and adjustments. PhD theses. etc..

Corpus linguistics is concerned with the collection in electronic format and the analysis of large amounts of naturally occurring spoken or written data ‘selected according to external criteria to represent. they find it difficult to ‘hedge’ appropriately and the structure of their texts may be influenced by their first language (see Flowerdew. which includes a number of text-handling tools to support quantitative and qualitative textual data analysis. Computer corpora are analysed with the help of software packages such as WordSmith Tools 4 (Scott. corpus-linguistic methods focus more on the co-text of selected lexical items in academic texts. Flowerdew (2002) identified four major research paradigms for investigating academic discourse. namely (Swalesian) genre analysis. they require highly explicit. Wordlists for two corpora can be compared automatically so as to highlight the vocabulary that is particularly salient in a given corpus. fiction and newspaper reportage. in other words its linguistic environment in terms of preferred co-occurrences and grammatical structures. 2005: 16). 1988: 121–60). its keywords. While the first three approaches to English for Academic Purposes (EAP) emphasize the situational or cultural context of academic discourse. 1999). a language or language variety as a source of linguistic research’ (Sinclair. Because it causes major difficulties to students and scholars alike. 1999) provides a comprehensive description of the range of distinctive grammatical and lexical features of academic prose. phrases or structures are most typical of the genre and how they are generally used. The Longman Grammar of Spoken and Written English (Biber et al. The research paradigm of corpus linguistics is ideally suited for studying the linguistic features of academic discourse as it can highlight which words. ethnographic approaches and corpus-based analysis..2 Academic Vocabulary in Learner Writing These problems include the fact that they have less facility of expression and a poorer vocabulary. Biber’s (1988) study of variation across speech and writing has shown that academic texts typically have an informational and non-narrative focus. Corpus-based studies have already shed light on a number of distinctive linguistic features of academic discourse as compared with other genres. Wordlists give information on the frequency and distribution of the vocabulary – single words but also word sequences – used in one or more corpora. conceptual or technical subject matter (Biber. academic discourse has become a major object of study in applied linguistics. i. Concordances are used to analyse the co-text of a linguistic feature. as far as possible. compared to conversation. contrastive rhetoric. 2004). text-internal reference and deal with abstract. Common features of this genre include a high rate of ..e.

2000) was compiled on the basis of corpus data to meet the specific vocabulary needs of students in higher education settings. to resolve this tension. researchers such as Hyland and Tse (2007) question the widely held assumption that students need a common core vocabulary for academic study. They argue that the different disciplinary literacies undermine the usefulness of such lists and recommend that lecturers help students develop a discipline-based lexical repertoire. This book is an attempt to resolve the tension between the particularizing trend which advocates the teaching of a more restricted. As a result. as well as idioms and collocations to develop a substantial lexical arsenal to improve their writing in English’ (Hinkel. as well as discipline-specific vocabulary. activity verbs.Introduction 3 occurrence of nouns. private verbs. I first argue that. that. 2001: 187–216). and the lack of academic vocabulary development contribute to a situation in which nonnative students are simply not prepared to write academic texts. noun phrases with modifiers. among the top priorities. This book aims to provide a better description of the notion of ‘academic vocabulary’. studies of vocabulary have emphasized the importance of a ‘sub-technical’ or ‘academic’ vocabulary alongside core words and technical terms in academic discourse (Nation. 2002: 247). It takes the reader full circle. there is a wide range of words and phraseological patterns that . nominalizations. She provides a list of priorities in curriculum design and writes that. discipline-based vocabulary syllabus. By contrast. verbs with inanimate subjects. the term has been used in various ways to refer to different (but often overlapping) vocabulary categories. the concept of ‘academic vocabulary’ must be revisited. on the basis of corpus data. Recent corpus-based studies have emphasized the specificity of different academic disciplines and genres. I demonstrate. and the generalizing trend which recognizes the existence of a common core ‘academic vocabulary’ that can be taught to a large number of learners in many disciplines. from the extraction of potential academic words through their linguistic analysis in expert and learner corpus data. derived adjectives. attributive adjectives. the relative absence of direct and focused grammar instruction. The Academic Word List (Coxhead. first and second person pronouns. agentless passive structures and linking adverbials. But what is ‘academic vocabulary’? Despite its widespread use. In addition. that-deletions and contractions occur very rarely in academic texts. to the pedagogical implications that can be drawn from the results. ‘NNSs [non-native students] need to learn more contextualized and advanced academic vocabulary. Hinkel (2002: 257–65) argues that the exclusive use of a process-writing approach.

or to perform important discourse-organizing or rhetorical functions in academic writing. Polish. I make use of Granger’s (1996a) Contrastive Interlanguage Analysis to test the working hypothesis that upper-intermediate to advanced EFL learners. range. and evenness of distribution. of scientific knowledge. Spanish. However. The learner corpus used is the first edition of the International Corpus of Learner English (ICLE). not all learner specific-features can be attributed to developmental factors. to select academic words that could be part of a common core academic vocabulary syllabus. which is among the largest non-commercial learner corpora in existence. and which may therefore be transfer-related. Finnish. and more generally. share a number of linguistic features that characterize their use of academic vocabulary. even those at the high-intermediate or advanced levels. German. It then proposes a data-driven procedure based on the criteria of keyness. A comparison of the ten subcorpora then makes it possible to identify linguistic features that are shared by learners from a wide range of mother tongue backgrounds. Swedish) are compared with a subset of the academic component of the British National Corpus (texts written by specialists in the Humanities) to identify ways in which learners’ use of academic vocabulary differs from that of more expert writers. for productive purposes. I made use of Jarvis’s (2000) unified framework to investigate the potential influence of the first language on French learners’ use of academic vocabulary in English. The comparison of several ICLE sub-corpora helps to pinpoint a number of patterns that are characteristic of learners who share the same first language. It contains texts written by learners with different mother tongue backgrounds. and they are often novice writers in their mother tongue as well. a category which has so far been largely neglected in EAP courses but which is usually not fully mastered by English as a foreign language (EFL) learners. Italian. The book is organized in three sections. The EFL learners are all learning how to write in a foreign language. academic vocabulary is more usefully defined as a set of options to refer to those activities that characterize academic work. . Dutch. and build the rhetoric of academic texts. Ten ICLE sub-corpora representing different mother tongue backgrounds (Czech. French. reviewing the many definitions of the term and arguing that.4 Academic Vocabulary in Learner Writing are used to refer to activities which are characteristic of academic discourse. The first scrutinizes the concept of ‘academic vocabulary’. Russian. irrespective of their mother tongue background. organize scientific discourse. A large proportion of this lexical repertoire consists of core vocabulary. and therefore possibly developmental.

comprises a set of 930 potential academic words. These factors include a limited lexical repertoire. One important feature of the methodology is that. the AKL includes the 2. . called the Academic Keyword List (AKL).Introduction 5 The resulting list. lack of register awareness. This section offers a thorough analysis of these lexical devices as they appear in the International Corpus of Learner English. The AKL is used in Section 2 to explore the importance of academic vocabulary in expert writing and to analyse EFL learners’ use of lexical devices that perform rhetorical or organizational functions in academic writing. The final section briefly comments on the pedagogical implications of these results. sentence-initial positioning of adverbs and transfer effects. summarizes the major findings. semantic misuse. thus making it possible to appreciate the paramount importance of core English words in academic prose. and points the way forward to further research in the area. unlike Coxhead’s (2000) Academic Word List.000 most frequent words of English. describing the factors that account for learners’ difficulties in academic writing. infelicitous word combinations.

This page intentionally left blank .

range and evenness of distribution. it can be understood in a variety of ways and used to indicate different categories of vocabulary. In this section. Nevertheless. viz.Part I Academic vocabulary ‘Academic vocabulary’ is a term that is widely used in textbooks on English for academic purposes and Second Language Acquisition (SLA) reference books. 2007b). Chapter 2 proposes a data-driven methodology based on the criteria of keyness. The AKL is used in Section 2 to analyse EFL learners’ use of lexical devices that perform rhetorical or organizational functions in academic writing. Chapter 1 therefore tries to identify the key features of academic vocabulary and to clear up the confusion between academic words and other vocabulary. This list is very different from Coxhead’s Academic Word List and has already been used to inform the writing sections in the second edition of the Macmillan English Dictionary for Advanced Learners(see Gilquin et al. my objectives are to clarify the meaning of ‘academic vocabulary’ by critically examining its many uses. . and to build a list of words that fit my own definition of the term. and uses this to build a new list of potential academic words. the Academic Keyword List (AKL)..

This page intentionally left blank .

2008). The very existence of academic words has recently been challenged by several researchers in English for Specific Purposes (ESP) who advocate that teachers help students develop a more restricted. disciplinary specificity in teaching vocabulary for academic purposes.000 words of English is not very useful for productive purposes in higher education settings and argue for a function-based definition of the term. I will show why a definition of academic vocabulary that excludes the top 2. In this chapter. consist. hypothetical. bug. Recent titles include Essential Academic Vocabulary: Mastering the Complete Academic Word List (Huntley. compare. Academic vocabulary is also sometimes used as a synonym for subtechnical vocabulary (e. I will round off this chapter by situating the book in ongoing debates over generality vs.g. But what is academic vocabulary? The term often refers to a set of lexical items that are not core words but which are relatively frequent in academic texts. likewise. mouse. Coxhead. colleague. sub-technical words and discourseorganizing words. as witnessed by the increasing number of textbooks on the topic.Chapter 1 What is academic vocabulary? Academic vocabulary is in fashion. core words. contrast.1. discipline-specific lexical repertoire.g. 2006) and Academic Vocabulary in Use (McCarthy and O’Dell. Examples of academic words include adult. differ. 2000). regardless of the discipline. transport and volunteer (cf. core vocabulary and technical terms Numerous second language acquisition studies have investigated whether there is a threshold which marks the point at which vocabulary knowledge . chemical. I set out to review the many definitions of academic vocabulary that have been given and to clear up the confusion between academic words. feature. technical terms. nuclear. cause. 1. equivalent. parallel. solution) or discourse-organizing vocabulary (e. Unlike technical terms. and identify). they appear in a large proportion of academic texts. Academic vocabulary vs.

Hirsh and Nation.000 high-frequency words. It comprises the most useful function words (e. The best-known list of core words is West’s (1953) General Service List of English Words (GSL). be. do not agree that vocabulary categories can be described as if they were clearly separable. a. 1992). They have no cultural or geographical associations. do. Next to frequency and coverage. academic vocabulary and technical terms are described and illustrated. other . Nuclear words have a ‘purely conceptual. They are also neutral with respect to tenor and mode of discourse: they are not restricted to formal or informal usage or to a specific medium of communication. Stubbs describes nuclear words as an essential common core of ‘pragmatically neutral words’ (1986: 104) and lists five main reasons for their pragmatic neutrality: 1.10 Academic Vocabulary in Learner Writing becomes sufficient for adequate reading comprehension.000 word families. They are used in preference to non-nuclear words in summarizing tasks. 5. 1992) has shown that at least 95 per cent coverage is needed to ensure reasonable comprehension of a text. Laufer (1989. Core vocabulary A core (or basic or nuclear) vocabulary consists of words that are of high frequency in most uses of the language. and up to 76 per cent of academic texts (Coxhead.g. 4. lesson. emotional or evaluative connotations’ (ibid. In this section. person. some and to) and content words like bag.g. he. 3. and technical terms.1. by. 2. Some researchers. the notions of core vocabulary. I. Percentage figures are given for different word meanings and parts of speech of each headword. The criticisms levelled at the division of vocabulary into mutually exclusive lists are then reviewed.g.e. with no necessary attitudinal. In a variety of studies. its domain of experience and social settings. They give no indication of the field of discourse from which a text is taken. put and suggest. written or spoken language. however. plus some academic words. it is commonly believed that students in higher education settings need to master three lists of vocabulary: a core vocabulary of 2. To achieve this coverage.). e.1 which was created from a five-million word corpus of written English and contains around 2. the GSL provided coverage of up to 92 per cent of fiction texts (e.1. 1. logical or propositional meaning. 2000). cognitive. about. i.

2006). Huntley. the second set of word families in the GSL provided coverage of less than 10 per cent. however. ornament and vessel) but does not contain very common words such as computer. The GSL has had a wide influence for many years and served as a resource for writing graded readers and other material.. 1984). 1. A number of criticisms have. and West’s careful application of criteria other than frequency and range’ (Nation and Waring 1997:13).. chapters from university textbooks and laboratory manuals. the GSL includes many words that are considered to be of limited utility today (e. Ghadessy. because of changes in the English language and culture.g. While the first 1. Carter.g.2. law and science. several researchers have pointed out that.g. Major. Engels (1968) criticized the low coverage of the second 1. It is now included in vocabulary textbooks (e. for educational purposes. The Academic Word List (AWL) was created from a corpus of 414 academic texts by more than 400 authors and totals around 3. Leech et al. Each sub-corpus is further subdivided into seven subject areas as shown in Table 1. It is divided into four sub-corpora of approximately 875.000 word families covered between 68 and 74 per cent of the words in the ten texts of 1. most particularly at its coverage and age. 2005. 2001: ix–x. Praninskas. necessity and style were also used in making the selection (West 1953: ix–x). it still remains the best of the available lists because of ‘its information on frequency of each word’s various meanings.000 running words he analysed.1. The Academic Word List (Coxhead. astronaut and television (see Nation and Hwang. 1971. and dictionaries (e. commerce. 1972. coal. Lynn. 2000) is the most widely used today in language teaching.g. .1. In addition. Schmitt and Schmitt.000 words representing broad academic disciplines: arts.What is academic vocabulary? 11 criteria such as learning ease. 2001).5 million words. 1973. The Academic Corpus includes journal articles.g. been levelled at the GSL. vocabulary tests (e. However. Campion and Elley. Academic vocabulary A number of academic word lists have been compiled to meet the specific vocabulary needs of students in higher education settings (e. 2006). 1995: 35–6. West also wanted the list to include words that are often used in the classroom or that would be useful for understanding definitions of vocabulary outside the list. testing and the development of pedagogical material. Xue and Nation. Schmitt et al.000 word families. 1979. crown. 1998: 207). computer-assisted language learning (CALL) materials.

international law.000 most frequent words of English as listed in West’s (1953) General Service List. Frequency: a word family had to occur at least 100 times in the Academic Corpus. The resulting list consists of 570 word families and covers at least 8.214 879. management. presuming. presumption. computer science.513. Range: a word family had to occur in all four academic disciplines with a frequency of at least 10 in each sub-corpus and in 15 or more of the 28 subject areas. It is divided into 10 sublists ordered according to decreasing word-family frequency.330 113 414 Like the General Service List. politics. industrial relations. 2. marketing. criminal law. history. environment. formula. the Academic Word List is made up of word families. pure commercial law. research. 2000: 225). psychology. it accounts for a very small percentage of words in other types of texts such as novels. context. suggesting that the AWL’s word families are closely associated with academic writing (Coxhead. issue. presumptions and presumptuous are all members of the same family.1 Academic Vocabulary in Learner Writing Composition of the Academic Corpus (Coxhead 2000: 220) Running words Texts 122 107 72 Subject areas education.723 Science Total 875. rights and remedies biology. finance. which includes all the inflections and the most frequent and productive derivational affixes. quasi-commercial law.12 Table 1. labour. benefit. geography. presumed. significant and . sociology accounting. Coxhead (2000) selected word families to be included in the AWL on the basis of three criteria: 1. economics. the words presumably. geology. public policy constitutional law. family law and medico-legal. presumes. Each family consists of a headword and its closely related affixed forms according to Level 6 of Bauer and Nation’s (1993) scale. Specialized occurrence: a word family could not be in the first 2.5 per cent of the running words in academic texts. psychology. chemistry. physics Arts Commerce Law 883.547 874.846 3. For example. Some of the most frequent word families included in Sublist 1 are headed by the word forms analyse. 3. mathematics. presume. By contrast.

Technical terms Domain-specific or technical terms are words whose meaning requires scientific knowledge. 2006: 238–9). Technical vocabulary is difficult to quantify.1. cytoplasm and abiotic. depress. Corson. Academic words are likely to be problematic for native as well as nonnative students as a large proportion of them are Graeco-Latin in origin and refer to abstract ideas and processes. resistance to semantic change and absence of exact synonyms (cf. As explained by Nation (2001: 203). likewise. 1997).000 headwords or less per subject area. technical dictionaries contain probably 1. chromatid. colleague. Language teachers are not specialists in chemistry. law or economics and may have a great deal of difficulty with technical words. Scarcella and Zimmerman (2005: 127) have also shown that mastery of derivative forms makes academic words particularly difficult for foreign language learners who often fail to analyse the different parts of complex words. In biology. 1. the boundaries between them are fuzzy . it is possible to identify them on the basis of their frequencies of occurrence. some practitioners consider that it is not the English teacher’s job to teach technical terms.3. These words are very unlikely to occur in texts from other disciplines or subject areas.What is academic vocabulary? 13 vary. 1. enormous. 1973: 228). They are typically characterized by semantic specialization. Fuzzy vocabulary categories Although core words. These words are best learned through the study of the body of knowledge that they are attached to. for example. Technical terms occur with very high or at least moderate frequency within a very limited range of texts (Nation and Hwang. genotype.4. academic words and technical terms are described as if they were clearly separable. According to Coxhead and Nation (2001).3) and to use them as a way of characterizing text types (Yang. persist and undergo.1. 1995). thus introducing additional propositional density to a text (cf. 1986). Since technical terms are highly subject-specific. learners who specialize in the field may have little difficulty in understanding these words (Strevens. Research suggests that knowledge of domain-specific or technical terms allows learners to understand an additional 5 per cent of academic texts in a specific discipline. computer science. we find words such as alleles. Examples of the least frequent word families in Sublist 10 are assemble. By contrast. range and distribution (see Section 2. Mudraya.

interaction. chest. Chung and Nation consider items at Steps 3 and 4 to be technical terms. Table 1. organs.5 per cent in the applied linguistics texts (e. moderate or low frequency. that is. cage. supports. amounts. Step 4 Words that have a specific meaning to the field of anatomy and are not likely to be used in general language. Examples are: thorax. ‘any division is based on an arbitrary decision on what numbers represent high. adjacent. ribs. academic or technical in context. such as the regions of the body and systems of the body. associated. trunk. breast. hematopoietic. 16. viscera. because vocabulary frequency. Yang. structures. or features of the body. directly. 1986. vertebrae. acquisition. demifacets. trachea. mammary. skin. They refer to structures and functions of the body. pectoralis.3 per cent of the word types at Step 3 are from the GSL or AWL (e. Words in this category may be technical terms in a specific field like anatomy and yet may occur with the same meaning in other fields where they are not technical terms. Mudraya. forms. periosteum. but not items at Steps 1 and 2. They refer to parts.2. In the anatomy texts. liver. meaning. Chung and Nation (2003) investigate what kinds of words make up technical vocabulary in anatomy and applied linguistics texts. part. muscles. intervertebral. commonly. Examples are: the. Examples are: chest. They classify technical terms on a four-level scale designed to measure the strength of the relationship of a word to a particular specialized field. it. Results for vocabulary in anatomy texts are given in Table 1. abdominal. . heart.g. as applied to the field of anatomy Step 1 Words such as function words that have a meaning that has no particular relationship with the field of anatomy. sternum. As Nation and Hwang remark. These words have clear restrictions of usage depending on the subject field. A large proportion of technical words belong to the 2. Beheydt. abdomen. shoulder. words independent of the subject matter. early and especially Step 2 Words that have a meaning that is minimally related to the field of anatomy in that they describe the positions. Examples are: superior. wall. or wide or narrow range. between. A major result of this study is that a word can only be described as general service. pairs.2 Chung and Nation’s (2003: 105) rating scale for finding technical terms.g. The words may have some restrictions of usage depending on the subject field. This increases to 50. neck. review). surrounds. protects. bony.14 Academic Vocabulary in Learner Writing (cf. input. shoulder). costal. Step 3 Words that have a meaning that is closely related to the field of anatomy. common. lodges. breathing. by. 2005). cavity. fascia. structures and functions of the body. cage. lungs.000 most frequent word families of English as given in the GSL or to the AWL. pectoral. movements. Such words are also used in general language. constantly. neck. 2006. pedicle. coverage and range figures for any text or group of texts occur along a continuum’ (1995: 37). is.

implement. They also comment that the fact that ‘items such as study appear in the GSL (but not in the AWL) and items such as drama in the AWL (but not in the GSL). As early as 1937.g. The division of vocabulary into three mutually exclusive lists becomes problematic. 2007a). find. panel.What is academic vocabulary? 15 Similarly. drama. 2009: 192). West argued that ‘both as regards Selection and still more as regards detailed Itemization..) . Hancioglu et al. The one is Recognition of a lot. however. reason. For example. use. relatively frequent in academic texts and students will most probably encounter them quite often while reading. normalize. there is a need of a divorce between receptive and productive work’ (West. result. Reading and speech bear the same relation to each other as musical appreciation and actual execution on the piano. (ibid. show) (cf. Originating from research on vocabulary needs for reading comprehension and text coverage. 2008: 462). On the other hand. argument. They should therefore be the focus of an academic reading course. whereas reading and speaking are the Hare and the Tortoise.g. These words are. the AWL includes words that are extremely common outside academia (e. however. it has been shown that the GSL contains words that appear with particularly high range and frequency in academic texts (e. suggests that the division of vocabulary into mutually exclusive lists is likely to be an activity that for all its initial convenience may prove inherently problematic in the long run’ (ibid. argue that ‘the assump˘ tion that any high frequency word outside the GSL coverage in the academic corpus would be a de facto academic item perhaps accounts for the distinctly “un-academic” texture of some of the items on the list’ (Hancioglu ˘ et al. Martínez et al.. which are not very common in everyday English. policy. These words may be used differently in academic discourse. sex. adult. concept. Partington (1998: 98) has shown that a claim in academic or argumentative texts is not the same as in news reporting or a legal report. Most English for Academic Purposes (EAP) students recognize core words but are not familiar with the meaning of academic words such as amend. when it is transposed to academic writing courses and the need arises to distinguish between knowing a word for receptive and productive purposes.: 463). 1937: 437) and regretted that teachers were giving composite lessons aiming at teaching reading and speaking simultaneously. tape) (Paquot. principle and rationalize. the other is Skill in using a little. example. the division between core words and academic words is very practical for assessing text difficulty and targeting words that are worthy of explanation when reading an academic text in the classroom.

Huntley. 2001: 27–8). (3) the first 2.000 word families which contains both technical terms and all the general words necessary for reading comprehension and shows that it provides 95 per cent coverage of many basic engineering texts (see also Mudraya. have tried to revise the General Service List. Schmitt and Schmitt. Selection is thus a key issue in teaching vocabulary for academic writing and speaking. (6) the Longman Wordwise of commonly used words and (7) the Longman Defining Vocabulary.000 words of the British National Corpus. (4) the first 5. and use it with words that commonly occur with it (Nation. ˘ See Stein (2008) for a similar approach.g. the University Word List (Xue and Nation. as well as most word lists for learners of English. . (2) the AWL.000 words of the Brown corpus.g. Gillett’s website about vocabulary in EAP < http://www. 2006).709 word families categorized according to the number of lists in which they were represented. e.. Ward (1999). Luton’s Exercises for the Academic Word List < http://www. Other examples include the GSL. groups words into families. academic vocabulary and technical terms by a single list. . It is questionable whether all the words from the AWL should be the focus of productive learning. This procedure led to the emergence of only 176 word families that were not in either the GSL or the AWL. Others. [.com> and Haywood’s AWL Gapmaker <) Several scholars have suggested replacing separate lists of general service words.uefap. regardless of specialization. 1984) and recent domain-specific lists such as those developed by Ward (1999) and Mudraya (2006). being able to pronounce and/or spell it correctly. (5) the revised version of the GSL. 2006) and CALL materials (see. 2008: 466). A final criticism that can be levelled at the AWL is related to the notion of a word family. produce it to express the intended meaning in the appropriate context. ] much of the AWL would be absorbed into it’ (Hanciog lu et al. Knowing a word productively involves. thus confirming that ‘if the GSL was enlarged by even a relatively small degree. for built an engineering word list of 2. The resulting Billuroglu-Neufeld-List (BNL) consists ˘ of 2. The AWL..16 Academic Vocabulary in Learner Writing Learning vocabulary for productive purposes has been found to be much more difficult than learning for receptive uses. And yet this strategy lies at the heart of several recent textbooks (e. for example. Coxhead (2000: 218) argues that this practice is supported by psycholinguistic evidence suggesting that morphological relations between words are represented in . to ensure maximum utility for any learner. by contrast.htm>. either a more specialized list or a larger common core vocabulary. Billuroglu and Neufeld (2007) ˘ combined into one list all the words from: (1) the GSL. 2005.

17 occurrences per million words respectively. Unlike Coxhead.29 occurrences per million words in the academic part of the British National Corpus (see Section 3. does not tell us whether the word forms issue and issues (under the headword issue) are more often used as nouns or verbs in EAP. not all members of a word family are likely to be equally helpful in academic writing. 2006). Nation (2001: 187–96) uses the term ‘academic vocabulary’ to refer to words that are not in the top 2. Table 1. and ‘specialised non-technical lexis’ .3).3 link linkage linkages linked linking links 17 Word families in the AWL proceed procedural procedure procedures proceeded proceeding proceedings proceeds issue issued issues issuing evident evidenced evidence evidential evidently item itemisation itemise itemised itemises itemising items stress stressed stresses stressful stressing unstressed utilize utilisation utilised utilises utilising utiliser utilisers utility utilities utilization utilize utilized utilizes utilizing the mental lexicon. ‘non-technical terms’ (Goodman and Payne. with relative frequencies of 0. 1981). these two lemmas are quite rare in academic writing.3 shows several word families taken from the AWL: the only information provided is that the words in italics are the most frequent form of their family. A related problem is that parts-of-speech are not differentiated. However. ‘semi-technical vocabulary’ (Farrell. he also uses it to label a whole set of lexical items also known as ‘sub-technical vocabulary’ (Cowan 1974. which has a relative frequency of 134.000 words of English but which occur reasonably frequently in a wide range of academic texts.2. 1988. For example. This may well be true and may justify the use of word families for receptive purposes. 1.06 and 1. Yang.What is academic vocabulary? Table 1. However. Mudraya. however. This. however. under the headword item. Baker. we find the noun itemisation and word forms of the verb itemise. Academic vocabulary and sub-technical vocabulary Like Coxhead (2000). 1990). 1986.

They do not offer a precise definition of the term. time sequence. measurement. Cowan defines sub-technical vocabulary as ‘context independent words which occur with high frequency across disciplines’ and comments that. Clearly some of what I am calling sub-technical vocabulary would be encompassed in the existing word frequency counts like Thorndike Lorge. for example. alternatively. Cohen et al.g. Michael West’s General Service List and the recent one million word computer analysis by Henry Kuc era and Nelson Francis. for example. thus causing problems of lexical cohesion at the level of synonymy. viz. 1990: 37). Cohen et al. wage and cage that would be categorized as technical terms according to Chung and Nation’s (2003) four-level rating scale of technicality or field-specificity (see Table 1.18 Academic Vocabulary in Learner Writing (Cohen et al. (Cowan. However. In Li and Pemberton’s (1994) view. They show that a large proportion of vocabulary items which indicate time sequence or frequency in a genetics text are unknown to their informants (e. repair processes and repair mechanism in a genetics text). or truth validity. sub-technical vocabulary as defined by Trimble (1985) is an important subset of academic vocabulary. which is a characteristic of enzymes. They showed that first-year computer science students are better able to recognize the technical meanings of sub-technical words than their non-technical meanings. Trimble (1985) extends Cowan’s (1974) usage to include ‘those words that have one or more “general” English meanings and which in technical contexts take on extended meanings’ (Trimble 1985: 129). Trimble’s definition thus encompasses words such as junction. A second area of difficulty arises because non-technical words may be used in contextual paraphrases to refer to the same concept (e. (1988) regard the extended meanings of what they call ‘non-technical’ words as a major area of difficulty for non-native readers who may only be aware of one of their meanings. In biology. consecutively. the adjective specific may also be used with reference to the genetic notion of specificity.2) (see also Farrell. For example. intermittently. circuit. ensuing. 1988). subsequent and successive). 1974: 391) ˇ Cowan’s definition of sub-technical vocabulary applies to those words that have the same meaning in several disciplines. all these terms have been used quite differently in the literature. they are quite familiar with the technical meaning of the verb compile in computer science and tend to interpret it as ‘convert . but explain that this lexis includes vocabulary items indicating.g. (1988) identify a subset of non-technical vocabulary as a third area of difficulty. ‘specialized non-technical lexis’.

Martin uses the term academic vocabulary as a synonym for sub-technical vocabulary to refer to words that ‘have in common a focus on research. i.g. Items which express notions shared by all or several specialized disciplines. 1988: 92). to describe or comment on technical processes and functions. 2. This is problematic as the non-technical meaning of a sub-technical word is often more common than its technical meaning (see Mudraya. These are ‘items which signal the writer’s intentions or his evaluation of the material presented’ (Baker. solution in mathematics and chemistry). The vocabulary of analysis includes high-frequency verbs and two-word verbs that are ‘often overlooked in teaching English to foreign students but . morphological in linguistics. 4. in preference to other semantically equivalent items. analysis and evaluation – those activities which characterize academic work’ (1976: 92). Expressed in botany is therefore not associated with emotional or verbal behaviour as is the case in general language’ (Baker. Baker thus comments that take place and occur can be regarded as subtechnical words. even in a chemical engineering thermodynamics textbook. state the hypothesis and expected results. Baker (1988) has argued that this middle area between core and technical vocabulary is itself made up of several different types of vocabulary: 1.What is academic vocabulary? 19 or translate a language into a machine code’ or ‘translate’ regardless of the context in which the word occurs.e. present the methodology. in addition to a different meaning in general language (e. The vocabulary of the research process consists primarily of verbs. the word solution is more frequently used in its non-technical sense in engineering textbooks.g. botany and biology). In botany.g. Items which are not used in general language but which have different technical meanings in different disciplines (e. For example. as opposed to being masked. develop a model). are more apparent physically. Items which are used in academic texts to perform specific rhetorical functions. nouns and their co-occurrences (e. Items which have a specialized meaning in a particular field. 5. method and function. ‘genes which are expressed have observable effects. an examination of biology textbooks showed that photosynthesis does not happen but takes place or occasionally occurs. 1988: 92). plan or design the experiment. Examples include factor. 2006). General language items which are used. 6. 3. General language items which have restricted meanings in one or more disciplines. For example. bug in computer science.

group. analyse data and evaluate results (Martin 1976).1. develop. The AWL also contains several sub-technical words as defined by Martin (1976) (e. Baker (1988) uses the term as a broad category for different types of lexical sets including both Trimble’s (1985) sub-technical vocabulary and Martin’s (1976) academic vocabulary. The noun error refers to ‘the quantity by which a result obtained by observation or by approximate calculation differs from an accurate determination’ in mathematics. The same is true of Baker’s (1988) category of words that perform rhetorical functions: case.g. result from. and study are among the top 2. found that out of 508 lemmas occurring more than five times in a corpus of electronic texts.000 most frequent words of English. For example. plan. Sub-technical vocabulary is generally defined as a category of words which are frequent across disciplines and account for a significant proportion of word tokens in academic texts.20 Academic Vocabulary in Learner Writing which graduate students need in order to present information in an organized sequence’ (ibid: 93). Adjectives and adverbs make up a large proportion of the vocabulary of evaluation. which are of various sizes and may share certain characteristics. the adjective nuclear has extended senses in astronomy. . to turn on’). psychoanalysis. Figure 1. model. compare. e. In summary. significant. cause. 44 per cent were sub-technical. base on. the many definitions of sub-technical vocabulary proposed in the literature cover very different sets of lexical items. according to the Oxford English Dictionary (OED). consist of. explanation. for example. sociology. Many of these are general service words (e. be noted for.2 partially overlap. report. function) but a large number of them do not fall within Coxhead’s definition of academic vocabulary. observe. linguistics and phonetics.g. referring to words that take on extended meanings in specific academic disciplines (Trimble. Coxhead’s (2000) Academic Word List includes a large proportion of the words that take on extended meanings in specialised fields (cf. group. Trimble’s definition of ‘sub-technical vocabulary’). This category will be the focus of the next section as it is itself made up of various sets of lexical items and Baker (1988) suggested that it is the most difficult type of sub-technical vocabulary to teach and acquire. the verb enable has a specialized meaning in computer science (‘to make (a device) operational. Farrell (1990). cause. hypothesis. derive.1 shows that the various definitions of sub-technical vocabulary and academic vocabulary as defined in Section 1. bring about. 1985). describe. or to words that allow scholars to conduct research.g. result). Definitions of sub-technical vocabulary also differ widely. method. biology. medicine. cause.

study compile base fast mouse dog bug solution 'expressed' (genes) 'masked' (genes) thereby briefly welfare hence widespread participant transport. plan.1 The relationship between academic and sub-technical vocabulary . group.What is academic vocabulary? Baker's (1988) sub-technical vocabulary 21 Coxhead's (2000) academic vocabulary Martin's (1976) academic vocabulary psychology colleague nevertheless enormous cause. model. case derive result. civil. text error enable morphological Trimble's (1985) sub-technical vocabulary Figure 1. journal. significant. show. nuclear. explanation function increase. consist present. develop. appropriate experiment. method result. hypothesis interesting. decade. observe factor. remarkable.

1997. connect. . . case. 1999). we often need to do ‘something similar to what we do when we encounter words like it. Vocabulary and the organization of academic texts Baker (1988) gave the following examples of sub-technical words that are used to perform rhetorical functions: ‘One explanation is that…’. Labels have traditionally been described as content words. differ. Vocabulary 1 consists of ‘subordinators’ which either connect clauses together (e. identify. aspect. Vocabulary 3 items serve to establish semantic relations in the connection of clauses or sentences in discourse. . anyway. and way). These words ‘may be used to make the relation explicit by saying what the relation is’ (Winter.22 Academic Vocabulary in Learner Writing 1. 1977: 15). look back in the text to find a suitable referent.3. let alone . affirm. As such. verb.or post-modified. . Luzón Marco. as far as. when we encounter them in a text. hypothetical. except that. therefore. not . in other words. ‘Others have said …’. feature. . reason. . whereas) or embed one clause within another (e.’. as . for example. .g. he and do in texts: we either refer to the bank of knowledge built up with the author. They behave grammatically like subject. that . Vocabulary 2 comprises ‘sentence connectors’ which ‘make explicit the clause relation between the matrix clause and the preceding clause or sentence’ (Winter. thus. Examples include addition. matter.. move. cause. hence. approach. . Vocabulary 3 items include a large proportion of nouns that are inherently unspecific and require lexical realization in their co-text. e. or forward. Francis (1994) refers to this type of lexical cohesion as ‘advance’ and ‘retrospective labelling’: labels2 allow the reader to predict the precise information that will follow when they occur before their lexical realization and they encapsulate and package a stretch of discourse when they occur after their realization (e. that is to say. although.). However. i. area. compare. they are part of what Widdowson (1983) called ‘procedural vocabulary’. .g. specify and subsequent. unless.e. contradict. explanation.g. Winter (1977: 14–23) distinguished between three types of words that are commonly used to create cohesion or structure in discourse and that are essential to the understanding of clause relations. . problem. highly context-dependent items with very little lexical content which serve to do things with the content-bearing words and draw attention to the function that a stretch of discourse is performing (see also Harris. consequence. either beforehand or afterwards. . . analogous. Each group is distinguished by its clause-relating functions. not so much . indeed. alike. and ‘It has been pointed out by . 1977: 22). These words bear a strong relationship with what Winter (1977) called ‘Vocabulary 3 items’ and Widdowson (1983) ‘procedural vocabulary’. method. object or complement and can be pre. result.g. .

summary.g. Within the category of labels. etc. hypothesis. instance. and adjectives. view. etc. recommendation. observation. assumption.What is academic vocabulary? 23 anticipating that the writer will supply the missing content’ (Carter and McCarthy. conviction. attitude. although there is some overlap between them: 1. Text nouns refer to the formal textual structure of discourse. analysis. position. comparison. Flowerdew. example. description. effect and result (see also Jordan. numbers. words. statement. section. they are also characterized by their specific collocational environment as shown by Francis: there is a tendency for the selection of a label to be associated with common collocations. difficulty. opinion. 1993. hinder. reasoning. The following words typically cluster round the elements of problem-solution patterns: concern. ‘the language learner who has trouble with such words may be disadvantaged in the struggle to decode the whole text as efficiently as possible and as closely as possible to the author’s designs’ (McCarthy. Mental process nouns refer to cognitive states and processes and the results thereof. they can occur in various parts of a sentence and they have a significant constant meaning’ (2001: 212). 1991: 76). belief. 1994. answer. definition. quotation. remark. consequence. labels ‘additionally give us indications of the larger text-patterns the author has chosen. Many labels are built into a fixed phrase or ‘idiom’ . interpretation. illustration. insight. Illocutionary nouns are nominalizations of verbal processes. assertion. 1988: 206–7). Francis identified a set of nouns which are ‘metalinguistic in the sense that they label a stretch of discourse as being a particular type of language’ (1994: 89). 1984. 4. contrast. argument. response. concept. suggestion. As well as representing text segments. Language-activity nouns refer to language activities and the results thereof. respond. e. detail. and build up expectations concerning the shape of the whole discourse’ (McCarthy. dilemma. idea. unlike pronouns. Nation. Metalinguistic labels are of four types.g. e. etc. finding. e. the strength of labels as discourse organizing vocabulary is that ‘they have a referential function and variable meaning like pronouns but. e. thesis. 2. 2001: 211). As pointed out by Nation.g. excerpt. theory.g. As explained by McCarthy. Hoey. term. 2008. 1991: 76). 3. reply. advice. claim. obstacle. reference. phrase. they can be modified by demonstrative pronouns. Labels not only cluster around elements of macro patterns. proof.

relations between entities (e. indicate. for example. They determine the status of the (more or less technically phrased) propositions that are laid down in it. representing a single choice. parallel. likewise. and yield are used to perform specific rhetorical functions in academic discourse. and cause/effect’ (Zwier. involve). . . ). obviously. stem from. for example. . ’. method. link. Zwier focused on lexical items that are ‘particularly useful in the kinds of writing most common in EAP writing classes – general description. . 1988: 103).g. rise. increase. proposal. . In Building Academic Vocabulary. ‘the move follows . 2004). The claim-counterclaim pattern. epistemic relations between the subject matter and the scholar (e. i. may. comprise. raise. the term ‘academic vocabulary’ has been used extensively in the literature to refer to various sets of lexical . these words express temporal deixis (e. alike.24 Academic Vocabulary in Learner Writing (in the widest sense of the word). arising from. 2002: xiii) and described the way in which words such as consist of.g. fluctuation). comparison/ contrast. original. since. later) (ibid: 10–11). Labels are not the only indicators of text patterns in academic discourse.g. Meyer’s ‘non-technical vocabulary’ is therefore closely related to the notion of metadiscourse. . distinguish. rejected/ denied the allegations’. Similarly. quantitative changes of entities (e. . ‘. above. 2002: xi) (see also Swales and Feak. . For example. ‘a specialized form of discourse which allows writers to engage with and influence their interlocutors and assist them to interpret and evaluate the text in a way they will see as credible and convincing’ (Hyland. the preposition according to and adverbs such as apparently and arguably. description of processes (especially those involving changes). classifiers of entities (e.g. . characteristic). show. to solve a problem’. (1994: 100–1) More generally. He devoted particular attention to verbs because ‘accurate verb use is especially difficult for academic writers’ (Zwier. Baker comments that sub-technical words which perform specific rhetorical functions and structure the writer’s argument ‘should not be taught in isolation but in context and as central elements in typical collocations’ (Baker. and ‘. to reverse the trend’. . the label occurs in a compatible lexical environment. adjectives like false and likely. ‘.g. follow. is often organized with verbs such as assert and state. seem). where the retrospective label is found in predictable company (. 1997: 9).e. theory.g.g. scholarly speech acts (e.g. suggest. Meyer (1997) commented that nontechnical words ‘provide a semantic-pragmatic skeleton for the text. define). and the relations between them’ (Meyer. 2005: 60). Frequent collocations include. likely). and textual deixis (e. currently). problem. Even where the collocations are less fixed. modality (e. As shown in the previous sections.

Of the 570 AWL families. An investigation of a set of potential homographs in the AWL revealed a considerable amount of semantic variation across fields. often appears in the multi-word unit marketing strategy in business. a majority of the occurrences located in just one domain. while in the social sciences it often simply means ‘considering something carefully’ (ibid: 244). with. physics and computer science (sciences sub-corpus). business studies and applied linguistics (social sciences subcorpus). 1. and social sciences sub-corpora. Hyland and Tse questioned the widely held assumption that ‘a single inventory can represent the vocabulary of academic discourse and so be valuable to all students irrespective of their field of study’ (Hyland and Tse. 227 (40%) have at least 60 per cent of all occurrences concentrated in just one sub-corpus. 534 (94%) have irregular distributions across the sciences. Is there an ‘academic vocabulary’? In an article entitled ‘Is there an “academic vocabulary”?’. mechanical and electronic engineering (engineering sub-corpus). In addition.3 million words from a range of academic disciplines is not evenly distributed. They also showed that the verb analyse tends to refer to ‘methods of determining the constituent parts or composition of a substance’ in engineering. and sociology.What is academic vocabulary? 25 items.4. learning strategy in applied linguistics and coping strategy . in many cases. They made use of Coxhead’s (2000) Academic Word List and showed that the coverage of AWL items in a corpus of 3. They gave the example of the word process which is far more likely to be encountered as a noun by science and engineering students than by social scientists. for example. The disciplines that make up the corpus are biology. Hyland and Tse further argued that ‘all disciplines shape words for their own uses’ (ibid: 240) as demonstrated by their clear preferences for particular meanings and collocations. Overall. engineering. only 36 word families were found to be relatively evenly distributed across the sub-corpora. 78 families were extremely infrequent in one sub-corpus. words may take on additional discipline-specific meanings as a result of their regular co-occurrence with other items. Science and engineering students. The noun strategy. for example. By contrast. 2007: 238). its very existence has recently been challenged by several ESP researchers. 63 in two subcorpora and 6 in all three. Of these. are very unlikely to come across the noun volume in the meaning of ‘a book or journal series’ unless they are reading book reviews. However.

with the ‘data’ and the ‘methods or tools’ varying across academic fields. cotext. for example. The authors concluded that ‘By considering context. this central concept can be defined as ‘to examine data using specific methods or tools in order to make sense of it’. 2007: 247) while also subscribing to Eldridge’s claim that ‘though one function of research is to unravel what distinguishes different fields and genres. As regards the verb analyse. An investigation of the verb analyse in a corpus of 1.351 words of business. 2009a). in an article on second language teaching for academic achievement. lexico-grammar and phraseological patterns (Granger and Paquot. Two decades ago. 2003. It is only by invoking more general definitions of this type that EAP tutors will help L2 learners deal with the various uses of verbs that they may come across even within a single discipline. 2008: 111). 2002. These findings pose a tremendous challenge to the growing number of students who enrol in interdisciplinary programmes and to English teachers who are regularly faced with mixed groups of students. phrase or sentence. Saville-Troike insisted that ‘vocabulary knowledge is the single most important area of second language competence when learning content through that language is the dependent variable’ (1984: 199). This balanced approach aims to reconcile research findings and the reality of EAP teaching practice. Wang and Nation commented that ‘learners should be encouraged to look for the central concept behind a variety of uses’ (2004: 310). The lexico-grammatical environment of the verb will help differentiate its ‘distinct (though not unconnected)’ (Hoey. analyse is also often used in the sense of carrying a statistical analysis. EAP courses need to ensure that sufficient attention is given to vocabulary development (cf. and use. 1987. academic vocabulary becomes a chimera’ (ibid: 250). another function is to find similarities and generalities that will facilitate instruction in an imperfect world’ (Eldridge. most notably in international EAP programmes (cf. the problem is to determine what words EAP tutors should teach a mixed groups of students. . if academic vocabulary is a chimera. 2008). That being the case. Eldridge.26 Academic Vocabulary in Learner Writing in sociology. Huckin.701. Sutarsyah et al. submitting data to computer-aided analyses or distinguishing the constituents of a word. 1991). 1994: 37). Granger and Paquot (2009a) advocate a ‘happy medium’ approach which concurs with Hyland and Tse’s rejection of approaches of EAP as ‘an undifferentiated unitary mass’ (Hyland and Tse. linguistics and medicine articles has shown that it is possible to identify both the common core features of an academic word and its discipline-specific characteristics in terms of meaning. 2005: 105) senses (see also Sinclair. Bhatia. In linguistics.

however.000 most frequent words of English and looked at academic texts to see what words not in the core vocabulary occur frequently across a range of academic disciplines.What is academic vocabulary? 27 1. As shown by Martínez et al. Defining academic vocabulary in opposition to core words. 2008: 464). expressing modality. and demonstrate and are often more frequent than these three AWL verbs in academic texts. however. nevertheless. I do not. Summary and conclusion There have been several studies that have investigated the vocabulary needed for academic study. This list is very useful for students entering university. textual deixis. 2000) consists of 570 word families that are not in West’s (1953) General Service List but which have wide range and occur reasonably frequently in a 3. the same rhetorical function of reporting research as establish. Some of them have assumed that learners already knew the 2. find and report are not presented as academic words because they are part of the GSL. Beheydt. irrespective of the disciplines. More recently.000 word corpus of academic texts.5. Meyer (1997) focused on words that provide a semantic-pragmatic skeleton for academic texts and identified a number of lexical subsets that fulfil important rhetorical and organizational functions in academic discourse (e. That definition should rely on the work of researchers such as Martin (1976) and Meyer (1997) who focused on the nature and role of words that occur across subject-oriented texts. The Academic Word List (Coxhead. Martin (1976) discussed words that are useful instruments in the description of activities that characterize academic work. It proves helpful in setting feasible learning goals and assessing vocabulary learning. They perform. (2009: 192). The construct of academic vocabulary remains a useful one which is. Martínez et al. analysis and evaluation. 2008: 468). conclude. scholarly speech acts). however. that is. providing words useful . I agree with Hanciog lu and her colleagues that EAP practitioners ˘ should ‘avoid taking the GSL as any kind of “given” in the compilation of more specialized wordlists’ (Hancioglu et al.g. research. ˘ subscribe to the idea according to which we ‘should seriously consider putting aside the idea of a distinct discrete-item Academic Word List’ (Hancioglu ˘ et al. as well as being an excellent resource for preparing for the reading test in International English Certificates such as TOEFL and IELTS. is of limited use when the role words play in academic discourse is examined. These GSL verbs therefore also deserve careful attention in the academic writing classroom. commented that academic words should serve ‘to build the rhetoric of a text. 2005). in need of a more precise definition (cf. the verbs show.500.

linguists and computer scientists writing in higher education settings but not to novelists. keyword analysis. I will investigate whether academic words can be automatically extracted from corpora. lawyers.28 Academic Vocabulary in Learner Writing for the construction of the argument of science’ (2009: 193). – Words that are reasonably frequent in a wide range of academic texts but relatively uncommon in other kinds of texts will not be granted the status of academic words automatically. There are. this lexical set should therefore be reasonably frequent in a wide range of academic texts but relatively uncommon in other kinds of texts. it seems reasonable to argue that. for productive purposes. sociologists. organize scientific discourse and build the rhetoric of academic texts. Academic words in their ‘functional’ sense should be useful to biologists. I will use a number of corpus linguistics techniques. and more particularly. Following Coxhead (2000). poets or playwrights. however. historians. This frequency-based criterion is not regarded as a defining property of academic words but as a way of operationalizing a function-based definition of academic vocabulary. physicists. agronomists. two major differences between this proposal and Coxhead’s work: – The 2.000 most frequent words of English may be part of a list of academic vocabulary. academic vocabulary would be more usefully defined as a set of options to refer to those activities that characterize academic work. The next step is to build a list of academic words according to this definition and it remains to be seen whether this can be done automatically. All in all. In the next chapter. . economists. To weed out those words that are not specific to academic texts.

I describe the data-driven approach used to extract potential academic words from corpora. They examine ‘linguistic (lexical or grammatical associations of the feature). Other corpus-based studies invert this relationship and investigate the characteristics of whole texts or language varieties. lemma. The term ‘potential academic words’ is used to refer to words that are reasonably frequent in a wide range of academic texts but relatively uncommon in other kinds of texts and which. Rayson proposed a different approach: ‘decisions on which linguistic features are important or should be studied further are made on the basis of information extracted from the data itself. Build: corpus design and compilation. organize scientific discourse and build the rhetoric of academic texts. Biber. by examining how certain linguistic features appear in a text (e. might be used to refer to those activities that characterize academic work. . The majority of corpus-based studies tend to focus on a particular linguistic feature. Annotate: manual or automatic analysis of the corpus.g. Common to all corpus-based studies is the prior selection of which linguistic features to study. it is datadriven’ (2008: 521). as such. multiword expression or a grammatical construction. in other words. and so be granted the status of academic vocabulary.Chapter 2 A data-driven approach to the selection of academic vocabulary In this chapter. This model is set out in five main steps: 1. The method used to extract potential academic words is based on Rayson’s (2008) data-driven approach. 1988). which draws on both the ‘corpus-based’ and the ‘corpus-driven’ paradigms in corpus linguistics. Rayson (2008) identified two general kinds of research question that can be investigated using a corpus-based paradigm. 2. possibly a word. and non-linguistic aspects (distribution of the feature across different types of texts or speech)’ (Rayson 2008: 520).

Studies that make use of the data-driven approach first focus on whole texts (Step 3) and then refine the research question or suggest specific linguistic features to study in further detail (Step 4). Rayson (2008) uses the term ‘data-driven’ to distinguish this approach from the corpus-driven paradigm.. I first detail the corpora used (Step 1) and the type of annotation adopted (Step 2). The keyword procedure is first used to retrieve a set of words which are distinctive of academic writing. 2006: 8). Interpret: interpretation of the results or confirmation of the accuracy of the model. In particular the latter approach is best viewed as an idealized extreme’ (McEnery et al. I give a description of the final list of potential academic words.30 Academic Vocabulary in Learner Writing 3. However. corpus-driven approaches to language studies is overstated. and investigate whether its constituents fit my definition of academic vocabulary. In the last part of the chapter. I then focus on the different steps undertaken to retrieve potential academic words. in which the corpus is the main informant.. 5. This provides a check on the accuracy of the retrieval procedure (Step 5). 2001: 48). The model is also testimony to the fact that the ‘distinction between the corpus-based vs. Research questions emerge from iterative analyses of the corpus data. . the Academic Keyword List. Retrieve: quantitative and qualitative analyses of the corpus. Rayson’s (2008) data-driven method thus combines elements of both the corpus-based and the corpus-driven approaches. The model bears some similarity to corpus-driven linguistics as presented by Tognini-Bonelli (2001: 85). Corpus-driven linguists question the ‘underlying assumptions behind many well established theoretical positions’ (Tognini-Bonelli. 4. Following Rayson’s (2008) data-driven approach. new categories and formulate new hypotheses on the basis of the patterns that were observed’ (De Cock. 2003: 197). Question: devise a research question or model (iteration back to Step 3). stating that pre-corpus theories need to be re-examined in the light of evidence from corpora. The advantages and disadvantages of a keyword list are discussed and the criteria of range and evenness of distribution are proposed to refine the list of potential academic words (Steps 3 and 4). They also have strong objections to corpus annotation (see McEnery et al. 2006: 9–10). It relies on pre-existing part-of-speech tagsets but considers corpus data as ‘the starting point of a path-finding expedition that will allow linguists to uncover new grounds.

(2004: 440) comment. corresponding to five broad academic domains (e.302 1.1 Corpus MC The corpora of professional academic writing Variety of English mainly British English Text type books Number of words 1. Table 2.476 196.000 words each.1. As Nesi et al.g. 2000.596 203. journal articles and textbooks.067 Arts Belief and religion Science Applied science Social science B-BNC Humanities Politics.612 219.g. As shown in Table 2. social science. 1978. both corpora consist of five sections of about 200. 2006) have principally considered book sections. ‘novice writers do not (. Coxhead.’ The automatic selection of potential academic words for this study was therefore made on the basis of an analysis of both professional and student writing. . Their early attempts at academic writing are more likely to be assessed texts produced in the context of a course study. Corpora of academic writing Corpus-based studies of vocabulary in academic discourse (e. Johansson.026.678 283. The professional academic corpora used are the Micro-Concord Corpus Collection B (MC) and the Baby BNC Academic Corpus (B-BNC). arts. science).496 199. . Texts in the B-BNC were written by British scholars while the MC also includes texts written by American researchers. includes other kinds of text than professionally edited articles and books. 1990). 2009: 62–5) is particularly well suited to extracting words that are used by all members of the ‘academic discourse community’ (Swales.316 202.021.005. This division into ‘knowledge domains’ (Hyland.060 180. Mudraya. The corpora contain about a million words of published academic prose each. Academic writing. The MC comprises 33 book sections and the B-BNC is made up of 30 book sections and extracts from scientific journals. notably student essays. or for a readership of strangers. education and law Social science Science Technology and engineering TOTAL British English books and periodicals .041 2.007 262. however.322 132.1.Selection of academic vocabulary 31 2.490 146. ) begin by writing for publication.

173. The part used for this study consists of argumentative essays written by university students and totals 168.041 Number of words 1. among others. Argumentative essay titles include. 2004).322 132. Two corpora of student writing were also used: part of the Louvain Corpus of Native Speaker Essays (LOCNESS) and a selection of texts from the British Academic Written English (BAWE) Pilot Corpus. ‘Crime does not pay’ and ‘Money is the root of all evil’.490 146.316 283. The BAWE Pilot Corpus1 contains about one million words of proficient assessed student writing. 27 per cent of the contributors .593 words.596 203.476 196. It is across this dividing line that ‘we tend to see the clearest discoursal variation and rhetorical distinctiveness’ (Hyland.2 The re-categorization of data from the professional corpus into knowledge domains Corpus ProfSS MC Arts MC Belief and religion MC Social science BNC Humanities BNC Politics.400 words) (see Granger. British university students (95.000 words in length (Nesi et al.2.612 202.209 words). ‘The National Lottery’.443 219. Together they constitute the Student Writing Corpus. 1996a. education and law BNC Social science ProfHS MC Science MC Applied science BNC Science BNC Technology and engineering 852. 1998a for further details). in the form of 500 assignments ranging from 1. ‘The death penalty’. As shown in Table 2..000 to 5.32 Academic Vocabulary in Learner Writing Table 2.695 words) and American university students (168.886 180. and humanities and the social sciences (‘soft sciences’).678 For centuries the traditional dividing line in the history of academia has been between the natural sciences and technology (‘hard sciences’). two corpora were compiled from the MC and the B-BNC: a corpus of professional ‘soft science’ (ProfSS) and a corpus of professional ‘hard science’ (ProfHS). ‘Nuclear power’.304 words and consists of argumentative and literary essays written by British A-level students (60. ‘Euthanasia’.302 262.496 199. LOCNESS totals 323. ‘Fox hunting’. There seem to be good reasons for taking knowledge domains as the point of departure for identifying potential academic words. 2009: 63).

given that their assignments have been awarded high grades’. however.3 that the procedure used to extract potential academic words largely overcomes this limitation. 2003: 1066). The Student Writing Corpus is thus quite representative of university students’ writing in that it comprises different types of writing tasks (skillsbased writing and content-based writing.300 201. It is. politics. The ‘Language studies’ sub-corpus consists of essays produced for courses in English studies. without regard for their language background’. French studies. multilingual environment.257 168. argumentative and expository writing). and in their departments students are assessed on merit. skewed towards humanities and social sciences. For the purpose of identifying potential academic words. Unlike the final BAWE corpus2.013. (2004: 444) comment that ‘the University of Warwick is a multicultural.3 Corpus BAWE Language studies Social sciences Psychology History LOCNESS Student Writing Corpus mainly American English argumentative essays The corpora of student academic writing Variety of English British English Text type assignments Number of words 845. It will be seen in Section 2. law.946 258.344 221.4 for examples). Texts in business. disciplines are not equally represented in the pilot corpus and the majority of student assignments come from the humanities and social sciences. and add that ‘all contributors are proficient users of English.841 163.3.937 . and literature.Selection of academic vocabulary 33 were not native speakers of English. Italian studies. I decided to only make use of assignments written by British students as Hinkel (2003) has shown that even English as a Second Language (ESL) students ‘continue to have a restricted repertoire of syntactic and lexical features common in the written academic genre’ (Hinkel. the texts were grouped into four sub-corpora which represent a discipline or a set of disciplines. sociology and economics were grouped together as social sciences as there were not enough texts per discipline to build separate corpora.593 1. As shown in Table 2. Essay topics in the BAWE pilot corpus are very diverse and seldom repeated (see Table 2. Nesi et al. Table 2. It could be argued that the Academic Keyword List might therefore not fully represent academic vocabulary used in the ‘hard sciences’. theatre.

2.g.. 2006. ‘corpora are useful only if we can extract knowledge or information from them. 2009.2. Mudraya. None of them discuss issues arising from the format of the corpus.34 Table 2. Various levels of annotation can be distinguished.4 Academic Vocabulary in Learner Writing Examples of essay topics in the BAWE pilot corpus – Visual arts in Britain – Prince Arthur portrayed in books – Rise of aestheticism – Modes of writing essays – Housing policy – Teachers as professionals – Would you agree that subordination was inscribed in the life of domestic servant? – Clinical depression – Psychology as a science – Expressing attitude – Is attention merely a matter of selection? – Absolutism in early modern Europe – Why did America dominate the world film market by the 1920s? – Who was to blame for the Boxer rising? Language studies [98 essays] Social sciences [64 essays] Psychology [103 essays] History [136 essays] 2.1. Section 1. which concerns the labelling of the part-of-speech (POS) or grammatical category of each word in the corpus. Corpus annotation The Academic Word List (Coxhead. we often have to begin by building information in’ (1997: 4). Corpus annotation refers to the practice of adding linguistic information to an electronic corpus of language data. Most studies of vocabulary in the field of English for specific purposes (ESP) are based on raw corpora (e. Wang et al. 2009). POS tagging is . any corpus-based study that aims to identify a specific set of vocabulary items should consider the advantages and disadvantages of annotating corpora.2.2. The fact is that to extract information from a corpus.1.). However. A second type of annotation is the morphosyntactic level of annotation. or the different conjugated forms of a verb. such as the singular and the plural forms of a noun.. starting from the addition of lemma information to each word in the corpus. 2000) is a list of word forms that were manually classified into 570 word families (cf. Coxhead and Hirsh. 2007. Martínez et al. A lemma is used to group together inflected forms of a word. Ward. Issues in annotating corpora As Leech put it. 2008.

The actual loss of information takes place when. By providing information about the grammatical nature of a word. 2004a). While it is inevitable that annotation systems will sometimes get things wrong. Although the sets of categories and features used in annotating a corpus are generally chosen to be as uncontroversial as possible. for example. One of the most widespread criticisms is that annotation reflects. 1992. either as an adjective (‘my left hand’). However discourse annotation systems. see McEnery et al. the linguist processes . although valid. A number of criticisms have been directed at corpus annotation. The argument. at least to a certain extent. 2002. For more information on the different levels of annotation. Bowker and Pearson. 2001: 73–4). Aarts. but it should be taken as a warning against the naive assumption that using annotating software is a neutral act. once the annotation of the corpus is completed and the tagsets are attached to the data. but the problem is that they are bypassed in the normal use of a tagged text. The argument can be summarized as follows: It could be argued that in a tagged text no information is lost because the words of the text are still there and available. the interpretative nature of corpus annotation has been perceived as a way of imposing pre-existing models of language on corpus data (Tognini-Bonelli. These models of language date from a ‘pre-corpus’ time and some of them derive from descriptions which ignore empirical evidence altogether (Sinclair. Although annotated data is often described as ‘enriched’ data (Leech and Smith. POS tagging is a well-researched kind of linguistic annotation and taggers perform with very high levels of accuracy. (2006: 33–43). Other levels of annotation are syntactic annotation or parsing (the analysis of sentences into their constituents). semantic annotation (the labelling of semantic fields) and discourse tagging (the annotation of discourse relations within the texts). an adverb (‘turn left’) or a noun (‘on your left’). notably by distinguished contributors to corpus-driven linguistics. are more recent and still need to be substantially refined. the various levels of annotation distinguished above are performed with varying degrees of accuracy. 2002). some theoretical perspective.Selection of academic vocabulary 35 the most popular kind of linguistic annotation applied to text. is certainly not strong enough to counterbalance all the advantages of corpus annotation. Thus. and left as a word meaning the opposite of right. annotation has also sometimes been criticized for resulting in a loss of information (Sinclair. Another argument against annotation is that it may introduce errors. 2004a: 52). 1999. it distinguishes between left as the past tense or past participle of leave (‘I left early’). it makes it possible to extract information about its various meanings and uses.

Finally. by first extracting potential academic words from annotated corpora and then returning to raw data to analyse their use in context. Part-of-speech tagged corpora will thus facilitate the extraction of specific word classes. The main objective of this chapter is to select nouns. 2001: 73). 2007b). What is lost. The software The analysis was carried out using Wmatrix.g. Elsewhere (Paquot. The tools . verbs.2. consists. adjectives. characterise/characterize. This is the price paid for simplification. and lemma + morphosyntactic tag). 2. analyse /analyze..g. consist. Word forms of lexical items that have two alternative spellings (e. I have tested the extraction procedure described in this chapter on two corpus formats. a process that is so useful – but it is argued here that the interconnection between lexis and grammar is crucial in determining the meaning and function of a given unit: any processing that loses out on this is bound to lose out in accuracy. centre/center. it is worth stressing that this chapter does not attempt to meet a theoretical objective. behaviour/behavior) were lemmatized under the same headword (either the British variant or the most frequently used option). therefore. If lemmas are used. uni-functional items —tags — as the primary data. By doing this the linguist will easily lose sight of the contextual features associated with a certain item and will accept single. and shown that using lemmatised corpora makes it possible to identify some 31 per cent more lexical verbs that are typical of academic texts than using unlemmatised corpora.2. Even the linguists who have directed the most severe criticisms at annotated data acknowledge that the ‘good point of annotation lies in its value in applications’ (Tognini-Bonelli. (word form + morphosyntactic tag. the different inflectional forms of a verb (e. Rather it is content with an applied aim. is the ability to analyse the inherent variability of language which is realised in the very tight interconnection between lexical and grammatical patterns. consisted. (Tognini-Bonelli 2001: 73–4) The data-driven methodology adopted in this book aims to preserve the best of both worlds. and consisting) are merged and so a better frequency distribution for the lemma across texts is obtained.36 Academic Vocabulary in Learner Writing the tags rather than the raw data. a web-based corpus processing environment which gives researchers access to several corpus annotation and retrieval tools developed at the University Centre for Computer Corpus Research on Language (UCREL) at Lancaster University. adverbs and other function words that are commonly used in academic texts.

Part-of-speech tagging is essentially a disambiguation task. e. as if. in the light of. abound can only be a verb and kindness is always a noun). prepositions such as as opposed to. Co-occurrence probabilities are often automatically derived by training the software on manually disambiguated texts. Hoskins who came to dinner?)5. (see Ide. A tagger needs to determine which part-ofspeech is most probable.e. Although close to 90 per cent of English types4 can only be one part-of-speech (e. compounds such as tabula rasa. which can be a determiner (Do you remember that nice Mr. use. matter-of-fact and grown up. issue. Multiword expressions include adverbs such as a bit. at least. i.g. and conjunctions such as even though. all the same.Selection of academic vocabulary 37 available in Wmatrix include the Constituent Likelihood Automatic Word-tagging System (CLAWS) and the UCREL Semantic Analysis System (USAS) (see Rayson. 1997). 2005). 2003). and by and large. craft. 1988: 31). provided that.g. a conjunction (I can’t believe that he is only 17) or an adverb (I hadn’t realized the situation was that bad!). a relative pronoun (The people that live next door). contrary to. brand new. The Constituent Likelihood Automatic Word-tagging System A corpus uploaded to the Wmatrix environment is first grammatically tagged with the Constituent Likelihood Automatic Word-tagging System (CLAWS) (Garside and Smith. and so that. over 40 per cent of the running words (or tokens) in a corpus are morphosyntactically ambiguous (DeRose. Most current part-of-speech taggers use an approach to disambiguation which is at least partly probabilistic: they rely on co-occurrence probabilities between neighbouring tags. Typical rule-based taggers use context frame rules to assign tags to unknown or ambiguous words. The tagger makes use of a detailed set of 146 tags3 (CLAWS C7 tagset). It also uses two lexicons: (a) a lexicon of single words with all their possible parts of speech and associated lemmas. etc. Another very common source of ambiguity in English is homography between verbs and nouns. and so forth. given the immediate syntactic and semantic context of a homograph. given that x is a determiner. For example. they are spelt the same but belong to different word classes. the probability that the item to its immediate right is a noun or an adjective can be calculated. An example of a context frame rule is ‘if an ambiguous or unknown . and in comparison with. cause. and (b) a multiword expression lexicon. Many words are part-of-speech homographs. Non-probabilistic or rule-based taggers have also been making a comeback with systems such as that proposed by Brill (1992). abandon. This is largely due to the ambiguity of a number of high-frequency words such as that. because of.

belong). words are not always separated by blanks (e. 2003: 63). 3. A sentence is generally described as a string of words followed by a full stop. that is. viz. a word ending in *ly will be classified as an adverb.28).. Thus. The tagger is commonly described as going through five major stages (Garside. If a particular word is not found in the tagger’s lexicon. CLAWS is a hybrid tagger.38 Academic Vocabulary in Learner Writing word is preceded by a determiner and followed by a noun. An initial part-of-speech assignment: Once a text has been tokenised.e. If a word is ambiguous.8 or 14. The probabilistic tag-disambiguation program: The task of the probabilistic tag-disambiguation program is to inspect all the cases where a word has been assigned two or more tags and choose a preferred tag by considering the context in which the word appears and assessing the . and in situ as an adverb (see below for more details on ditto-tags and their advantages)).e. it is assigned a single tag. 1987): 1. i. combining both probabilistic and rule-based approaches. they’re). in contractions such as don’t.). belongs to only one part-of-speech category or word class (e. fire is first tagged as a noun and then as a verb.g. person. it’s. 4. a string of letters surrounded by white spaces. fire). A full stop does not. cause. as well as is tagged as a conjunction. the tagger assigns part-of-speech tags to all the word tokens in the text without considering the context. a task which is not trivial. it is assigned several tags listed in decreasing likelihood. tag it as an adjective’. if it can belong to more than one word class (e. e.g. A rule-based contextual part-of-speech assignment: This stage assigns a single ‘ditto-tag’ to two or more orthographic words which function as a single unit or multiword expression (e.g. fig. a word is generally considered as an orthographic word. If a word is unambiguous.g. This hybrid approach allows CLAWS to assign POS-tags with a very high degree of accuracy – 97–98 per cent for written texts (Rayson. in figures (5. Voutilainen (1999) has surveyed the history of the different approaches to word class tagging. however.g. A pre-editing or tokenisation phase: This stage prepares the text for the tagging process by segmenting it into words and sentence units. i. 2. morphological rules. Thus. because the probability of it being a noun is higher than that of it being a verb. a word ending in *ness will be classified as a noun.. However.). Dr. Similarly.g. boat. use. it is assigned a tag based on various sets of rules. always signal the end of a sentence (e. for tagging unknown items. etc.. title nouns (Mr.e. and other types of abbreviations (i.

– the probability that one tag follows another.5 shows a typical CLAWS vertical output: each line represents a running word in the corpus and gives its POS-tag and lemma.Selection of academic vocabulary 39 probability of any particular sequence of tags.6). for example. I have written a Perl program which takes this intermediate format as its input and creates a corpus with lemmas followed by their POS-tags (Table 2. and – the probability of a word being assigned a particular tag from the list of all its possible tags (Garside and Smith. Word form The whole point of the play seems to be an attack on the Church . Table 2. the word run has been assigned both a noun and a verb tag. The probability of a tag sequence is typically a function of. that on number given by the tags NN1 (singular common noun) or DD1 (singular Table 2.5 An example of CLAWS vertical output POS-tag AT JJ NN1 IO AT NN1 VVZ TO VBI AT1 NN1 II AT NN1 .6 is that the word forms are replaced by their lemmas. for example. 1997: 104). despite the fact that run is more often a verb than a noun. If. The intermediate format has the advantage of allowing researchers to select the information needed. while the POS-tags are too specific for our purposes. The problem with the format shown in Table 2. Output: The output data can be presented in intermediate format (vertical output for manual post-editing) or final format (horizontal and encoded in SGML). Redundant information includes. 5. Lemma the whole point of the play seem to be an attack on the church PUNC . it is less likely to be classified as a verb if it appears in the vicinity of another verb.

where II stands for a general preposition.7 CLAWS horizontal output [lemma + simplified POS tags] the_AT whole_JJ point_NN of_IO the_AT play_NN seem_VV to_TO be_VB an_AT attack_NN on_II the_AT Church_NN . for grammatical purposes. the annotation format has to be slightly modified to do this. VVZ: -s form of lexical verb._PUNC Where AT: article. As a result. JJ: adjective. frequency lists based on this format generate different frequencies for ‘example_NN1’ and ‘example_NN2’. The expression ahead of is an example of a sequence of two graphemic words treated as a single preposition. Finally. However. NN1: singular common noun. each CLAWS7 tag can be modified by the addition of a pair of digits to show that it occurs as part of a sequence of similar tags. It receives the tags: ahead_II21 of_II22. A word list based on CLAWS horizontal output would thus distinguish between the preposition in (in_II) and the preposition in used as the first word of three-word sequences (such as in terms of ) (in_II31).9 shows the CLAWS vertical output for the complex preposition in terms of. and the second digit the position of each graphemic word within that sequence. Ditto tags are very useful as they make it possible to extract complex prepositions. The simplification routines are presented in Table 2. II: general preposition. Table 2.6 after simplifi cation of the POS-tags. It would not be able to retrieve the complex . Such ‘ditto tags’ are not included in the lexicon but the program assigns them via an algorithm which is applied after initial part-of-speech assignment and before disambiguation by looking for a range of multiword expressions included in a pre-established list. Table 2. and complex conjunctions as well as single words that are typical of academic discourse. Each graphemic word of the complex preposition is tagged and lemmatised independently. The first of the two digits indicates the number of graphemic words in the sequence._PUNC determiner) and that on verbal forms given by the tags VVZ (.40 Table 2. are best treated as a single unit. infinitive.8.form of a s lexical verb) or VVG (-ing form of a lexical verb). POS-tags were therefore simplified by a Perl program to match the level of specificity of the lemmas. AT1: singular article. VBI: be. PUNC: punctuation Table 2. IO: of (as preposition). TO: infinitive marker ‘to’.6 Academic Vocabulary in Learner Writing CLAWS horizontal output [lemma + POS] the_AT whole_JJ point_NN1 of_IO the_AT play_NN1 seem_VVZ to_TO be_VBI an_AT1 attack_NN1 on_II the_AT Church_NN1 . representing a group of graphemic words which.7 shows the same sentence in Table 2.

VDN (done).g. inch) NP (proper nouns) NPD (weekday noun) NPM (month noun) MC1. infinitive). VHD (had).g. VBDR (were). e. hundred) NNT (temporal nouns. e.g. VVN (past participle). MC2 NN1. DAT (most. base form). be bound to). few) JJ (adjective) DAR (more. JJT Verb forms VB (be) VB0 (be. fewest) JJR. VVGK (-ing participle catenative. be going to). VHN (had). VBN (been). VVNK (past participle catenative.g. VBZ (is) VD0 (do. street) NNO (numeral nouns. VBI (be. VBM (am). VHG (having). infinitive).g. e. NPD2 NPM1. less). NNT2 NNU1.Selection of academic vocabulary Table 2. NN2 NNL1. VVZ (-s form) VD (do) VH (have) VV (lexical verbs) Table 2. VBR (are). e. VHZ (has) VV0 (base form of lexical verb). NNO2 NNT1. plural forms MC (cardinal number) NN (common nouns) NNL (locative nouns. VDG (doing).g. VDD (did). VDZ (does) VH0 (have. little. day. island. VDI (do. VVD (past tense). e. NPM2 41 Simplified POS tags Comparative and superlative forms DA (after-determiners. base form).8 Simplification of CLAWS POS-tags CLAWS7 POS tags Singular vs. VBDZ (was). infinitive).9 CLAWS tagging of the complex preposition ‘in terms of’ POS-tag II31 II32 II33 Word form in terms of Lemma in term of . e. week) NNU (units of measurement. VHI (have. e. base form). NP2 NPD1.g. VVG (-ing participle). NNL2 NNO. much. VVI (infinitive). NNU2 NP1.

travel and transport Numbers and measurement Substances.g. objects and equipment Education in general Language and communication Social actions. Letters Table 2. expand into 232 categories (see Archer et al. and any other words that are linked in other ways with the concept concerned. question. explain. The USAS tagset includes 21 major semantic fields (see Table 2. in-terms-of II). states and processes Science and technology Names and grammar .g. which.10). in_II31 terms_II32 of_II33) by the component words.42 Academic Vocabulary in Learner Writing preposition itself. materials. 1997: 54). separated by a hyphen and followed by their POS-tag (e. The UCREL Semantic Analysis System A second layer of annotation was applied by the UCREL Semantic Analysis System (USAS). location. states and processes Food and farming Government and public Architecture.10 Semantic fields of the UCREL Semantic Analysis System A B C E F G H I K L M N O P Q S T W X Y Z General and abstract terms The body and the individual Arts and crafts Emotional actions. This tool assigns tags representing the general semantic field of words from a lexicon of single words and multiword expressions. reply. and explanation. message. A semantic field is a theoretical construct which groups together ‘words that are related by virtue of their being connected – at some level of generality – with the same mental concept’ (Wilson and Thomas. anecdote. query. Another Perl program was therefore used to replace any sequence of words with ditto tags (e. This includes not only synonyms and antonyms of a word but also its hypernyms and hyponyms. states and processes Time World and environment Psychological actions. statement. house and the home Money and commerce in industry Entertainment. sports and games Life and living things Movement.. in turn. 2002). feedback. For example. the category ‘language and communication’ (Q) includes words such as answer. response.

states and processes – groups and affiliation’).2). It assigns a semantic field tag to every word in the text with about 92 per cent accuracy.2 I2.1 (‘language and communication – media – books’). semantic tagging can be subdivided broadly into a tag assignment phase and a tag disambiguation phase. The next stage consists of selecting the contextually appropriate semantic tag from the set of potential tags provided by the tag assignment algorithm. The program makes use of a number of sources of information in the disambiguation phase.1. states and processes – religion and the supernatural’) and T1.1 A1. the word chapter has been assigned the tags Q4.1.Selection of academic vocabulary 43 are used to denote the major semantic fields while numbers indicate field subdivisions.1[i1. For example.2 G2.1 S6+ Y1 Z5 S1.3 S9/S5 S5+ A1. S5 (‘social actions.2 Z5 Z5 X4. The program Table 2.2 L1- .1 T1.1 I2.2 M1 E1 S1.11 shows that in the sentence ‘This chapter deals with the approach of the criminal law to behaviour which causes or risks causing death’. Semantic tag M6 Z5 Z8 Q4. (‘time-period’). and contextual rules (Rayson.2 represents a word in the category ‘general and abstract words’ (A).K5. First. These categories are all marked with a Z-tag.1.3.1 G2. Word form This chapter deals with the approach of the criminal law to behaviour which causes or risks causing death .1.2. domain of discourse. the semantic tag A2.2 Z5 A15A2.2 F3/I2.1 Z5 Z5 G2. the subcategory ‘affect’ (A2) and more precisely the sub-subcategory ‘cause / connected’ (A2. Table 2.1 Z8 Z5 A2.1[i1. 2003: 67–8). a set of potential semantic tags are attached to each lexical unit.11 USAS vertical output POS-tag DD1 NN1 VVZ IW AT NN1 IO AT JJ NN1 II NN1 DDQ VVZ CC VVZ VVG NN1 . The semantic annotation does not apply to proper names and closed classes of words such as prepositions.1G2.1 A9.1.2.A5. notably POS-tags. Like part-of-speech tagging. conjunctions and pronouns. S9 (‘social actions.

12). In the USAS lexicon of multiword expressions.1 which_Z8 causes_A2. For example. The word families included had to be outside the first 2.2. the template ‘ma[kd]*_V* {JJ. at the drop of a hat.1.. Thus. and idioms (e.1 as the semantic tag with the highest correctness probability.1 law G2.1 with_Z5 the_Z5 approach_X4. makes little sense. This is displayed in the final output format (see Table 2.g.1[i1. 2. It thus retrieves all instances of the expression make sense and its variants make no sense. Multiword expressions are analysed as if they were single words. to bark up the wrong tree.12 Academic Vocabulary in Learner Writing USAS horizontal output This_M6 chapter_Q4. advisory committee. as represented by West’s (1953) General Service List.1.e.g.2 (see Table 2. criminal law is tagged as: criminal G2. The word chapter in the sense of ‘an ecclesiastical assembly of priests or monks’ is a case in point.000 most frequent words in English. by the skin of one’s teeth) are described as regular expressions or templates.11).2.1[i1. phrasal verbs (e.1[i1.3.2 death_L1._PUNC ranked these semantic tags and chose Q4.2.g. AT*} sense_NN1’ identifies all occurrences of the verb make directly followed by an optional adjective (JJ).44 Table 2. 1996) to select words that met three frequency-based criteria: 1. D*. It belongs equally to the semantic fields of ‘groups and affiliation’ and ‘religion’. The same occurrence of a word in a text may simultaneously signal more than one semantic field. i. compounds (e.2 to_Z5 behaviour_S1. take off). etc. . bank account). break out.2.2 or_Z5 risks_A15 causing_A2.1 deals_A1. made more sense. The two semantic tags are thus assigned in the form of a single tag S9/S5 (see Table 2. Automatic extraction of potential academic words Coxhead (2000) made use of the Range corpus analysis program (Heatley & Nation.1 law_G2.2 of_Z5 the_Z5 criminal_ G2.12).1[i1. using ditto-tags similar to those used in part-of-speech tagging. determiner (D*) or article (AT*) and the singular noun (NN1) sense. academic year. sequences of words. parts of words and grammatical categories used to match similar patterns of text and extract them.

the resulting list would have included a large number of function words and other high-frequency words that tend to be frequent in the English language as a whole (cf. Range 3.1) but which are not necessarily the most representative lexical items in the Academic Corpus. with a frequency of at least 10 occurrences in each sub-corpus and in 15 or more of the 28 subject areas. Two quantitative filters.1). These words and other high-frequency words will only occur in a keyword list “if their usage is strikingly different from the norm established by the reference text” (Archer.Selection of academic vocabulary 45 2. 1.2. A member of a word family had to occur in all 4 disciplines represented in the Academic Corpus.1. Section 1. Section 1. To address this limitation. the procedure described in this book is primarily based on keyness (Scott. applying Criterion 1 makes it impossible to identify high-frequency words that are particularly prominent in academic texts. namely range and evenness of distribution. Members of a word family had to occur at least 100 times in the corpus (cf. Keyness 2.). 3. Archer. 2009a: 3). a fully data-driven method that is often used in corpus linguistics to find salient linguistic features in texts (e.g.1.1 A three-layered sieve to extract potential academic words . 2001). are subsequently used to narrow down the resulting list of potential academic words (Figure 2. If Criterion 1 had not been used. Evenness of distribution Potential academic words Figure 2. On the other hand. 2009) and which does not require the use of a stop list to filter out function words.

are filtered out. the reference corpus was not chosen to represent all the varieties of the language6 but to serve as a ‘strongly contrasting reference corpus’ (Tribble. or are not significantly more frequent in the research corpus than in the reference corpus. words that are statistically prominent in the research corpus. What the text ‘boils down to’ is its keyness. The word list for the research corpus is reordered in terms of the keyness of each word type. 2001: 396). words that have strikingly low frequency in the research corpus in comparison to the reference corpus. A minimum frequency threshold is usually set at 2 or 3 occurrences in the research corpus. L (mystery and detective fiction). The procedure to identify keywords of a particular corpus involves five main stages (see Scott and Tribble. The K (general fiction). 2. and terminological items typical of specific sub-disciplines of English for information science and technology (Curado Fuentes. 5.e. Thus.3. 2006: 58–60): 1..e. Thus. the blah blah blah’. Words that occur less frequently than the threshold in the research corpus. ‘keyness is a quality words may have in a given text or set of texts. the adornment. as well as negative keywords. i. words typically used by men and women with cancer in interviews and online cancer support groups (Seale et al. As emphasized by Scott and Tribble (2006: 55–6). suggesting that they are important. 2006: 59). they reflect what the text is really about. e. 3.46 Academic Vocabulary in Learner Writing 2. and (b) be outstandingly frequent in terms of the reference corpus’ (Scott and Tribble. Keyness Keyword analysis has been used in a variety of fields to extract distinctive words or keywords. 4. usually the log-likelihood ratio. 2006). avoiding trivia and insignificant detail. . Frequency-sorted word lists are generated for a reference corpus and the research corpus. 2001). the two corpora of professional writing and the corpus of student academic writing described in Section 2. business English words (Nelson. 2000). i.1 were each compared with a large corpus of fiction on the grounds that academic words would be particularly under-represented in this literary genre. once we have steamed off the verbiage.1. The two lists of word types and their frequencies are compared by means of a statistical test. Software tools usually list positive keywords.g. then it (a) must occur at least as frequently as the threshold level. ‘for a word to be key. For the purposes of this research.

L.Selection of academic vocabulary Table 2.). P) FLOB (categories K. methane. N.322 4.946. Positive keywords are more numerous than negative keywords for each academic corpus. 2004). jurisdiction. simply because they are under-represented in fiction writing (e. not all of the keywords meet the definition of academic vocabulary in Section 1. in lemma + POS-tag format.5. the ProfHS corpus and the Student Writing Corpus. . This can be explained by the large amount of specialized vocabulary present in academic texts.01 with a critical value of 15.g.g. bacterium.14 Number of keywords Corpus ProfHS corpus ProfSS corpus Student Writing corpus Positive keywords 4.492 Negative keywords 837 1.13 (see Rayson et al.13 The fiction corpus Corpora LOB (categories K. N. M. L. enzyme. the BROWN corpus and the FROWN (Freiburg-Brown) corpus7 were combined with the Baby BNC fiction corpus (Table 2. The keyword procedure selects all words that occur with unusual frequency in a given text/corpus.025 Table 2. martyr. M. M. e. N.14 gives the number of positive and negative keywords for each corpus. 2004). factor and participant in student writing. rape. which means that there is less than 1 per cent danger of mistakenly claiming a significant difference in frequency. However. the FLOB (Freiburg Lancaster-Oslo/Bergen) corpus. L.688 1. law. DNA. Keywords were extracted for the ProfSS corpus. The resulting list is therefore likely to include technical words that do not occur in all types of academic texts. etc. N. L. N (adventure and western fiction) and P (romance and love story) categories of the LOB (Lancaster-Oslo/Bergen) corpus. archbishop. chromosome. offence and policy in law (soft science) and theory. P) FROWN (categories K. penicillin. compared to a reference corpus.201 956 M (science fiction). P) Baby BNC fiction TOTAL Number of words 946..13) to form the reference corpus for this study. formula.656 4. P) BROWN (categories K. cell and species in biology (hard science). Keyness values were calculated with the Keyness module of WordSmith Tools 4 (Scott. The significance of the log-likelihood test was set at 0. M. Table 2.337 47 999.

The criteria of range and evenness of distribution were subsequently used to refine the list of potential academic words still further. 1996).1 with the WordList option of WordSmith Tools (Scott. It is calculated on the basis of the 15 sub-corpora described in Section 2. 2. 2004). ‘which suggests that this word is key because of a single author’s use of a word in a specific case.2. I wrote a Perl program which automatically compares keywords for several corpora and creates a list of positive keywords that are shared in the ProfHS corpus. rather than being something that indicates a general difference in language use’ (2004: 350). the keyword status of wuz is more a function of the sampling decision to include one particular narrative in the corpus than evidence of the distinctiveness of the word in gay male erotic narratives (see also Oakes and Farrow. The words ability. able and about. Baker shows that wuz (used as a non-standard spelling of was) appears to be a keyword of gay male erotic narratives. Although the resulting number of keywords fell by more than 60 per cent. when in fact its use is restricted to one single text. In other words. lesbian erotic narratives. This tool can take several corpus files as input and range comes automatically with any word list it produces. 2007: 110).48 Academic Vocabulary in Learner Writing The keyword procedure relies on the conception of a corpus as one big text rather than as a collection of smaller texts. are shown to appear in all 15 sub-corpora. the procedure cannot distinguish between ‘global’ and ‘local’ keywords (Katz. 2. a phenomenon which Katz (1996: 19) has referred to as ‘burstiness’8. in 100 per cent of the corpora .e. Statistical measures such as the log-likelihood ratio are computed on the basis of absolute frequencies and cannot account for the fact that ‘corpora are inherently variable internally’ (Gries.3. Figure 2. For example.2 shows that there is a column headed ‘Texts’ which shows the number of texts each word occurred in. Range Range (i. 2007: 91). Global keywords are dispersed more or less evenly through the corpus while local keywords appear repeatedly in some parts of the corpus only. Scott’s (1997) notion of ‘key keywords’). in a keyword analysis of gay male vs. As a first step to overcome this inherent limitation of the keyword procedure. for example. the ProfSS corpus and the Student Writing corpus (cf. the number of texts in which a word appears) is used to determine whether a word appears to be a potential academic keyword because it occurs in most academic disciplines or because of a very high usage in a limited subset of texts.048 shared keywords were still identified. that is. As a consequence.

only words appearing in all 15 academic sub-corpora were retained as potential academic words. the meaning of which is more discipline or topic-dependent (e.g. the criterion of range excludes the words sector.3 and accurately reflect the difference between them. the law of gravity).2 WordSmith Tools – WordList option analysed. The large variation in the range of law can be explained by the peak frequency of occurrence of the noun in the professional soft science sub-corpora. while that of the word law varies between 11 and 812. range has an important limitation in that it gives no information on the frequency of a word in each sub-corpus. criminal law. which we intuitively regard as an academic word.Selection of academic vocabulary 49 Figure 2. The frequencies of these two words in each sub-corpus are shown in Figure 2. Used alone. . For the purposes of this study. Thus. paradigm and variance as they only appear in 11 sub-corpora but includes both the word example. The frequency of the word example ranges from 26 to 226 in the 15 sub-corpora. and the word law. canon law.

This is the last criterion I applied to restrict the list of potential academic words. the number of sub-corpora or texts) in the corpus. it will be well-distributed’ (Zhang et al.e..50 900 800 700 600 500 400 300 200 100 0 Academic Vocabulary in Learner Writing LAW EXAMPLE Figure 2. Its values range from 0 (most uneven distribution possible) to 1 (perfectly even distribution across the sectors of the corpus) (see Oakes (1998: 189–92) and Gries (2008) for more information on dispersion measures). but the exact number of times it appears’ (Oakes and Farrow 2007: 91).3.3. it will appear in different parts of the corpus. One such measure is Juilland’s D statistical coefficient.3 Distribution of the words example and law in the 15 sub-corpora 2. The variation coefficient V is given by V = s / x where x is the mean sub-frequency of the word in the corpus and s is the standard deviation of these sub-frequencies. Figure 2. And if the word is used commonly enough. 1964) and is calculated as D = 1 – V / √n-1 where n is the number of sectors (i. A number of studies have used a measure of dispersion to define a core lexicon on the basis that ‘if a word is commonly used in a language. Evenness of distribution Differences in range can be highlighted by a measure of the evenness of the distribution of words in a corpus. This measure takes into account ‘not only the presence or absence of a word in each subsection of the corpus. Juilland’s D was first used in the Frequency Dictionary of Spanish Words (Juilland and Rodriguez. 2004). The evenness of the distribution or dispersion of a word is ‘a statistical coefficient of how evenly distributed a word is across successive sectors of the corpus’ (Rayson. Juilland’s Ds were calculated for each word using the output list from WordSmith Tools Detailed Consistency Analysis.4 gives an example of a M BEL C BN SO C C BN HU C M BN PO C L SO C M MC C SC AP P BN SC BN C SC B C BA AW TE C W E E A H BA H RT W IST S E O PS R Y BA YC W HO E LO SO C C N ES S M C M C AR TS . 2003: 93).

4 WordSmith Tools Detailed Consistency Analysis 51 .Selection of academic vocabulary Figure 2.

conversely. These frequencies were copied into an Excel file and normalized per 100. whereas the noun law. Some words have skewed Juilland’s D values because of their polysemy. result and illustrate. the mean sub-frequency and the standard deviation) were computed in Excel and Juilland’s D values were then calculated for each word. the variation coefficient. Examples of words that have D values lower than 0. significance. Evenness of distribution is the only criterion used that could perhaps favour keywords that are more prominent in the different parts of the ProfSS corpus and the Student Writing corpus However. The measures necessary to calculate Juilland’s D values (i. Its general meaning is found in all academic sub-corpora while its technical meaning is restricted to scientific writing and accounts for its much higher frequency in the two professional scientific sub-corpora (MC-SC and BNC-SC in Figure 2. The noun solution is a case in point.83. and more specifically in the social science sub-corpus. appear. a relatively high minimum threshold of 0. perceive and isolate. and confirm that only example is of widespread and general use in this genre. and treatment. provide. highly and above. difference.52 Academic Vocabulary in Learner Writing detailed consistency analysis: the second column (Total) gives the total frequency of each word in the whole corpus.8 reduces the .8. the cut-off point of 0.9 For a word to be selected as a potential academic word. employment. effective. show. Dispersion values make it possible to avoid the mistaken conclusion that these two words behave similarly in academic writing.8 and were therefore not selected include the nouns health.e.8 appears to be too restrictive and words that would intuitively be considered as academic words are excluded. as it has both a general meaning (‘a way of solving a problem’) and a technical meaning (‘a liquid in which a solid or gas has been mixed’) with different frequencies and distributional behaviours. verbs such as prove. discuss. adjectives such as significant. similar and likely and adverbs such as particularly. was not selected. At times. The resulting list of 599 potential academic words includes nouns such as conclusion.69. and the verbs label. and consequence. The noun example was thus identified as a potential academic word as its dispersion value was 0. the third column (Texts) gives the range of each word and the following columns show its frequencies in each sub-corpus. with a Juilland’s D value of 0.000 words as the 15 sub-corpora are of different sizes. while the noun law is over-represented in the professional soft science corpus. extent. its Juilland’s D value had to be higher than 0. possibility of giving too much weight to words that would be particularly frequent in the ‘soft science’ sub-corpora but much less common in the ‘hard science’ sub-corpora. . personality.5).

Section 2. the category ‘general and abstract terms’ includes almost half the potential academic words. circumstance. Table 2. Examples include the nouns activity. The category ‘numbers and measurement’ accounts for more than 10 per cent of the potential academic words and includes nouns (e.3. I therefore made use of a semi-automatic procedure to identify words that did not pass the dispersion criterion but were semantically related to the 599 potential academic words.5 Distribution of the noun ‘solution’ These two peak frequencies of occurrence are responsible for the relatively low D value (54. Some words were automatically classified into more than one category but the figures given are based on the semantic tag most frequently attributed to each word.2 discussed how a text uploaded to the web-based environment Wmatrix is morphosyntactically and semantically tagged. the adjectives detailed and particular and the adverbs similarly and conversely.4. and limitation as well as the verbs perform and cause. high. amount. adjectives (e. measure.Selection of academic vocabulary 300 250 200 150 100 50 0 AR TS C BE L M C SO BN C C H U BN M C PO BN LI C M SO C C SC M IE C AP N L S BN C C BN SC C I BA TE C W H E A BA RT S BA WE H W IS E PS T C BA HY W E LO SC C N ES S M 53 Figure 2. in particular. It is notable that 87 per cent of the 599 potential academic words fall into just six of the categories.2.g.g.6) of the noun solution. 2. the 15 sub-corpora are relatively small and the frequencies of occurrence of words may be skewed by a particular topic or author’s preferred turn of phrase. M C .15 shows the distribution of the 599 potential academic words across these semantic classes. Broadening the scope of well-represented semantic categories More generally. degree. The semantic analysis was conducted with the UCREL System which classifies words and multiword units into 21 major semantic categories. extent).

Money and commerce in industry K. analyse. Social actions. whether). subsequently. encourage. states and processes’ (e.3 per cent. Substances.g. extend. Emotion F. On this basis.g. Arts and crafts E.3 9.g. 8.6 0. and ‘language and communication’ (e. location.8 but which formed part of one of the six semantic categories described above were added to the list of potential academic words.7 0. define.3 0. Numbers and measurement O. Architecture.0 0. Time W.0 12. assumption. social. suggest) represent 9. prepositions (such as. states and processes’ (e.7 5.g. verbs (e. in addition to).3 100 Semantic categories A. materials.7 per cent and 5. wide). therefore)). Government and public H. Language and communication S. which is morphologically related to the potential academic verb analyse. was retrieved by the semantic criterion although .2 0.3 0. World and environment X. Many of the words that were retrieved by this additional criterion are morphologically related to words that had already been automatically selected. house and the home I.54 Table 2. since. Food and farming G. thus. impose). adverbs (e. Science and technology in general Z.3 1. 331 keywords that did not have Juilland’s D values higher than 0. according to. Movement.2 0. frequently.0 0.3 0. attempt). Psychological actions. reduce).0 2. ‘names and grammar’ (mainly consisting of connective devices such as conjunctions (or. Education in general Q. states and processes T. Live and living things M.3 8. The body and the individual C.7 0.2 per cent.g. claim.g.7 per cent of the potential academic words respectively. General and abstract terms B. argue. The categories ‘psychological actions. For example. sports and games L. states and processes Y.2 0. objects and equipment P.15 Academic Vocabulary in Learner Writing Automatic semantic analysis of potential academic words Number of words 267 2 2 4 0 4 2 7 0 0 12 74 7 4 34 47 26 2 55 2 50 599 Percentage of words 44.4 1. ‘social actions.7 4. increase. Entertainment. facilitate. 7. interpretation. the noun analysis. conclusion. Names and grammar TOTAL large. also) and prepositions (e. travel and transport N. during) and adverbs (moreover.7 7.

Second.17. states and processes’.4.3 identified 930 potential academic words on the basis of four criteria. states and processes’ and ‘language and communication’ – were also included. they had to be well-distributed across the corpora and have a Juilland’s D value higher than 0. ‘names and grammar’. the first of which is keyness. The criteria of minimum frequency and range still apply: the noun analysis was retrieved only because it is very frequent in academic prose and appears in a wide range of academic texts.e.35 9. First. Table 2. The resulting list of potential academic words has been named the Academic Keyword List (AKL) to emphasize the fact that it is the output of a data-driven set of criteria. This is consistent with Biber et al. However this is not an argument for using word families instead of lemmas. Other morphologically related words such as analyst or analysable were still excluded from the list. they had to be characterized by wide range. ‘numbers and measurement’.8. ‘social actions.05 19.17 per cent of all potential academic words.35 8.16 presents a breakdown of the AKL by grammatical category. the words had to be keywords in professional (both ‘hard’ and ‘soft’ disciplines) and student academic writing. Third. Keywords that did not match this last criterion but belonged to one of the six best represented semantic categories – ‘general and abstract’.’s (1999) finding that nouns are particularly frequent in academic prose.17 25. to appear in all 15 sub-corpora representing different academic disciplines. The complete list is given in Table 2. i. and not a list of academic vocabulary in its functional sense. A large proportion of the nouns in the list are Table 2.06 100 . Nouns make up 38. ‘psychological actions.16 Distribution of grammatical categories in the Academic Keyword List Number Nouns Verbs Adjectives Adverbs Others Total 355 233 180 87 75 930 Percentage 38.8. The Academic Keyword List The (semi-)automatic extraction procedure described in Section 2. 2.Selection of academic vocabulary 55 its Juilland’s D was below 0.

phenomenon. target. procedure. cause. assertion. aspect. effect. tradition.Table 2. classification. perception. quality. rate. unit. analysis. diversity. viewpoint. feature. interaction. learning. survey. volume. destruction. bias. failure. improvement. structure. influence. discrimination. form. resolution. examination. code. logic. attitude. factor. mode. achievement. convention. property. reason. link. strategy. extreme. literature. difficulty. absence. maintenance. criticism. emphasis. support. alternative. author.17 The Academic Keyword List 355 nouns ability. communication. outcome. majority. success. risk. recognition. theme. variety. representative. measure. isolation. adoption. basis. figure. community. assistance. importance. norm. separation. birth. whole. choice. belief. essence. doctrine. member. occurrence. set. crisis. control. solution. division. exclusion. conduct. standard. understanding. level. observation. minority. movement. guidance. case. finding. option. participant. policy. culture. presence. part. space. research. benefit. challenge. difference. approach. loss. validity. past. similarity. spread. introduction. sample. experiment. compromise. commitment. implication. parallel. perspective. transition. list. application. attention. combination. resource. advice. source. damage. creation. operation. manipulation. behaviour. constraint. correlation. stress. person. data. definition. material. fact. description. interest. aim. country. tolerance. kind. possibility. gain. growth. event. publication. complexity. integration. trend. reproduction. uncertainty. contrast. dimension. capacity. hypothesis. comparison. defence. tension. section. argument. need. establishment. action. relevance. function. restriction. amount. world 56 Academic Vocabulary in Learner Writing . evaluation. discovery. extent. construction. sex. adult. likelihood. example. protection. practice. instance. knowledge. disadvantage. group. reduction. concern. report. concentration. error. advance. limit. pressure. mankind. selection. centre. change. guideline. organisation. scope. advantage. parent. character. population. female. statistics. use. process. proportion. indication. progress. output. variation. institution. resistance. impact. network. account. identity. assessment. reader. reality. point. contradiction. range. observer. conclusion. production. society. summary. means. existence. investigation. reference. model. association. series. purpose. rule. review. search. element. evidence. technique. degree. information. colleague. addition. condition. limitation. work. activity. experience. consensus. scale. class. force. result. intervention. exposure. individual. exception. dilemma. contribution. age. content. characteristic. idea. topic. consideration. type. awareness. decline. attempt. programme. effectiveness. number. tendency. expansion. team. increase. scheme. position. requirement. theory. conflict. discussion. problem. explanation. environment. service. motivation. version. percentage. evolution. shift. insight. term. relationship. pattern. male. demand. rise. reasoning. lack. potential. interpretation. circumstance. stimulus. committee. balance. respect. conception. debate. notion. situation. distinction. concept. determination. decision. formation. value. view. criterion. period. system. sense. future. question. method. assumption. skill. medium. development. relation. personality. consequence. category. subject. issue. provision. being. task. opportunity. significance. act. analogy. role. proposition.

consistent. dependent. increase. possess. significant. appear. demonstrate. natural. large. encounter. generate. inherent. reveal. show. different. be. contrast. past. initial. write. explain. expand. imply. true. relative. far. express. pose. systematic. main. induce. decline. correspond. interesting. create. constitute. apparent. aim. allow. combine. pursue. severe. ensure. abstract. differentiate. describe. destroy. experimental. negative. mental. potential. restrict. crucial. profound. general. surprising. initiate. excessive. dominate. local. logical. damage. enhance. final. function. exclude. achieve. require. fixed. outline. measure. isolate. operate. precede. major. successive. respond. maintain. inadequate. apply. regulate. valuable. accessible. sustain. future. formal. adopt. confine. single. conduct. alternative. indirect. establish. other. inferior. effective. unlike. impose. separate. identical. normal. become. explicit. common. previous. transform. great. assist. immediate. solve. include. emerge. fail. support. identify. attribute. related. important. record. link. separate. undertake. label. note. obvious. real. deal. maximum. submit. typical. undermine. average. early. locate. permanent. vary. concentrate. minor. equal. physical. stable. possible. connect. rely. basic. comprise. misleading. can. extend. predict. valid. cite. extreme. indicate. conform. publish. comprehensive. formulate. passive. sufficient. control. design. prevent. govern. form. strict. central. expose. coincide. parallel. total. integrate. fundamental. wide. high. individual. suffer. examine. particular. conventional. result. follow. promote. reduce. produce. ideal. contain. participate. report. standard. consist. convert. retain. conclude. attain. traditional. selective. employ. characterise. active. unsuccessful. refer. reject. distinct. assert. extensive. critical. subsequent. obtain. secondary. exemplify. cause. additional. responsible. treat. suitable. rational. overcome. evaluate. frequent. theoretical. investigate. use. involve. argue. summarise. discuss. prove. stress. recent. limited. difficult.233 verbs accept. aid. resolve. improve. distinguish. psychological. seek. adapt. influential. useful. considerable. reflect. tend. complete. partial. develop. incomplete. realistic. available. certain. unlikely. successful. evident. various. equivalent. define. dominant. present. suggest. essential. late. unique. differ. claim. neglect. benefit. allocate. advance. represent. direct. compare. term. introduce. receive. productive. symbolic. restricted. leading. regard. display. interpret. minimal. vital. concern. male. primary. arise. following. stimulate. render. limit. assume. assess. emphasize. assign. experience. favourable. focus. exist. classify. reproduce. should. acceptable. progressive. clarify. original. state. remain. highlight. relate. internal. adequate. favour. relevant. eliminate. diminish. contribute. mutual. supply. replace. act. study. correct. specific. necessary. principal. simple. varied. compete. widespread (Continued) Selection of academic vocabulary 57 . determine. incorporate. lead. practical. strengthen. may. select. tackle. sexual. human. associate. perform. acquire. permit. divide. evolve. affect. yield 180 adjectives absolute. positive. propose. construct. applicable. prime. complex. encourage. gain. provide. alter. substantial. rapid. prominent. scientific. random. actual. social. independent. lack. clear. acute. avoid. reinforce. perceive. overall. new. attempt. likely. competitive. facilitate. effect. modern. illustrate. quote. consider. present. exceed. preserve. attend. view. account (for). base. detailed. so-called. enable. representative. similar. remove. special. depend. appropriate. derive. occur. influence. advocate. finance. radical. arbitrary. analyse. choose. specify. visual.

widely 75 others according to. as well as.. or. in that.17 (Cont’d) 87 adverbs above. since. originally. specifically. typically. in. consequently. third. less. by. generally. potentially. the. virtually. same. in response to. both. themselves. which. given that. clearly. either. these. primarily. versus. fewer. upon. subsequently. subject to. unlike. as. most. notably. however. purely. extremely. commonly. of. namely. directly. basically. successfully. rather than. most. in common with. because. their. conversely. provided. socially.g. contrary to. prior to. in relation to. wholly. simply. depending on. because of. despite. similarly. therefore. although. only. fully. former. in the light of. within Academic Vocabulary in Learner Writing . to. whether or not. per. this. than. adequately.58 Table 2. necessarily. several. more. in particular. for instance. as opposed to. first. accordingly. less. that. strongly. whereas. especially. equally. those. traditionally. such. increasingly. even though. closely. its. indirectly. independently. readily. at best. accurately. in terms of. latter. also. approximately. highly. effectively. hence. often. ultimately. each. whether. initially. far. explicitly. thereby. thus. particularly. greatly. e. in general. including. normally. such as. recently. moreover. secondly. some. previously. in addition to. in favour of. essentially. an. fairly. mainly. as to. little. significantly. over. other than. inevitably. partially. solely somewhat. relatively. frequently. between. second. from. itself. indeed. for. for example. largely. considerably. correctly. further. during. many. due to.

differ. become.g. conversely. potential. show.g. prove) and logico-semantic relationship verbs (e. interpret. inadequate. concept. argue. (3) words of the Academic Word List. inevitably. In order to calculate the percentages of GSL and AWL words in the list of potential academic words. explain.g. different. thus) and evaluative (e. according to. correctly. following. because of). analyses the vocabulary in the text. additional. There are also complex conjunctions such as whether or not and given that. comparison. conjunctions.2.g. appear. interesting. articles. linking verbs (e. argument. parallel. deal. illustrate. mental verbs (e. possibility (e. definition. The category ‘other’ includes prepositions. This web interface takes a text file as its input.2): some 40 per cent of the potential academic prepositions are complex (e.g. hypothesis.Selection of academic vocabulary 59 abstract terms which belong to Francis’s (1994) category of metalinguistic labels: illocutionary nouns (e. determiners and ordinal numbers. assess. alternative. The majority of adjectives express value judgements. So it is logical that. similar).g.g. contrast.g. include). provide).000 most frequent words of English. use. articles and determiners are not fully lemmatised by CLAWS.g. theory. such as. The syntactic function of adjectives is to modify nouns and noun phrases. likely. increasingly. pronouns.35 per cent of the potential academic words in this list. these and those are not analysed as word forms of the same determiner and are categorized as separate lemmas. appropriate.g. in terms of. As Conrad (1999) pointed out. adequately. if nouns are frequent in academic writing. imply. analysis. reporting verbs (e. useful). term). the Academic Keyword List was uploaded to the Web Vocab Profile developed by Tom Cobb10. arise. incomplete. (2) the second 1. clear. respond). contrast. unlikely) and logico-semantic relationships (e.g. either positive or negative (e. the category of potential academic adverbs (9. hence.000 most frequent words of English. discuss. Pronouns. Verbs account for 25 per cent of the list and include verbs that Hinkel (2004) classified as activity verbs (e. and (4) other words. equivalent.g. Before uploading the list of potential academic words. and classifies the words into four main categories: (1) the first 1. and description). final. possible. adjectives represent 19. prime. Although usually disregarded in academic textbooks and teaching materials. note).g. mental process nouns (e. it was necessary to remove multiword units (such as the adverbs for example and for instance and the complex prepositions . examine. this. consequently. highly. significantly) words.g. be.g. language-activity nouns (e. effectively. Inspection of this list illustrates the value of CLAWS ditto tags (Section 2. and view) and text nouns (e.35%) consists essentially of linking (e. For example. follow. question). numerous adjectives will be used to qualify them. however. adequate. section. cause. emphasize.

2. emphasize. effect. degree. prior to. scope. Other well-represented semantic categories . render.5. result and stimulate. For example. produce. 1993: 237). cause. typically accurately. reason. viewpoint AWL 40% Other 3% in addition to. influence. criticism. the adjectives dependent.18 Lists GSL % 57% Academic Vocabulary in Learner Writing The distribution of AKL words in the GSL and the AWL Examples aim.18). proposition. cause. hypothesis. relation. in favour of and because of) since Web Vocab Profile cannot deal with them and would simply decompose multiword units into their parts. lead. link. argument. proportion. effect. adequate. result and stimulus. Table 2. example. depend. point. result. compare. versus. importance. because. base. in view of. influence. while 57 per cent are among the 2. namely. connect. The results show that only 40 per cent of the potential academic words in my list also occur in the Academic Word List. measure.60 Table 2. explanation. extent. therefore. which provides the retrieval procedure with some evidence of face validity. result). in response to. An important application of word frequency lists for course design is to uncover ‘functional and notional areas which might be important for the syllabus’ (Flowerdew. include. and thus. implication. correlation. evidence. define. in relation to. the semantic sub-category named ‘A2. in the light of. The Academic Keyword List is a list of potential academic words. likelihood. frequency-based criteria are not used as a defining property of academic words but as a way of operationalizing a function-based definition of academic vocabulary. therefore. approach. in respect of. difference. the prepositions depending on. Affect: cause – connected’ consists of the nouns basis. explain. argue. increasingly. in terms of and subject to. show. motivation. relevance. in the light of. The automatic semantic analysis has shown that a number of semantic sub-categories are particularly wellrepresented. The comparison is thus based on single words only. condition. provided and since. derive.g. the adverbs consequently. analysis. summary. validity. impact. attribute. induce. conclusion. comparison. the verbs associate. differ. assess. generate. and the conjunctions because. consequence. link. because of. related and resulting. survey. reason. discuss. on account of. exception. These results highlight the important role played by general service vocabulary in academic writing. due to. inherent. result. method.17 shows that a high proportion of AKL words fit the functional definition of academic vocabulary (e. whereas assertion.000 most frequent words of English as described in West’s (1953) General Service List (see Table 2. factor. hence. consequence. conclude. with respect to. reference. experience. comprise. As explained in Section 1. thereby. due to. given that. theory. combine. correlation. tackle. exemplify. relate.

primary). include). content. discussion. words that are reasonably frequent in a wide range of academic texts but relatively uncommon in other kinds of texts and which.1. The Academic Keyword List is not a final product and does not in itself ‘carry any guarantee of pedagogical relevance’ (Widdowson. Importance’ (e. organize scientific discourse and build the rhetoric of academic texts’. extent.g. groups.1. scope. suggest) and ‘X2. arise. kind.5. limited. female. etc’ (e. parent. favourable. a sense that will not be useful to all scholars who are writing in higher education settings. General kinds. consider. progress. similar. It still needs ‘pedagogic mediation’ (Widdowson. The main objective of Chapter 2 has been to operationalize this function-based definition using frequency-based criteria. 2002: 151).2. Evaluation’ (e.Selection of academic vocabulary 61 include ‘A1. ‘A4.1. circumstance. event. The (semi-)automatic method used relies on Rayson’s (2008) data-driven model. instance. Summary and conclusion In Chapter 1. considerable. Similarly.g.1. 1993: 237). refer.g. ‘N5. view. ‘A11. emphasize. category. Quantities’ (e. examples’ (e.g. widely. example. making. 1991: 20–1). fundamental. ‘A1. illustrate). The fact that these AKL words are used to serve particular functions in academic prose has ‘to be corroborated by concordancing’ (Flowerdew. ‘A5. exclude.g. also means ‘to put pictures in a book’. classify. significance. sex and world). and to retrieve potential academic words (i. Comparing’ (e.8. improve. Thought and belief’ (e.e. assume. General actions. amount. . the nouns country. major. the verb illustrate and the preposition like are often employed as exemplifiers but also have different uses.g. several). academic vocabulary was defined as ‘a set of options to refer to those activities that characterize academic work. The keyword procedure was first used to retrieve a set of words which are distinctive of academic writing. conversely). as such. account. Inclusion / exclusion’ (e. for example.g. activity. male. might count as academic vocabulary) from several corpora of academic prose.g. consist.g. difference. perform). the Academic Keyword List includes several words that are relatively well distributed keywords in academic prose but should probably not be part of a list of academic vocabulary (e. positive). 2. definition. The verb illustrate. ‘Q2.g. figure. enhance. case. which combines elements of both ‘corpus-based’ and ‘corpusdriven’ paradigms in corpus linguistics. quote. Words such as the noun illustration. formulate). Speech acts’ (e. 2003) and is thus better conceived of as a ‘platform from which to launch corpus-based pedagogical enterprises’ (Swales. unlike. describe. contrast. ‘A6.

This methodology has limitations. confirming the accuracy of the retrieval procedure. ‘the one thing that EAP seems to lack is a corpus that includes all levels of data – from presessional students’ writing and speech to academic lectures. and published research articles and books. some of which may be shared with the corpus under study. the criteria of range and evenness of distribution were used as a sieve to refine the list of potential academic words. AKL words have been shown to fall in semantic categories such as cause and effect. a high proportion of the words included in the Academic Keyword List match the definition of academic vocabulary. Such a corpus would need to include as many disciplines as possible. ‘psychological actions. at present no corpus exists that represents all the varieties of academic discourse. a limitation inherent in the keyness approach is the use of a reference corpus. the minimum number of texts in which a keyword must appear. As Krishnamurthy and Kosem (2007: 370) comment. ‘social actions. it is likely that a few potential academic words passed unnoticed because they are also used in fiction. a number of keywords that had failed the dispersion test were selected on the basis of a semantic criterion: they belonged to one of the six semantic categories (‘general and abstract’. or the building of the argument of academic texts. evaluation. and speech acts. A reference corpus is itself characterized by a set of distinctive linguistic features. Despite these limitations. ‘names and grammar’. First. with sufficient detailed categorization to enable the users (teachers or students) to select a customized set of corpus texts appropriate for their needs’. comparison. 2007: 92. ‘numbers and measurement’. inclusion/exclusion. Third. There is thus a strong case for using ‘strongly contrasting reference corpora’ (cf. PhD theses.62 Academic Vocabulary in Learner Writing Second. 2001: 396). The method provides a good illustration of the usefulness of annotation for the development of practical applications such as the Academic Keyword List. Paquot and Bestgen. importance. states and processes’. The Academic Keyword List might have looked quite different if other cut-off points had been adopted. the results are dependent on a number of arbitrary cut-off points: the probability threshold under which log-likelihood ratio values are not considered significant. states and processes’ and ‘language and communication’) that included most potential academic words. which clearly relate to academic work. irrespective of differences in meaning or function. 2009). However. and the minimum coefficient of dispersion (see Oakes and Farrow. Third. quantities. The method has also made . However. Second. the organization of scientific discourse. the corpora used are relatively small and contain a limited number of text types. Tribble.

.. The Academic Keyword List still needs validation. 2008: 461). A number of general service words take on prominent rhetorical and organizational functions in academic discourse and their absence from lists such as Coxhead’s (2000) Academic Word List may be highly problematic in academic writing courses.Selection of academic vocabulary 63 it possible to appreciate the paramount importance of general service words in academic prose. and sentence position. phraseology. Each AKL word ˘ should be subject to a careful corpus-based analysis to confirm its status as an academic word and establish how it is used in academic prose in terms of meaning. However it unquestionably offers ‘a portal into the complex behaviour and intricate relationships of individual lexical items’ (Hanciog lu et al.

This page intentionally left blank .

The function of ‘exemplification’ is presented in detail. share a number of linguistic features that characterize their use of academic vocabulary. and proposed a data-driven methodology to extract potential academic words from corpora.Part II Learners’ use of academic vocabulary Having clarified what academic vocabulary is by providing a critical overview of its many definitions. In Chapter 4 the Academic Keyword List is shown to include a large number of lemmas that can be used to organize academic texts and structure their content around logico-semantic relations. Chapter 4 also focuses on the phraseology of academic words. not all the features of their writing are shared across the . This illustrates the type of data and the results obtained when the whole range of lexical strategies available to expert writers for establishing cohesive links in their texts is examined. and can therefore be assumed to be developmental. I focus on AKL words that are used to organize discourse and build the rhetoric of a text. Chapter 3 offers a detailed description of the corpora used and the method adopted to compare them: Granger’s (1996) Contrastive Interlanguage Analysis (CIA). I will now examine EFL learners’ use of academic vocabulary. The EFL learners are all learning how to write in a foreign language. However. analysis and evaluation. The learner corpus used is a collection of texts taken from the International Corpus of Learner English. rather than on words that focus on research. Many of the linguistic features that characterize learner writing are shared by learners from a range of mother tongue backgrounds. Chapter 5 aims to test the working hypothesis that upper-intermediate to advanced EFL learners. irrespective of their mother-tongue background. and they are often novice writers in their mother tongue as well. As learner texts are short argumentative essays.

66 Academic Vocabulary in Learner Writing different language backgrounds. Chapter 5 therefore ends with a brief discussion of the potential influence of their mother tongue on French speakers’ use of academic vocabulary in English. . and the differences may be due to transfer from the writers’ mother tongues.

Belgium. I will focus on learners’ use of academic words that serve typical organizational or rhetorical functions in academic discourse. 3. Learner-profile questionnaires give two types of information: learner characteristics and information on the type of task. 2003). share characteristic ways of using academic vocabulary. which is among the largest non-commercial learner corpora currently existing. Each learner text is documented with a detailed learner-profile questionnaire. The method of analysing learners’ use of academic vocabulary and comparing it with that of expert writers relies on Granger’s (1996a) Contrastive Interlanguage Analysis. Granger. A computer learner corpus is an electronic collection of (near-) natural language learner texts assembled according to explicit design criteria (see Granger. which all the learners were requested to fill in. . 2009 for the design criteria. In the second part of this book. In this chapter.. The International Corpus of Learner English The learner data used consist of ten sub-corpora of the International Corpus of Learner English version 1 (ICLE) compiled at the University of Louvain. A subset of the British National Corpus is used as a control corpus and helps bring to light features of learner language. academic vocabulary has generally been described as a major source of difficulty for EFL learners.1. I will investigate whether upper-intermediate to advanced EFL learners. 2005 for a discussion of different types of learner data).Chapter 3 Investigating learner language Whatever the definition adopted. 2002. and contains texts written by learners with different mother-tongue backgrounds. I describe the corpora and methodology used to pursue these objectives. irrespective of their mother tongue backgrounds. and Ellis and Barkhuizen. under the supervision of Sylviane Granger (Granger et al. The learner corpus used is the International Corpus of Learner English.

2003: 539). ‘Money is the root of all evil’ and Feminists have done more harm to the cause of women than good. They were learners of English as a Foreign Language rather than as a Second Language and were in their third or fourth year of university study. as part of an exam or not. French.g.2 A large proportion of the learner texts are argumentative. In the words of the old song.1. genre (academic essay). Finnish.1 Learner productions have quite a few task variables in common. Czech.1 ICLE task and learner variables (Granger et al. Dutch. Polish. essay topic and task settings). notably in terms of medium (writing). Russian. which were used as corpus-design criteria. Italian. On the basis of these external criteria. The ICLE learners represent 11 different mother tongue backgrounds: Bulgarian. their level is described as advanced although ‘individual learners and learner groups differ in proficiency’ (Granger. mother tongue background of learners. German. Topics include Most university degrees are theoretical and do not prepare students for the real word. 2002: 13) As shown in Figure 3.68 Academic Vocabulary in Learner Writing International Corpus of Learner English Shared features Variable features Learner variables Age Learning context Proficiency level Task variables Medium Field Genre Length Learner variables Gender Mother tongue Region Other FL L2 exposure Task variables Topic Task setting Timming Exam Reference tools Figure 3.. L2 exposure. field (general English rather than English for Specific Purposes) and length (between 500 and 1.000 words). but the ICLE essays cover a wide range of topics. Other variables differ (e. ICLE texts share a number of learner and task variables. All the learners who submitted an essay to the ICLE were university undergraduates and were therefore usually in their twenties. Spanish and Swedish. Essays differ in task conditions: they may have been written in timed or untimed conditions. with reference tools such as .

potentially written with the help of reference tools. of words Average no.1 Breakdown of ICLE essays No. .768 162.521 165.Investigating learner language Table 3. More information on the types of analyses that can be performed with WordSmith Tools can be found on Mike Scott’s webpage (http: //www. Most studies of the ICLE data to date have not taken these task settings into consideration (see Ädel’s (2008) analysis of timed and untimed essays for an exception). researchers in second and foreign language acquisition and teaching insist that the influence of task type and condition is important (Shaw.119 48. Although essays written without the help of reference tools would arguably have been more representative of what advanced EFL learners can produce. The software was used to examine the occurrences of potential academic words in context.292 136. Learner essays in each sub-corpus were carefully selected in an attempt to control for the task variables which may affect learner productions: all the texts are untimed argumentative essays.165. 2004).1. 1990). As shown in Table 3. of words per essay 890 828 750 598 612 604 636 855 665 593 697 69 Czech (ICLE-CZ) Dutch (ICLE-DU) Finnish (ICLE-FI) French (ICLE-FR) German (ICLE-GE) Italian (ICLE-IT) Polish (ICLE-PO) Russian (ICLE-RU) Spanish (ICLE-SP) Swedish (ICLE-SW) TOTAL 147 196 167 228 179 79 221 194 149 81 1641 130. 2004.524 grammars and dictionaries or not. However.739 140. of essays No.3 The ICLE sub-corpora compiled for this study were analysed with the Concord tool of the computer software WordSmith Tools 4 (Scott. Task and learner variables can be used to compile homogeneous and in Scott and Tribble (2006). Kroll.343 109.060 1. The Clusters option proved very useful for identifying the most frequent n-grams or lexical bundles that contained the words being studied. I chose to select untimed essays with reference tools as they represent the majority of learner texts in ICLE.937 99. lexically. this study makes use of ten ICLE subcorpora.243 125.556 47. representing different mother tongue backgrounds.

Firth and Wagner. 1998). i.70 Academic Vocabulary in Learner Writing 3. CIA compares varieties of one language: native and non-native varieties (L1/L2). A strong argument that can be invoked in defence of the CIA model is that the native speaker norm used in learner corpus research is explicit and corpus-based (Mukherjee. which consists of comparing two or more languages. L1/L2 comparisons bring out the features of non-nativeness in learner productions. the English of French speakers compared to that of Dutch speakers). Contrastive Interlanguage Analysis The methodology most frequently used to analyse learner corpora is Contrastive Interlanguage Analysis (CIA) (Granger 1996a.2. 2002). Gilquin 2000/2001). 1996a) . Granger. 2001.. however. or shared by several learner populations (and therefore likely to be developmental or due to other causes such as teaching methods) (cf. Leech describes the native control corpus as ‘a standard of comparison. or different non-native varieties (L2/L2) (see Figure 3. 1997).e. for comparing learner language to a native speaker norm and thus failing to analyse interlanguage in its own right (see Lakshmanan and Selinker. on the other hand. In his preface to Learner English on Computer (Granger. Unlike contrastive analysis.and under-use of linguistic items or structures as a question of downright errors’ (Granger et al.g. a norm against which to measure the characteristics of the learner corpora’ (Leech. 1983). 1998: xv). ‘which at an advanced level are as much (if not more) a question of over. make it possible to assess whether these features are peculiar to one language group (and thus possibly due to the influence of the learner’s mother tongue).2). 2005) rather than implicit and intuitionbased as has been common in second language acquisition (SLA) studies CIA L1 > < L2 L2 > < L2 Figure 3.2 Contrastive Interlanguage Analysis (Granger. Comparisons of different interlanguages (e. 2009: 41). In second language acquisition research. L1/L2 comparisons have generally been criticized for being guilty of the ‘comparative fallacy’ (Bley-Vroman.

2009: 20). 2009) and that other varieties of English are sometimes used as control corpora. Lakshmanan and Selinker’s point may be understood as a plea for more natural language data (i. (2) who know to a large extent what is acceptable in a given communication situation and speak/write accordingly.e.Investigating learner language 71 (see also Granger. lexico-grammaticality. For example.e. 2005: 14) Mukherjee advocates a corpus-approximation to the native speaker norm and argues that corpus data can be used to describe this norm by ‘generalizing and abstracting from a vast amount of representative performance data’ (ibid: 15). that the existence of a variety of norms is recognized in learner corpus research (see Granger. Mukherjee. 2001. Another criticism of L1/L2 comparisons is directed at the idea of the ‘native speaker’ as the target norm (e. (Mukherjee. Piller. Gilquin and Granger (2008) compared the Tswana component of the second edition of the International Corpus of Learner English to a corpus of South African English editorials.g. (3) whose usage is largely idiomatic in terms of linguistic routines commonly used in a given speech community. Although they do not dwell on it. He proposes a usage-based definition of the native speaker based on three aspects that he regards as central to native-like performance. this speaker is thus taken to exemplify the abstract native speaker model on grounds of his/her language use. 2001: 401). the corpus-approximation to the native speaker norm is based on British and American English corpora. 2005). Tan. acceptability and idiomaticity (see Pawley and Syder. a ‘useful myth’ in Davies’s (2003) terms. . It should be noted. In this book. however. however. argues that ‘nativeness’ remains a useful construct both for linguistics and for the ELT community. If we refer to an individual speaker as a native speaker. 1983): The term ‘native speaker’ should be used for an abstraction of all language users (1) who have good intuitions about what is lexicogrammatically possible in a given language and speak/write accordingly. Lakshmanan and Selinker (2001) address the issue of the comparative fallacy and warn against the danger of ‘judging language learner speech utterances as ungrammatical from the standpoint of the target grammar without first having compared the relevant interlanguage utterances with the related speech utterances in adult native-speaker spoken discourse’ (Lakshmanan and Selinker. corpus data) and a warning against hasty conclusions based on a single researcher’s intuitions. The control corpora used in this study are described below. i.

It is. on the type of material that is best suited to serve as a control for a learner corpus. On the one hand. It has been suggested that the ICLE might be compared with ‘a corpus of newspaper editorials. 2000). native students have been shown to produce more dangling participles than EFL learners (Granger. ‘native-speaking students do not necessarily provide models that everyone would want to imitate’ (Leech. however.3. As Leech puts it. however. 1998: xix). Native student writing is arguably a better source of comparable data to EFL learner writing if the aim of the comparison is to describe and evaluate interlanguage(s) as fairly as possible. 2006: 206–7) The International Corpus of Learner English. footnote 10). 1999a: 14) and taking a stand against the ‘unrealistic standard of “expert writer” models’ (Hyland and Milton. (Ädel. In this study.72 Academic Vocabulary in Learner Writing 3. In a number of . arguing that it is ‘both unfair and descriptively inadequate’ (Lorenz. As Ädel put it. On the other hand. we need to use native-speaker writing that is also produced by students for comparison. 1998a: 18. There is no general agreement. The question of the norm is best addressed by considering the aim of the comparison. a text type which combines the advantages of being argumentative in nature and written by professionals’ (Granger. it can also be argued that professional writing represents the norm that advanced foreign learner writers try to reach and their teachers try to promote. 1997) and different types of spelling mistakes (Cutting. consists of argumentative texts and ‘argumentative essay writing has no exact equivalent in professional writing’ (De Cock. it can be argued that in order to evaluate foreign learner writing by students justly. Professional writing has a major role to play in learner corpus research if instruction and pedagogical applications are the goals of the comparison between learner and native-speaker productions. however. For example. a useful corpus for comparison is one which offers a collection of what Bazerman (1994: 131) calls ‘expert performances’. Several researchers have criticized the use of professional writing in learner corpus research. In this respect. learner writing was compared to expert academic prose. A comparison of learner vs. doubtful whether the findings from such comparisons could make their way into the classroom. expert writing Carrying out L1/L2 comparisons implies choosing an L1 corpus to be used as some kind of ‘norm’ with which the learner corpus data can be compared. 2003: 196). 1997: 184).

or technical writing. time and medium. 2008). having chosen to sample such things as popular novels. etc. letters. The British National Corpus (BNC) was created to be a balanced reference corpus of late twentieth century British English.. and medium refers to the type of publication (books. if we want to analyse texts produced by scholars specializing in natural sciences. as well as linguistic information (morphosyntactic tags. Neff and her colleagues used both native-speaker student writing and newspaper editorials as control corpora (Neff et al. 2004a. newspaper articles. General corpora have also been used in learner corpus research. by selecting instances of those types which have a wide distribution.). lemmas. Nesselhauf (2005). and developed a new resource called the BNC Index which contains genre labels for all BNC texts. Table 3. 1998: 28) The BNC mark-up conforms to the Text Encoding Initiative (TEI) recommendations (Burnard.). Thus. we can select all BNC texts classified under ‘W_ac_nat_science’. etc. Three criteria were originally used to select written texts to design a balanced corpus: domain.g. if we are not interested in discipline-specific differences and want to examine . (Aston and Burnard. time refers to the period when the text was written. The written component totals about 90 million words and includes samples of academic books. Domain refers to the subject field of the texts. newspapers. 2007). best-seller lists and library circulation statistics were consulted to select particular examples of them. The text selection procedure has been described as follows: In selecting texts for inclusion in the corpus. Lee (2001) criticized the domain categories for being overly broad and not sufficiently explicit. periodicals. by sampling a wide variety of distinct types of material. made use of the written part of the British National Corpus to determine the degree of acceptability of verb-noun combinations that had been extracted from the German subset of ICLE. and reception. file description.Investigating learner language 73 studies based on ICLE texts produced by Spanish EFL learners. 2004b.2 gives the breakdown of the genre categories in the BNC written corpus and shows that genre labels are often hierarchically nested. popular fiction. university essays and many other text types. Mark-ups include rich metadata on a variety of structural properties of texts (e. account was taken of both production. The BNC contains both written and spoken material. headings. Neff van Aertselaer. sentences and paragraphs). On the other hand. text profile. for example. Thus.

9% 5.3% 0.528.592 686.111.6% Letters 0.0% 4.3% 0.1% 0.867 1.451 15.892 52.632 87.3% 0.8% Non-academic prose 19.2 BNC written Academic Vocabulary in Learner Writing BNC Index – Breakdown of written BNC genres (Lee 2001) No.3% 0.480 66.413 3.3% Fiction 18.261 436.133 3.530 65.346 4.1% 4.5% 0.6% 4.364 % ‘Super genre’ No.391 1.895 297.045 146.156 351.1% 1.004 219.895 101.677 1.508.564 3.8% 0.366 213.8% 4.5% 0.9% 0.843 728.247.742 1.140.4% 0.757 222.5% 3.840 4.737 239.421.1% 10.2% News 7.156.5% 0.3% 100% Academic prose 17.946 558.031 9.6% 2.2% 0.933 1.1% 0.2% 0.144 W_ac_humanities_arts W_ac_medicine W_ac_nat_science W_ac_polit_law_edu W_ac_soc_science W_ac_tech_engin W_admin W_advert W_biography W_commerce W_email W_essay_sch W_essay_univ W_fict_drama W_fict_poetry W_fict_prose W_hansard W_institut_doc W_instructional W_letters_personal W_letters_prof W_misc W_news_script W_news_brdsht_nat_arts W_news_brdsht_nat_commerce W_news_brdsht_nat_editorial W_news_brdsht_nat_misc W_news_brdsht_nat_reportage W_news_brdsht_nat_science W_news_brdsht_nat_social W_news_brdsht_nat_sports W_news_other_arts W_news_other_commerce W_news_other_report W_news_other_science W_news_other_social W_news_other_sports W_news_tabloid W_non_ac_humanities_arts W_non_ac_medicine W_non_ac_nat_science W_non_ac_polit_law_edu W_non_ac_soc_science W_non_ac_tech_engin W_pop_lore W_religion TOTAL 3.865 498.6% 1.396 2.8% 0.831 4.171 546.1% .2% 4.7% Unpublished essays 0.477.5% 1.717.8% 1.8% 1.355 65.751.3% 0.143.74 Table 0.679 2.1% 0.1% 1.829 1.284.3% 1.759. of words 3. of files 87 24 43 186 138 23 12 60 100 112 7 7 4 2 30 432 4 43 15 6 11 500 32 51 44 12 95 49 29 36 24 15 17 39 23 37 9 6 111 17 62 93 128 123 211 35 3.256 4.5% 1.1% 0.3% 0.024 1.293 81.796 7.292.943 663.376.3% 5.811 424.640.957 1.2% 0.649 1.2% 0.1% 0.388 45.444 54.926.3% 18.4% 8.1% 0.258 415.

student essays rarely total more than 1. British literature since 1945. Soviet relations with Latin America.947 words and includes a wide variety of spoken registers.. ICLE texts were produced by students of humanities. The British National Corpus was accessed via the BNCweb (CQP-edition) interface developed by Stefan Evert and Sebastian Hoffmann. 2000) and the Corpus Query Processor (CQP). among others. interviews and lectures. Nietzsche on tragedy. This web interface is the result of a ‘marriage of two corpus tools’ (Hoffmann and Evert. Second. Table 3. the topics in BNC-AC-HUM differ from those in ICLE (described in Section 3. topics in BNC-AC-HUM appear only once.000 words). There are. The BNC spoken corpus (BNC-SP) consists of 10. i. The spoken part of the British National Corpus was also regularly consulted to check whether words and word sequences that were found in learner writing are more typical of speech or academic writing. Unlike the ICLE. we can select all texts whose categorizing labels begin with ‘W_ac’. a web-based client developed at the University of Zurich which allows users to access the BNC by means of a Web browser (see Lehmann et al. What is this thing called science?.000 to 45. They also have the advantage of corresponding to the type of writing that learners will try to produce during their university studies. a central component of IMS Open Corpus Workbench. This sub-corpus was chosen for two main reasons. First. Europe in the central middle ages. the BNCweb.e.2) were sometimes used to compare the frequency of words and phrases across ‘super genres’. however. texts in the BNC-AC-HUM are arguably quite close to the type of text these students might have come across in their first few years at university.000 words while samples in the BNCAC-HUM are much longer (from 25. 2004a).1 above).Investigating learner language 75 texts produced by professional writers in higher education settings. major differences between the two corpora. henceforth BNC-AC-HUM) totals 3. It was used as the comparison corpus to ICLE in this study.. China’s students. etc. 2006). The BNC sub-corpus of academic prose in humanities and arts (W_ac_humanities_arts. which ‘allows sophisticated searches both for individual words (which can be matched against regular . The W_fic and W_news sub-corpora (cf. broadcast documentaries and news.321.334. First. ICLE is a corpus of unpublished university student essays while BNCAC-HUM consists of samples of published articles and books.4 Third.867 words. National liberation. Interpreting the results in the light of genre analysis thus required special care: differences between student essays and expert writing may simply reflect differences in their communicative goals and settings (Neff et al. The morality of freedom. They include The people’s peace.

inclusion of lemma and part-of-speech information.01. e. log-likelihood and log-log measures rank co-occurrences in very different ways (Evert. Significant co-occurrents are sorted by decreasing log-likelihood values (right column). use of a threshold value of 15. (2004).’s (2004) suggestion. McEnery et al. maximum window span. Association measures are the most widely used method of distinguishing between casual and significant co-occurrences. log-log and MI3 tests appear to provide more realistic collocation information’ (McEnery et al. which picks out significant co-occurrents of the search word on the basis of a number of measures of association. log-likelihood and log-log measures. It is a marriage between the efficiency and flexibility of CQP queries. The log-likelihood test was therefore used to study the phraseology of academic words in expert and learner writing. focused on the comparison of word frequencies between corpora and suggested that. which indicates the strength of the association relative to that expected by chance. however.76 Academic Vocabulary in Learner Writing expressions) and for lexico-grammatical patterns (using linear grammars that have access to all levels of annotation)’ (Hoffmann and Evert. One tool that is particularly useful is Collocations. 2006: 180). minimum frequency of the co-occurrent. (2006) compared the various statistical measures provided by BNCweb and reported that ‘MI and z-scores tend to put too much emphasis on infrequent words. MI3. z-score. Co-occurrence frequencies can be quite low and I therefore followed Rayson et al. MI3. 2004).g.13 is preferred at p < 0. etc. . A number of other settings are customizable. They compute an association score for each pair of words extracted from a corpus. Co-occurrences were analysed in windows of one to three words to both the left (3L-1L) and the right (1R-3R). the log-likelihood. Users of the BNCweb can decide to use any of five different measures: mutual information. Figure 3.. (2008) proved further information on the British National Corpus and the BNCweb interface. 2006: 220). The CQP edition of the BNCweb combines the strengths of both software packages while overcoming their limitations. in order to extend applicability of the frequency comparisons to low expected values. minimum frequency of the co-occurrence. Mutual information.5 They can also sort co-occurrents by decreasing frequency. 1998: 176). z-score. Rayson et al. and the user-friendliness of BNCweb with its wide range of query options and display facilities. The frequency of the co-occurrence is given together with the number of texts in which it appears. Hoffmann et al. The log-likelihood scores can be directly compared with critical values of a chi-square distribution table (see Oakes. In contrast.3 displays a collocation query result.

Investigating learner language Figure 3.3 BNCweb Collocations option 77 .

Word pairs in the ICLE sub-corpora were classified into three groups according to their co-occurrence status in professional academic writing: – word pairs that are statistically significant co-occurrents in the academic sub-corpus of the BNC (BNC-AC). In his study of the statistics of word co-occurrences. – word pairs that do not appear in the BNC-AC. Word pairs that did not appear in the BNC-AC were presented to a native speaker of English for acceptability judgments. 2004) or frequency-based approach (Nesselhauf.78 Academic Vocabulary in Learner Writing Log-likelihood measures are strongly dependent on corpus size and word frequencies.2. The ICLE sub-corpora are in fact too small for a statistical analysis of co-occurrences to be meaningful. I found that learners’ word pairs were sometimes not statistically significant in BNC-AC-HUM just because the co-occurrence was not frequent enough. the hapax and dis-legomena. – word pairs that appear in the BNC-AC but are not statistically signifi cant co-occurrents. I therefore decided to use the larger BNC-AC instead of the BNC-AC-HUM to judge the acceptability and typicality of EFL learners’ phraseological sequences (see Section 5. . Co-occurrence statistics are therefore not comparable across corpora of different sizes such as the British National Corpus and the International Corpus of Learner English. Evert argued that ‘data with co-occurrence frequency f < 3. the co-occurrence proved significant. 2004) was adopted to examine the phraseology of academic words in learner writing. i. Academic words are not high-frequency words such as make. 2004: 133) as expected frequencies and p values for low frequency words are distorted in unpredictable ways. Summary and conclusion This chapter has described the data and methodology used to investigate the use of academic vocabulary in writing by EFL learners.e.3). should always be excluded from the statistical analysis’ (Evert. As soon as more data was used. A distributional (Evert.4. In a pilot study. do and take and co-occurrences often appear less than three times. Special care has been taken to select a set of learner essays from the International Corpus of Learner English that is as homogeneous as possible and to control for a number of variables that have been found to influence such writing. 3.

Investigating learner language


The learner corpus can be compared to the humanities and arts academic sub-corpus of the British National Corpus to identify learner-specific features of the use of academic vocabulary. The BNC spoken corpus can also sometimes be useful to check whether specific words and phrases that appear in the learner sub-corpora are more typical of speech or writing. The method used to investigate learners’ use of academic vocabulary is based on Contrastive Interlanguage Analysis (CIA) and combines comparisons of learner and native-speaker writing, and comparisons of different learner interlanguages. CIA is very popular among researchers in the field of learner corpus research and has helped to highlight an unprecedented number of features that characterize learner interlanguages. To date, however, most studies have used the technique only to compare a learner corpus and a native reference corpus, rather than to explore different learner corpora in the same target language. The studies that have compared more than one interlanguage have usually focused on learners from one mother-tongue background, and used data from one or two other learner populations only to check whether the features they have identified are L1-specific (and thus possibly transfer-related) or are shared by other learners. L2/L2 comparisons involving many different first languages are, however, indispensable if we want to identify the distinguishing features of learner language at a given stage of development (Bartning, 1997). In the following chapters, I try to make the most of CIA by comparing academic vocabulary in ten learner corpora representing different mother-tongue backgrounds.

This page intentionally left blank

Chapter 4

Rhetorical functions in expert academic writing

This chapter deals with academic vocabulary that serves specific rhetorical and organizational functions in expert academic writing. Section 4.1 focuses on the Academic Keyword List and shows that a high proportion of AKL words can fulfil these functions in academic prose. It lists several steps which are necessary to turn the AKL into a tool that can be used for curriculum and materials design (most notably a phraseological analysis of AKL words). Section 4.2 presents a detailed analysis of exemplificatory devices in academic writing. This serves as an illustration of the type of data and results obtained when the whole range of lexical strategies available to expert writers to organize scientific discourse are examined. For lack of space it is impossible to describe in similar detail all the functions that were analysed in the BNC-AC-HUM so as to provide a basis for comparison to EFL learner writing. Section 4.3 briefly comments on the types of lexical devices used by expert writers to serve four additional functions: ‘expressing cause and effect’, ‘comparing and contrasting’, ‘expressing a concession’ and ‘reformulating: paraphrasing and clarifying’ and aims to characterize the phraseology of rhetorical functions in academic prose.

4.1. The Academic Keyword List and rhetorical functions
The functional syllabus has a long tradition in English language teaching (see Wilkins, 1976; Weissberg and Buker, 1978). Jordan (1997: 165) reports that most of the textbooks that were published in Britain in the 1980s and 1990s that followed a product approach to academic writing were organized according to language functions such as explanation, definition, exemplification, classification, cause and effect, and comparison and contrast (e.g. Jordan, 1999). However, they were rarely based on principled


Academic Vocabulary in Learner Writing

selection criteria, relying instead on the writers’ perceptions of good practice in academic writing. Unlike textbooks adopting a functional approach, courses which use vocabulary as the unit of progression, introduce new words according to principles such as frequency and range of occurrence. Nation explains that such courses generally combine a “series” and a “field” approach to selection and sequencing. In a series approach, the items in a course are ordered according to a principle such as frequency of occurrence, complexity or communicative need. In a field approach, a group of items is chosen and the course covers them in any order that is convenient, eventually checking that all the items are adequately covered. Courses which use vocabulary as the unit of progression tend to break vocabulary lists into manageable fields, (. . . ), according to frequency, which are then covered in an opportunistic way. (Nation, 2001: 386) Most pedagogical applications of Coxhead’s (2000) Academic Word List to date have adopted this particular approach, using the frequency-based AWL sub-lists as fields (e.g. Obenda, 2004; Huntley, 2006). There is a need for teaching materials that merge the two types of syllabus design, thus adopting a ‘functional-product’ approach (Jordan, 1997: 165) to academic writing while introducing new vocabulary according to principled criteria such as frequency and range of occurrence. This is precisely where the Academic Keyword List has a role to play. As explained in Section 2.4, the Academic Keyword List requires pedagogic mediation: it is a platform which can inform a functional syllabus for academic writing, but it needs to be organized. As argued by Martinez et al. (2009: 193), ‘a list based on semantic and pragmatic criteria would perhaps be more useful than lists built solely on frequency criteria.’ Sinclair, however, warns us that ‘there is no assumption that meaning attaches only to the word’ (Sinclair, 2004b: 160). Similarly, Siepmann (2005: 86) comments that ‘neat compartmentalizing of meanings or functions can do no more than partially capture a complex reality’ in which any word or multi-word sequence may express more than one discourse relation. This being said, the results of the automatic semantic analysis of the Academic Keyword List revealed that a significant proportion of AKL words fall into semantic categories that correspond to the rhetorical functions typical of scientific discourse, e.g. A2.2. Affect: Cause – connected, A4. Classification, A5. Evaluation, A6. Comparing, Q2.2. Speech acts (see Section 2.4). A close

Rhetorical functions in expert academic writing


examination of the words classified into these semantic categories made it possible to identify twelve rhetorical functions that dovetail with the functions typically treated in EAP textbooks adopting a functional approach to academic writing: 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. Exemplification, e.g. example, for example, illustrate Cause and effect, e.g. cause, consequence, result Comparison and contrast, e.g. contrast, difference, same Concession, e.g. although, despite, however Adding information, e.g. first, further, in addition to Expressing personal opinion, e.g. appropriate, essential, major Expressing possibility and certainty, e.g. likely, possibility, unlikely Introducing topics and ideas, e.g. discuss, examine, subject Listing items, e.g. first, second, third Reformulating – paraphrasing and clarifying, e.g. namely Quoting and reporting, e.g. define, report, suggest Summarizing and drawing conclusions, e.g. conclude, conclusion, summary.

The next step is to analyse all words that may serve one of these twelve functions in context, with special emphasis on their phraseology. Multiword sequences have been shown to provide ‘basic building blocks for constructing spoken and written discourse’ (Biber and Conrad, 1999: 185) and to correlate closely with the complex communicative demands of a particular genre, thus contributing to its lexical profile (Biber et al, 1999; 2004; Luzón Marco, 2000). A phraseological analysis will also make it possible to investigate how academic vocabulary contributes to this ‘shared scientific voice or “phraseological accent” which leads much technical writing to polarise around a number of stock phrases’ (Gledhill, 2000: 204). It will examine phrasemes, i.e. syntagmatic relations between at least two lemmas, contiguous or not, written separately or together, which are typically syntactically closely related and constitute ‘“preferred” ways of saying things’ (Altenberg, 1998: 122). This is because such phrasemes: – form a functional (referential, textual or communicative) unit (e.g. Burger, 1998); and – display arbitrary lexical restrictions (e.g. Mel’cuk, 1998); ˇ

In this chapter and the next. They include speech act formulae. Nor is it reasonable to focus on functions such as ‘reporting and quoting’ and ‘expressing personal opinion’. The phraseological analysis used here is based on Granger and Paquot’s (2008a) classification of phraseological units into three main categories: referential phrasemes. cause and effect. similes. the . they include grammaticalized sequences such as complex prepositions and complex conjunctions. proverbs and slogans. I focus on the vocabulary of five rhetorical functions — exemplification. The use of academic words is compared in the BNC-AC-HUM corpus (a corpus of book samples and journal articles written by experts in the fields of arts and humanities) and the International Corpus of Learner English (a corpus of short argumentative essays produced by EFL learners of English). either to focus their attention. They include lexical and grammatical collocations. attitudinal formulae. Textual phrasemes are typically used to structure and organize the content (i. include them as discourse participants or influence them. Unlike experts writing in their field. referential information) of a text or any type of discourse. irreversible bi. Communicative phrasemes are used to express feelings or beliefs towards a propositional content or to explicitly address interlocutors. idioms. phenomena or real-life facts.g. textual phrasemes and communicative phrasemes. and reformulating — with occasional forays into other functions. concession. Referential phrasemes are used to convey a content message: they refer to objects. linking adverbials and textual sentence stems. Apart from being essential rhetorical functions in academic prose. – display a certain degree of syntactic fixity. these functions should be among the least sensitive to the text type differences discussed in Section 3.e. Barkema. commonplaces.and trinomials. comparison and contrast.3. – display arbitrary restrictions on the word forms that can be used to instantiate at least one of the lemmas involved.84 Academic Vocabulary in Learner Writing and/or – are characterized by a certain degree of semantic non-compositionality (e. As the BNC includes truncated texts. This function is typically localized in the last paragraphs of a piece of academic writing and might thus be absent from a number of BNC texts. compounds and phrasal verbs. 1996). it would not be reasonable to quantitatively compare the words that are used to serve the function of ‘summarizing and drawing conclusions’.

It is only by examining their frequency and patterns of use in expert and learner corpora that I shall be able to assess whether these words and phrasemes should be part of an academic vocabulary and added to the AKL. Conrad. the AKL includes a number of words and phrasemes that are commonly used as exemplifiers: the wordlike units for example and for instance. have a very restricted meaning or prove particularly difficult for learners. 1998b. the adverb notably and the abbreviation e. however. 2008: 4). What is your opinion?’ and ‘In the 19th century. may not be the sole criterion to include lexical items in the curriculum (see Section 1. the verbs illustrate and exemplify. They have often found considerable mismatches between naturally-occurring language and the type of language that is presented as a model in teaching materials (Römer. 2006. dominated by science. they were explicitly encouraged to give their personal opinions (topics for the essays include: ‘Some people say that in our modern world.1. 2004a. there is no longer a place for dreaming and imagination. the noun example. I therefore consulted several EAP textbooks (Harris Leonhard. the nouns illustration and a case in point and the preposition like. 2004. Jordan. For example. Some of these non-AKL words may be used in very specific lexico-grammatical environments. 2004b. 2003. technology and industrialism.1). Other lexical items listed in EAP materials but not found in the AKL are the expressions by way of illustration and to name but a few. 2002). By contrast. 2002. Carter. 2005. The textbooks-derived list appeared to be very different from the words found in the Academic Keyword List. the preposition such as. 1999.g. Victor Hugo said: “How sad it is to think that nature is calling out but humanity refuses to pay heed. Lonon Blanton. Frequency. it is not quite clear why these items are taught to novice writers and EFL learners while other much more frequent lexical items that are used to express the same rhetorial functions are missing. First. Several researchers in applied corpus linguistics have examined language features in general reference corpora and compared the distributions and patterns found in actual language use with the presentations of the same features in teaching materials such as textbooks or grammars (e. Römer.” Do you think it is still true nowadays?’) (see Section 3.Rhetorical functions in expert academic writing 85 learners who produced the argumentative essays were not supposed to show that they were familiar with the subject by referring to or quoting from the literature. I decided to include the lexical items found in EAP textbooks in my study of academic vocabulary for two main reasons. Zemach and Rumisek. 2001. Oshima and Hogue. . and listed all the lexical items that are commonly taught to serve rhetorical functions.g.1). Zwier. Ruetten. 2005).

2002: 30) This approach should help us move towards ‘understanding the intersection of form and function’ (Swales. always. but also how often it could have occurred. and simply reflect the fact that a word is more/less frequent in learner writing. so that the frequency of each exemplificatory lexical item can be calculated as a proportion of the total number of exemplifiers. maybe. The resulting list was analysed to identify words that might serve one of the 12 rhetorical functions listed above. As Wray stated in her book on formulaic language. . Learner-specific word sequences are discussed in Section 5.3. especially if conceptual frequency is to be investigated. we need to know not only how often that form can be found in the sample. their inclusion in the description of a specific function in academic writing allows us to approximate as closely as possible what Hoffmann (2004: 190) referred to as ‘conceptual frequency’. 2003 for a keyword analysis of the French sub-corpus of ICLE). sure. I. These two terms are neutral. For example. EFL learners may use different lexical devices than native writers to serve rhetorical functions. which are quite rare in academic prose. In learner corpus research. really.3. let. To capture the extent to which a word string is the preferred way of expressing a given idea (for this is at the heart of how prefabrication is claimed to affect the selection of a message form).86 Academic Vocabulary in Learner Writing Second. which has limitations for an analysis of learner writing. thirdly. In other words. The Academic Keyword List is based on native corpora only.1 for the automatic extraction of potential academic words was therefore adopted to identify words and word sequences that EFL learners frequently use. but. but which are not favoured by expert academic writers. say. in brief and all in all for summarizing and concluding. Examples of overused words which do not belong to the AKL but are employed to serve rhetorical functions in learner writing include like. thanks. 2002: 163) in academic prose.2. A keyword procedure such as that described in Section 2. we need a way to calculate the occurrences of a particular message form as a proportion of the total number of attempts to express that message. thing. say. positive keywords are often referred to as overused words and negative keywords are said to be underused. they repeatedly use word-like units such as in a nutshell. opinion. The ICLE corpus was compared to the BNC-AC-HUM to extract distinctive words in the learner corpus. secondly. so and why (see De Cock. firstly. (Wray.

similarly. though (conj. parallel. thus. differently. like. The words in italics are not part of the Academic Keyword List. parallelism. Exemplification: example. on account of. provoke.. because of. bring about. so. that is to say. responsible. *in contrary to. prompt. or more accurately. in spite of. in the same way. versus. as against. exemplify. stem from.). as. comparatively. that is. a case in point. on the grounds that. likewise. owing to. yet. because. like. therefore. implication. compared with/to.). consequent. contrasting. notwithstanding Reformulating – paraphrasing and clarifying: i. differentiation. by/in contrast. reason. by way of illustration Comparison and contrast: analogy. differentiate. difference. make sb/sth do sth. accordingly. contrast. distinct. or rather . result. comparable. emerge. by/in comparison with. in other words. look like. thanks to. as a result of. in parallel with. for instance. as a result. root. illustrate. analogous. contrastingly. reversely. They are included in the corpus-based analyses presented in this chapter and in Chapter 5 to assess the adequacy of the treatment of rhetorical functions in EAP textbooks and investigate whether the AKL should be supplemented with additional academic words. cause (v. resemble.g. factor. contribute to. contrary to. result in/from. origin. distinguishable. while. due to. similar. correspond. yield. unlike. parallel. whereas. distinctively. *on the opposite. namely. by way of contrast.. consequence. for. arise from/out of. (the) opposite. compare. although. as opposed to. derive. *on the other side.e. give rise to. even if. in contrast to/with. (the) contrary. as … as. different. common. analogously. generate. in view of. on the other hand. contrariwise. reverse. despite. to name but a few. lead. nevertheless. trigger. on the grounds that. conversely. as. follow from. They were identified on the basis of a close examination of EAP materials and a keyword analysis of learner corpora. opposite. parallely. in consequence of. (the) reverse. identically. such as. distinguish. effect. same. consequently. though (adv. in (the) light of. distinctiveness.. correspondingly. induce. notably. even though. parallel. unlike. by/in comparison. similarity. in the same way as/that Cause and effect: cause (n. since. so that. source.). differing.). distinctive. identical. hence. distinction. for example. thereby. in consequence. resemblance. viz. differ. comparison. e. is why Concession: however. illustration. albeit. (the) same.Rhetorical functions in expert academic writing 87 The final lists of words that may be used to serve one of the five selected rhetorical functions are given below. nonetheless. contrast. on the contrary. by implication. alike. as a consequence of. as a consequence. or more precisely. contrary. outcome.

e. In Figure 4. This section deals with the latter and focuses on the lexical items used by expert writers to give an example. and then focus on the exemplificatory use of nouns and verbs. the lexical items are ordered by decreasing relative frequency in the academic corpus. the noun example. illustrate. The most frequent exemplifiers in professional academic writing are the mono-lexemic phrasemes such as and for example. plus the noun example.g. 2005: 111). see Siepmann (2005: 112–18). and the wording of the example (i.000 words) while the adverbials to name but a few and by way of illustration as well as the noun case in point appear very rarely in the BNC-AC-HUM.1.2. I will now discuss my main findings on the exemplificatory functions of prepositions. the verbs illustrate and exemplify.88 Academic Vocabulary in Learner Writing 4. the adverb notably and the abbreviation e. the preposition such as.g. are the expressions by way of illustration and to name but a few.000 words. The verb exemplify and the noun illustration are less frequent (around 2.3 occurrences per 100. the choice of exemplifiers). that as an object of study. Table 4. Siepmann (2005) showed that exemplificatory discourse markers occur in all kinds of discursive prose.e. . like. when and why are examples introduced into a text). For a rhetorical perspective on exemplifiers in native writing. the nouns illustration and a case in point and the preposition like. but not found in the AKL. adverbs and adverbial phrases. subordinate to major discoursal stratagems such as “inferring” and “proving”’ (Siepmann. Coltier (1988) remarked that examples and exemplification merit close investigation at two levels: the exemplificatory strategies adopted (i. The Academic Keyword List (AKL) includes a number of words and multiword sequences that are commonly used as exemplificatory discourse markers: the mono-lexemic or word-like units for example and for instance.000 words and the percentage of exemplificatory discourse markers they represent. The function of exemplication This section presents a detailed analysis of the academic words that are used by expert writers to serve the rhetorical function of exemplification. He argued. Almost half of the exemplifiers — for instance. and notably — occur with a relative frequency of between 5 and 20 occurrences per 100.e. and are particularly frequent in humanities texts.000 words. Other lexical items commonly listed in textbooks and EAP/EFL materials.1 gives the absolute frequencies of these words in the BNC-AC-HUM as well as their relative frequencies per 100. ‘exemplification continues to be the poor relation of other rhetorical devices’ and that ‘such neglect has led to a commonly held view in both the linguistic and the pedagogic literature that exemplification is a minor textual operation. however. which occur more than 35 times per 100.

2 38. Nouns example illustration (BE) a case in point TOTAL NOUNS Verbs illustrate exemplify TOTAL VERBS Prepositions such as like TOTAL PREP.3 0.7 7. notably to name but a few by way of illustration TOTAL ADVERBS TOTAL 1263 609 259 77 4 3 2215 5959 21.3 7.1 0.0 18.9 34.6 1.1 Ways of expressing exemplification found in the BNC-AC-HUM Abs. 89 50 45 40 35 30 25 20 15 10 5 0 pl e re xa m pl fo e ri ns ta nc e g.8 2.7 179.3 23.1 66.g.3 0.1 0.5 % Rel.0 61. su ex Figure 4.0 16. no ta bl y ex em pl ify illu BE st a ra ca tio se n to in na po m in e by t bu w ta ay fe of w illu st ra tio n lik e ra te as am ch illu st e.3 0.Rhetorical functions in expert academic writing Table 4.5 41.0 8.7 2.4 1. freq.3 1.1 Exemplification in the BNC-AC-HUM fo .2 4.4 10.2 10.3 0.0 259 79 338 4. freq. Adverbs for example for instance e.8 2.2 100 38.3 5.2 1285 77 18 1380 21.4 1494 532 2026 25.1 37.0 45.

4.3) as they are used to ‘supply additional information.e.4. pointing forward to the example) as shown in Examples 4. Unlike in other genres (such as speech and fiction). especially when executed by a showman like Salvador Dali. . to ensure the reader is able to recover the writer’s intended meaning’ (Hyland. to whom abominable practices were also ascribed. explaining. The term ‘buggery’.2. . who married the former wife of the poet Paul Éluard. Such associations of sexual deviance and political threat have a long history sedimented into our language and culture. ].2. or elaborating what has been said. so that the associations between faunal types and ecology are well documented [ . schools. .4. 4.3.1. while remaining essentially cataphoric in nature (i. i. by rephrasing. are distinct from grassland faunas. 4. the complex preposition such as is the most frequent exemplifier in the BNC-AC-HUM (see Example 4. derives from the religious as well as sexual nonconformity of an eleventh-century Bulgarian sect which practised the Manichaean heresy and refused to propagate the species. enclosed by commas. Surrealist painting had publicity value. for example and for instance are typically used within the sentence. In a phraseological approach to academic vocabulary.1. Code glosses are ‘interactive resources’ in Hyland’s typology of metadiscourse: they are features used to ‘organize propositional information in ways that a projected target audience is likely to find coherent and convincing’ (ibid.e. 4. These two adverbials are commonly classified as ‘code glosses’ in metadiscourse theory (see Section 1. Using prepositions. for example. 50). it is much more frequent than the preposition like in professional academic writing (Example 4. cultural organisations. and run by the separate catholic and protestant communities. multi-word units that are equivalent to single words and which fill only one grammatical slot. welfare services. they fall into the category of textual phrasemes as they are mono-lexemic multi-word units. with an organizational – exemplificatory – function. In the BNC-AC-HUM. for instance. But they can also follow the subject of the exemplifying sentence.1). and hospitals are duplicated.2).1. the OED tells us that it was later applied to other heretics. and when these and more precise distinctions are made it is possible to correlate and even define ecological zones by their small mammal faunas. especially after the subject. and tropical faunas distinct from temperate faunas. 2005: 52). The small mammals living today in many different habitats and climatic zones have been described. for example is twice as frequent as for instance.3 and 4. adverbs and adverbial phrases to exemplify As shown in Figure 4.90 Academic Vocabulary in Learner Writing 4. This is the arrangement in Holland whereby various institutions such as media. Similarly. Woodland faunas.

1) and see (f[n.2 The use of ‘for example’ and ‘for instance’ in the BNC-AC-HUM Cataphoric marker for example for instance 1. Such is the relation which Nicaragua bears to El Salvador. given that counterrevolutionary response to any successful formula will ensure that it will be that much more difficult to apply the same tactics in another situation.6. It is worth pausing here momentarily to observe that such legally provided remedies can be morally justified even when applied to people who are not subject to the authority of the government and its laws. log-likelihood = 19. however. mono-lexemic adverbial phrases can also have their own phraseological patterns. Three verbs. Enforcing such a duty against a person who refuses to pay damages is morally justified because it . the debates over how far to forge a strategy either for winning power or for promoting economic development in a post-revolutionary society have not been satisfactorily resolved. Laruelle (2004: 96–7) writes that for example should be placed in the initial position if the whole sentence has an exemplificatory function. Example 4. consider (f[n. log-likelihood = 71. In Mieux écrire en anglais. 4. between commas. and indeed perhaps cannot be.3 clearly shows that for example need not be placed in the initial position to introduce an exemplificatory sentence. c] = 19. Assuming that it is what it should be.5). log-likelihood = 92. is not confirmed by corpus data.8%) 588 (96. Like nouns and verbs.Rhetorical functions in expert academic writing Table 4. This use is. if only the subject is the example.5. i. much less frequent (see Table 4.5. The verbs consider and take are typically used with for example to introduce an example that is discussed in further detail over several sentences: 4. Consider for example the law of defamation. They are used in the second person of the imperative.7) are significant left co-occurrents of for example in the BNC-AC-HUM.2%) 21 (3.5%) Endophoric marker 78 (6. for example. Thirdly. as illustrated in Example 4. take (f[n.185 (93. This statement.e.2). The duty to compensate the defamed person is itself a moral duty. c] = 7. c]1 = 13. it does no more than incorporate into law a moral right existing independently of the law. while the adverbial should follow the subject. however.5%) 91 For example and for instance can also function as endophoric markers and refer back to an example given before.

But while I am looking at it my eyes constantly wander from one flower to the next. literary criticism. chemical engineering. ‘devices that explicitly address readers.7. where readers are initiated into a new domain of argument. (1998) examined a corpus of research articles in ten disciplines (art history. I may move the vase closer. Take. 2002a: 217). philosophy. But the concept of compresence is far from clear. figures. 2002a. 2005:53). 63 per cent of the occurrences of the sequence see for example appear between brackets as in Example 4. The scene will keep constantly changing. 4. I shall experience a succession of different “complexes of qualities” but I shall still be looking at the same bunch of flowers. 2005. If it implies that no time-lag is detectable between elements of an experienced “complex”. the perceptual experience that I have while looking at this bunch of carnations arranged in a vase on the table in the middle of the room. or other sections of the article or to someone else’s ideas or publications (Hyland. As a result. One need not invoke the authority of the law over the defamer to justify such action.e. or directed to understand a point in a certain way’ (2002a: 217). The law may not have authority over him. He categorizes them as interactional resources.92 Academic Vocabulary in Learner Writing implements the moral rights of the defamed. Hyland describes this type of imperative as directives with a rhetorical purpose that ‘can steer readers to certain cognitive acts. i. picking out the details of their shapes and colours. Hyland and Tse. then this is true only in a very limited sense. I see this “complex” as one whole. communication studies. led through a line of reasoning. Swales et al.8: 4. In the BNC-AC-HUM. experimental geology. history. The verb see is frequently used in professional academic writing as an endophoric marker to refer to tables. referring them to another part of the text or to another text’ (Hyland.. Finally. political science and statistics) and found that second person imperative see was the most . The use of the second person imperative see ‘allow[s] academic writers to guide readers to some textual act. ignoring others. 1998.8. pausing at some. or walk around the table and look at the flowers from different angles. without taking my eyes off the flowers. Afro-Caribbean and Asian children are indeed painfully aware that many teachers view them negatively and some studies have documented reports of routine racist remarks by teachers (see for example Wright in this volume). 2007). either to focus their attention or include them as discourse participants’ (Hyland. linguistics. and more specifically as engagement markers. for example.

the use of the imperative varied across disciplines.9) and is qualified by the adverb most in 15. and more specifically engagement markers.2 The distribution of the adverb ‘notably’ across genres . the miners’ strike. notably Jenkins of Durham. and Hapgood of York. in his study of directives in academic writing. Similarly. It is typically preceded by a comma (Example 4. appears quite clearly here.10). have spoken out about deprivation in the inner cities. per million words 4040 30 20 10 0 academic writing news fiction speech Figure 4. and understanding of. The adverb notably can be regarded as a typical academic word: Figure 4. Sheppard of Liverpool. seven textbook chapters and 64 project reports written by final year Hong Kong undergraduates and found that the second person imperative see represented 45 per cent of all imperatives in that corpus. The sequences take/consider for example consist of two metadiscourse resources in Hyland’s (2005) categorization scheme: the imperative forms take and consider are interactional resources. and the need for government to show a greater compassion for. while for example is a code gloss.9. In our phraseological framework.2 per cent of its occurrences in the BNC-AC-HUM (Example 4. the sequences take/ consider/see for example are textual phrasemes as they form functional — textual — units and display arbitrary lexical restrictions. 4. Similarly. see is an endophoric marker in see for example. 50 freq. the poor.Rhetorical functions in expert academic writing 93 frequent imperative form across disciplines. Hyland (2002a) analysed a corpus of 240 published research articles. and hence metadiscourse resources.2 shows that it is much more frequent in academic writing than in the other genres. Note that in both studies. The advantage of adopting a phraseological approach to rhetorical functions. Some bishops.

Figures 4. it is preceded by a comma: 4. flood protection) are clearly within the domain of a soil conservation policy. e.10. these expressions are very infrequent in all types of discourse.g. or the air. government ministers.g. is used without brackets. The abbreviation e. At leading public schools. When e.4 show the distribution of the two phrasemes in four main genres of the British 1 freq.13. Primary industries are those which produce things directly from the ground. and prime ministers. As shown in Figure 4. Direct curative measures (e.3 and 4.12. the water. introduce one or more noun phrases rather than full clauses: 4. there is a tradition of providing MPs.3 The distribution of ‘by way of illustration’ across genres ac . farming.g. It is quite common in the BNC-AC-HUM. It may help to refer the patient to other agencies (e.g.g. In contrast to for example and for instance. the great majority of occurrences of e.11. (or less frequently eg) stands for the Latin ‘exempli gratia’ and means the same as for example. in which 65.1. social services.g.94 Academic Vocabulary in Learner Writing 4.7 per cent of its occurrences are between brackets: 4. self-help groups). In fact. the textual phrasemes by way of illustration and to name but a few are quite rare in the BNC-AC-HUM. per million words 0. most notably Eton.5 0 tio n s ic ne w ad em fic sp ee c h Figure 4. a psychosexual problems clinic.

which is much more common than illustration or a case in point. per million words 95 0. the textual phraseme for example. The verb be is the most frequent verb co-occurrent of example in windows of one to three words to both the left Table 4. in the BNC-AC-HUM. and speech.e. Using nouns and verbs to exemplify Nouns and verbs are used to give examples in specific phraseological patterns.4. and only one occurrence comes from speech. newspaper texts. 10 out of a total of 28) appear in academic texts.2. No instance of to name but a few was found in speech.43 for example Absolute freq. BNC-AC-HUM 1285 % 50.5 0 w s ic em tio ne ad fic sp ee ch n Figure 4. but only 12. Some 36 per cent of the occurrences of by way of illustration in the BNC (i.3 shows that it is as frequently used as its connective counterpart. fiction. 1263 % 49.2. The significant verb co-occurrents of the noun example in the BNCAC-HUM are listed in Table 4. namely academic writing.8 per cent of its occurrences (10 out of 78) appear in academic writing.3 The use of ‘example’ and ‘for example’ in the BNC-AC-HUM example Absolute freq. The noun which is most frequently used in this way is example. 4. The expression to name but a few is more frequent than by way of illustration in the whole BNC.57 ac .4 The distribution of ‘to name but a few’ across genres National Corpus (BNC). Table 4.Rhetorical functions in expert academic writing 1 freq.

4.16. of course. twice as frequent in the left window. Here is a simple example of the complexity at issue. it functions as an advance label which refers forward to a following example (underlined): 4.96 Academic Vocabulary in Learner Writing Table 4. used by fast-moving traffic. 84 14 21 15 12 6 7 8 6 16 15 13 (3L-1L) and the right (1R-3R). In addition.4 Significant verb co-occurrents of the noun ‘example’ in the BNC-ACHUM Left co-occurrents Verb be provide take give cite consider illustrate show see serve freq. 139 26 29 12 5 12 7 9 10 5 Right co-occurrents Verb be illustrate show give suggest quote include provide concern will can would freq.15. as Polygnotos showed the sack of Troy in its aftermath. . A car some distance ahead strikes a large dog but does not stop.14. with my children in the back seat. I am driving along a narrow main road. By contrast. This is the supreme surviving example of the early classical taste for stillness and indirect narrative. Be is. 4. Vision is a better example of a modular processing system.e.14) or to the demonstrative pronoun this which further points to a previous exemplifying sentence (Example 4. When example is preceded by the verb be. leaving the creature walking-wounded but in obvious distress. choices can result from lengthy weighing of odds. when the noun example is introduced by there + BE (11%) or here + BE (15%). it refers back to the exemplifying element which is given as the subject. The designer at Olympia chose to represent the race by the moment before it started. however. i.15). The noun example may refer back directly to a noun phrase (Example 4. it mainly functions as a retrospective label.

Copular clauses using the noun example consist of textual sentence stems (An example of Y is . which is entirely concerned with the evocation of feelings or attitudes. slowing down but then speeding up again.18.Rhetorical functions in expert academic writing 97 My children. . consider. . give. the exemplified item is the pronoun this which refers back to the previous sentences.17. I do not stop. seeing what occurred.g.4).19.19) and the exemplified item is generally introduced by the preposition of (Examples 4. consider. In Example 4. serve and include often co-occur with the noun example to form textual — exemplificatory — phrasemes. 4. 4. It is often qualified by an adjective (see Examples 4. while the verbs suggest. The verbs take. . The most striking example of this is perhaps the frontier in the Danubian plain between the Ottoman empire and the Habsburg territories in central Europe. quote and include and the modals will. 4.17 to 4. see. is an example of Y). can and would are significant right co-occurrents. even most. cry out. They consist of sequences of two or more clause constituents. namely provide. it always functions as an advance label. which do not contain any thematic element (e. cite. is another issue). give. European frontiers were very vague. zones in which the claims and jurisdictions of different rulers and their subjects overlapped and intersected in a complex and confusing way. . where many states were large and central governments were usually less effective at the peripheries of their territories than in the west. They typically have an empty slot for the following object or complement. The verbs provide. An example of Y is . see and serve are only significant left co-occurrents. or should not be. . The prime example is the Dada movement. . whose nihilistic work is now admired for its qualities of imagination. cite.19). When example is the subject of the verb be. illustrate and show (given in italics in Table 4. . e. and in which the writer’s and reader’s attention is not.18 and 4. The clearest example of emotive language is poetry. Until the seventeenth century many. and typically involve a subject and a verb. Rhemes typically consist of a verb and its post-verbal elements. take. ) and rhemes (.19. . Four other verbs. The verb . . directed at any of the objective relationships between words and things. I glance in the rear-view mirror to see other cars close behind. concern.g. Textual sentence stems are routinized fragments of sentences which serve specific textual or organizational functions. . are significant left and right co-occurrents of the noun example. This was especially true in eastern Europe.

23) and in imperative sentences (13.9%.24. but active structures in which the subject is the example (Example 4. for example.79%.20) are more frequent. To take one example. When used in the imperative. Take the example of following an object by eye-movements (so-called ‘tracking’). at the beginning of the project seven committees were established.98 Academic Vocabulary in Learner Writing provide can be used in active or passive structures. the verb consider is mainly used with the noun example in imperative sentences (70%).25. It is never used to introduce an example.22). above). 4. usually second person imperatives (Example 4. The verb see always co-occurs with the noun example in the second person of the imperative (Example 4.21. to investigate one of a range of competing architectural possibilities. it generally appears in the second person (Example 4. Example 4. In accordance with the theme of this chapter. Example 4. The verb cite is more often found in a passive structure in which example functions as a retrospective label (Example 4. each consisting of about six people. but always as an endophoric marker to direct the reader’s attention to an example elsewhere in the text. and I shall take four different examples of this kind of work as alternatives to the Prague School’s and Jakobson’s approach to the relationship between linguistics and literature.25) and less frequently first person plural. By contrast.26.20. The Magdalen College affair. I shall simply use ‘stylistics’ as a convenient label (hence the inverted commas) for the branch of literary studies that concentrates on the linguistic form of texts. 4.79%). provides a classic example of passive resistance. 4.24) and there is only one occurrence in the first person plural in the BNC-AC-HUM.23. 4. It also occurs in active structures with a personal pronoun subject (13. The most important vowel is set to two or more tied notes in a phrase designed to increase the lyrical expression (see Example 47.22.21). The two verbs often form rhemes with the noun example: 4. 4. 4. Consider the following example. The verb take is mainly used with the noun example in sentence-initial exemplificatory infinitive clauses (68. The verb include is used with the plural form of the noun example in subject position to introduce an incomplete list of examples in object position: . A famous passage of art criticism can be cited as one example entirely beyond dispute.26).

the whole novel being made up of dialogue and narrative units generated in waves by the central conversation. These significant co-occurrences illustrated in Examples 4. give more detail about them.29. suggest and show (Example 4. and make suggestions on their basis. this example [adv.29) to talk about conclusions that can be drawn from the examples. these co-occurrences are frequently used in adverbial clauses (e. outstanding. The adjectives above and following are used to situate the example in the text (Example 4.31.Rhetorical functions in expert academic writing 99 4. .) which describe examples.28).30 do not qualify as collocations as the meaning of the verb is not restricted by the noun example.g.g. typical. .28 to 4. The floral examples include a large lotus calyx and two ivy leaves joined by a slight fillet. Table 4. fine.30. classic.32. they function as endophoric markers in Hyland’s (2005) typology of metadiscourse features. 4. These include quote (Example 4.27. ) and sentence stems (e. there is a sense in which all observers see the same thing. Another set of verb co-occurrents of example is used to discuss the examples given in a text.] illustrates .33): used with the noun example. and the combinations are fully explicable in semantic and syntactic terms. 4. The example shows that the objector’s neat distinction between adjudicative and legislative authorities is mistaken. prime (Example 4. in all the examples quoted here. which pivots around a four-hour conversation between two characters. The prime example is the Dada movement. whose nihilistic work is now admired for its qualities of imagination. Thirdly. 4. 4. as the two men’s review of their past lives sparks off inner thoughts and recollections and conjures up other conversations and dramatised episodes.g. .30) to show what something is like or that something is true. The advantage of using the noun example rather than the adverbials for example or for instance is that it allows the writer to evaluate the example in terms of its suitability. 4.28. .g. as this example suggests .5 gives the 24 adjectives that significantly co-occur with the noun example in the BNC-AC-HUM. e. . This example clearly illustrates the theory dependence and hence fallibility of observation statements. However.32).31) or typicality. excellent (Example 4. and illustrate (Example 4. An outstanding example of this type of narrative is Vargas Llosa’s Conversation in the Cathedral. good. e.

i. However. The second element. Following Granger and Paquot (2008a: 43).e. . Collocations represent 8. the ‘collocate’ or ‘collocator’. i. i. standard. 9 8 8 9 7 6 7 5 6 5 5 6 4. the adjective is only commonly used in this sense with a very limited number of nouns— example. a word combination that is semantically fully compositional. mistake and case3.33. syntactically fully flexible and collocationally open: the adjective classic is used with a meaning that is listed as its first sense in the Longman Dictionary of Contemporary English (LDOCE4) (1.87 per cent of the tokens of adjective + example co-occurrences. The co-occurrence prime + example is a clear example of a collocation: the adjective prime has two core meanings – ‘most important’ and ‘of the very best quality or kin’ – but a prime example is ‘a very typical example of sth’.100 Academic Vocabulary in Learner Writing Table 4. Both lexemes make a separate semantic contribution to the word combination but they do not have the same status. 38 15 18 10 16 11 13 10 12 16 13 9 Adjective fine notable isolated interesting known excellent prime trivial previous remarkable numerous single freq. I classified co-occurrences of this type as collocations.e. 1996). approved as a model.e. TYPICAL: having all the features that are typical or expected of a particular thing or situation) and the Oxford English Dictionary Online2 (1. This is clearly an illustration of the difficulty of separating the senses that a word has in isolation from those that it acquires in context (see Barkema.3 per cent of the types and 6. is selected by and semantically dependent on the ‘base’. Consider the following example. There is a case for considering the co-occurrence classic example as a free combination. leading). usage-determined or preferred syntagmatic relations between two lexemes in a specific syntactic pattern. of the first class.5 Adjective co-occurrents of the noun ‘example’ in the BNC-AC-HUM Adjective good above following well-known obvious classic typical outstanding extreme clear simple striking freq. The ‘base’ of a collocation is semantically autonomous and is selected first by a language user for its independent meaning. of the highest rank or importance.

Right co-occurrents include the preposition of and the pronoun this. to analyse the common co-occurrences of a word in a large corpus is made clear by comparing the significant adjective co-occurrents of the noun example (listed in Table 4. however. Similarly. crass. glittering. The added value of using statistics. Second. emotive. and (. unidiomatic use in learners’ writing. these co-occurrents are best described as ‘singularities’ and do not represent ‘the habitual usages of the majority of users’.34). that they are pedagogically uninteresting. and more specifically association measures. another and one) are more frequent than the definite article the with example. First. they constitute ‘preferred ways’ of qualifying example as they are repeatedly used with this noun. at least. cautionary). hideous). consummate. Apart from verbs and adjectives. the meaning of an outstanding example is composed of the meanings of the adjective outstanding and the noun example.5. Thus. In 40 per cent of its occurrences. awe-inspiring. The first two examples discussed below illustrate different ways in which the linguistic model is used to develop a narrative model. Left cooccurrents include determiners and the pronoun this. 1999a) has pointed out that German learners’ use of adjectives. other significant co-occurrents of the noun example are found in professional academic writing. The pronoun this is typically used as a subject with the verb be to refer back to an example given in a previous sentence (see Example 4. alarming. in her study of verb + noun combinations. Lorenz (1998. which in turn is often determined by a demonstrative (Example 4. differs from that of native students. class or event exemplified. ).15 above). The is mainly used when the noun is qualified by a superlative adjective or preceded by ordinals such as first. . and adjectives which occur only once or twice in the corpus (exquisite.5) with attested adjectival collocates (as given by Siepmann. Siepmann listed a number of adjectives that do not appear even once in the 87-million word written part of the BNC (beguiling. 2005: 137). apposite. 4. well-worn. edifying. happy. the noun example is directly followed by the preposition of which introduces the idea. Indefinite determiners (a.31 above) or pronominalized to refer back to a previous sentence. In addition to most of the adjectives given in Table 4. next and last (Example 4. To use Sinclair’s (1999: 18) words. .Rhetorical functions in expert academic writing 101 Other adjectives form semantically and syntactically fully compositional sequences with the noun example. irrespective of their phraseological status. . This does not mean.34. eminent. anodyne. Nesselhauf (2005) has shown that free combinations are prone to erroneous or.

and it is possible that the appointment of Praeiectus and the retirement of Bonitus were less creditable than their hagiographers claim. etc’ (Example 4. 1993). 4. The verb exemplify is used with the meanings of ‘to be a very typical example of something’ and ‘to give an example of’. These findings also provide strong evidence against the use of stopword lists when extracting co-occurrences from corpora as there is a serious danger of missing a whole set of phraseological patterns (Clear. illustrated and published by Chris Trump). Nevertheless they do illustrate the complexities of local ecclesiastical politics.37. (BNC-NEWS) Figure 4. The verbs illustrate and exemplify can also be used as exemplifiers.5 also shows that the verb illustrate is more frequent than exemplify in professional academic writing. The verb illustrate is used with the meaning of ‘to be an example which shows that something is true or that a fact exists’ (Example 4. Also in the pipeline is an Australian children ‘s TV series based on Gumnut Factory Folk Tales (written. 4. The narratives of the Passio Praeiecti and of the Vita Boniti both have their peculiarities. Function words seem to display co-occurrence preferences just as content words do (also see Renouf and Sinclair’s (1991) notion of a ‘collocational framework’). My aim will be to illustrate different ways of approaching literature through its linguistic form.35) or ‘to make the meaning of something clearer by giving examples’ (Example 4.36.37). Exemplify is very rarely used in other genres. 4. Table 4. their word forms and tenses in the BNC-AC-HUM4 were computed in the way described by Granger (2006). but with the meaning of ‘to put pictures in a book.36) (LDOCE4). ways involving the direct application of linguistic theory and linguistic methods of analysis in order to illuminate the specifically literary character of texts. Both verbs are more frequent in academic writing than in any other genre.35.5 compares the relative frequencies of the two verbs in academic writing with three main genres represented in the British National Corpus. article. The frequencies of the two verb lemmas. The verb illustrate is not uncommon in news but a quick look at its concordances shows that a significant proportion of its occurrences are used not to introduce an example.6 shows that there is no .102 Academic Vocabulary in Learner Writing These findings support Gledhill’s (2000) view that there may be a very specific phraseology and set of lexico-grammatical patterns for function words in academic discourse. Figure 4.

39).000 words BNC-AC-HUM 97 36 61 84 7 0 77 63 15 2 13 259 3. Almost all occurrences of the past participle appear in the passive construction BE illustrated by/in (Example 4. illustrated and illustrates.77% 5% 100% major difference in proportion between the verb forms illustrate.321. figure.38).8 37.6 The use of the lemma ‘illustrate’ in the BNC-AC-HUM The lemma illustrate illustrate simple present infinitive illustrated simple past present/past perfect past participle illustrates illustrating continuous tense -ing clause Total Nr of words Relative freq.32% 5.89% 23. case or approach (Example 4. . table.79% 0.867 7.55% 32. When used in active structures.73% 24.43% 2. the verb is often preceded by a non-human subject such as example.45% 13. per 100.Rhetorical functions in expert academic writing 140 frequency per million word 120 100 80 60 40 20 0 Academic News illustrate Fiction exemplify Speech 103 Figure 4.5 The distribution of the verbs ‘illustrate’ and ‘exemplify’ across genres Table 4.7% 0% 29.

. The noun point is used as an object of illustrate which refers back to an idea put forward in a previous sentence:] in a 1R-3R window. or forms – for many would prefer to talk of ‘the schizophrenias’ – there is still no universally accepted set of criteria for diagnosis. In the course of their analysis the authors determined the extent to which their proposed criteria agreed with those contained in other existing diagnostic schemes – some ten or twelve of them. To illustrate this.40) represents 2. 4. This co-occurrence is even more marked in academic genres such as social sciences. To illustrate the point.) is used either as the subject of the verb illustrate or in the passive structure illustrated in Figure x. (W_ac_medicine BNC sub-corpus. Correlations varied over a very wide range. . 4.38. The contrast between the conditions on the coast and in the interior is illustrated by the climatic statistics for two stations less than 30 km (18. For most of this century it is those disorders gathered together under the heading of ‘schizophrenia’ that have been used as the paradigm for trying to describe and understand psychosis. example [LogL = 49. see Table 3.41.42 and 4. This example clearly illustrates the theory dependence and hence fallibility of observation statements. and with the nouns point [LogL = 168. proposing a new set of rules for diagnosing schizophrenia. In the BNC-AC-HUM. The noun figure (and the abbreviation fig. Figure 1 illustrates the spread of results for the alcoholics and the controls. . . ) Whatever the answer to such a question. How many observations make up a large number? (. one of the present authors was recently asked to review a paper submitted to a prominent psychiatric journal. natural sciences and medicine which rely extensively on figures. tables and diagrams (see Examples 4. I refer to the strong public reaction against nuclear warfare that followed the dropping of the first atomic bomb on Hiroshima towards the end of the Second World War.2) . examples can be produced that cast doubt on the invariable necessity for a large number of observations.78]. .7 per cent of the occurrences of the lemma illustrate in the BNC-AC-HUM.39. Yet even in this form. illustrate significantly co-occurs with the noun example [LogL = 112] in a 3L-1L window.65] and fig.43). The sentence-initial adverbial clause To illustrate this/the point/X.5 miles) apart. 4.104 Academic Vocabulary in Learner Writing 4. [LogL = 45. (Example 4.

7 shows that the lexico-grammatical preferences of the verb exemplify differ from those of illustrate. see Table 3.26% 55. While our discussion in this chapter is of the doctrine of neutrality as such.7 The use of the lemma ‘exemplify’ in the BNC-AC-HUM The lemma exemplify exemplify simple present infinitive exemplified simple past present/past perfect past participle exemplifies exemplifying continuous tense -ing clause Total Nr of words Relative freq.45. Table 4. 4.46.2) The adverbs well.000 words BNC-AC-HUM 9 5 4 53 8 1 44 15 2 0 2 79 3. the fallow stage is contributing to crop productivity as well as providing protection against soil erosion. Rawls ‘ treatment of it will serve to illustrate the problems involved. per 100. The history of the English monarchy well illustrates both the importance and the unimportance of war.53% 100% . better.44.7% 19% 2.38 11.44). like the Peruvian example discussed above.47).46). This prejudice against close involvement with the secular government may be illustrated by an anecdote related in the about Molla Gurani. can and may (Example 4.4% 6. 4. (W_ac_soc_science BNC sub-corpus. 4. The advantages of the system are illustrated in Fig.2 and. A large proportion of the occurrences Table 4.45).321. We recently did a simple experiment which happens to illustrate how children’s knowledge of where an object is determines their behaviour.Rhetorical functions in expert academic writing 105 4. best and clearly are sometimes used with illustrate to evaluate the typicality or suitability of the example (Example 4. and with the modals will. 4. with the verb serve (Example 4.33% 5% 67% 10% 1.47. 8.867 2. The verb illustrate also co-occurs significantly with how to introduce a clause (Example 4.43.53% 0% 2.

. without argument.) and rhemes (… is an example of . of the verb. for example . . Siepmann (2005) analysed a 9. is superior to forms of knowledge that do not share its methodological characteristics. the verb significantly co-occurs with the verb be and the conjunction as in a 3L-1L window. He enumerated every . the exceptions are a few collocations such as prime example and classic example. detailed and testable expression. In the BNC-AC-HUM.g. .50. the adverb notably. Examples include . imperative clauses (Consider. . . . however. the adverbials for example and for instance). the clerk Jankyn. but did not make use of statistical methods. who. as exemplified by physics. . provides a classic example of . exemplified in the assimilation-accommodation theory of infant learning mentioned above.48) and the lexico-grammatical pattern as exemplified by/in (Example 4. word-like units or mono-lexemic phrasemes (the preposition such as. 4.106 Academic Vocabulary in Learner Writing of the lemma exemplify are –ed forms. the abbreviation e.3.50). The function of exemplification can be fulfilled by a whole spectrum of single words (the preposition like. The association of this material with the clerk is clearly exemplified by Chaucer’s wife of Bath’s fifth husband.) and word combinations. and more precisely past participle forms. . reads antifeminist material to her from his book Valerie and Theofraste.). . . He assumed. These significant co-occurrents highlight the preference of the verb for the passive structure BE exemplified by/in (Example 4.) and sentence-initial infinitive clauses (To take one example.49. 4.2. that science. A large majority of these word combinations are semantically and syntactically fully compositional.49). in the Wife of Bath’s Prologue. . and with the prepositions by and in in a 1R-3R window. . characterized by their high frequency of use and can be described as ‘preferred ways’ (Altenberg. 4. . the verb exemplify does not co-occur significantly with nouns. Piaget’s claim that thinking is a kind of internalised action. 4.).5-million word corpus of academic writing. preceded by a comma (Example 4. Unlike illustrate. i. .e. 1998) of giving an example in professional academic writing. . Exemplify is also often used after a noun phrase. is really a global assumption in search of some refined. The corpus-based methodology adopted has highlighted a number of lexical items that are repeatedly used as exemplifiers in academic writing.48. Discussion The description of exemplifiers presented here does not aim at exhaustiveness in professional academic writing but at typicality. They are. sentence stems (An example of Y is X.

NP Consider a(n) (ADJ) example/instance take the example of (as examples of NP) consider (as an example) NP take. consider NP (2) Take (even) NP (2) Let us (now) take + (as) + DET + ADJ + example(s) Let us consider + DET + ADJ + example(s) Let me give (you) (but) one example Let me offer + DET (+ ADJ+) example Let us consider. only a very limited set of these are widespread in professional academic writing. Although a large range of exemplificatory imperatives may be available to language users. however. Siepmann. NP Total Frequency 200 54 16 7 5 3 1 2 2 4 4 2 1 1 302 % 66.3 2.8) shows. wrote that English authors have a large range of exemplificatory imperatives at their disposal.3 1. It may be argued that privileging exhaustiveness over typicality in corpus linguistic research is counter-productive. using the direct second-person imperative VP ~ as well as the less imposing hortative let us + VP and the inclusive let me + VP. for the sake of illustration.3 0. NP as an illustration (of this)/ by way of (brief) illustration.2 17.3 0.3 1.7 0. (Siepmann.5-million word corpus of professional academic writing he used. and that such an approach results in too much — unreliable — information.Rhetorical functions in expert academic writing 107 single occurrence of word sequences used to give an example and listed rare events such as the infinitive clauses to paint an extreme example and to pick just one example (a single occurrence in his corpus). for (another) example. 2005: 119) Imperatives in academic writing (for example/for instance) see (for example/for instance) NP (for example) consider (for example) NP take.3 0.8 The use of imperatives in academic writing (based on Siepmann. as an example.9 5. the co-occurrence example + is afforded by and the expression for the sake of example. Table 4. the former is around five times more frequent than the latter.7 0. for example.3 100 . Of these last two. showing a high degree of audience sensitivity among authors. First person imperatives are extremely rare and let me + VP only appeared three times in the 9.4 per cent of the imperatives Siepmann found.7 1. 2005:120) A closer look at his frequency data (reprinted in Table 4. that the co-occurrences see/take/consider + for example account for 89.7 1 0.

yet). and notably). ‘expressing a concession’ and ‘reformulating’ in an attempt to give a wider overview of the way academic vocabulary is used to serve specific rhetorical functions. in spite of). source. verbs and adjectives in specific phraseological or lexico-grammatical patterns. however. The expressions to name but a few and by way of illustration are rare in all types of discourse. these two functions can also be realized by means of nouns. although. e. but they are infrequent. effect. The phraseology of rhetorical functions in expert academic writing This section briefly comments on the types of lexical devices used by expert writers to serve the functions of ‘expressing cause and effect’.10). consequence.g. The exemplificatory lexical items which were extracted are of two types: — the most frequent exemplifiers in academic writing (such as. As shown in Table 4. result.g.5 per cent of the lexical means used to express a cause or an effect in academic writing.9 shows that the lexical means of expressing a concession consist of single word adverbs (e. However. cause.g.11. prepositions and conjunctions also represent a large proportion of the lexical devices used by expert writers to serve the functions of ‘expressing cause and effect’ (Table 4. for example and for instance) (see Figure 4.11) and ‘comparing and contrasting’ (Table 4. factor. ‘comparing and contrasting’. The nouns illustration and case in point are quite characteristic of formal textual genres. example.1 discussed earlier in this chapter).108 Academic Vocabulary in Learner Writing The analysis of exemplifiers presented here also validates the method used to design the Academic Keyword List. the abbreviation i. example. The preposition like can be used to fulfil an exemplificatory function in academic writing but it is much more common in other genres. (complex) conjunctions (e. exemplify. reformulation is most frequently achieved by means of the mono-lexemic units that is and in other words.g. for example and for instance. outcome and implication. It aims to characterize the phraseology of these rhetorical functions in academic prose. . Similarly.e. 4.g.12). despite. Table 4. Adverbs. nouns account for 32. nevertheless. e. — lexical items which are not as frequent as such as. but which are more frequent in academic prose than in any other genres (illustrate. and the adverb namely (Table 4. even though) and (complex) prepositions (e.3.

8 7.9 20.5 6.7 2.056 28.817 6. 109 Table 4.26 3.8 0.4 144.3 % Rel.6 .10 Ways of reformulating. freq.1 28.792 19.3 5.6 5.8 1.3 2. 9.8 1.6 0.9 100 Rel. Adverbs however nevertheless nonetheless though ADV yet TOTAL ADVERBS Conjunctions although though CONJ even though (even if) albeit TOTAL CONJ.7 182.4 6.Rhetorical functions in expert academic writing Table 4.6 100.6 0.46 353 2.727 5.5 4.8 0.353 676 66 144 1.3 7.2 16.5 100 20.4 0.5 51.9 11. TOTAL 681 159 39 879 11.2 1.0 51.3 2.2 26. Prepositions despite in spite of notwithstanding TOTAL PREP.2 15.314 % 25.2 2.9 Ways of expressing a concession in the BNC-AC-HUM Abs.292 1. or more precisely or more accurately or rather TOTAL 330 375 81 210 187 21 12 7 91 1. paraphrasing and clarifying in the BNC-AC-HUM Abs.86 69.0 14.6 2. i.9 0. freq.e.1 3.6 1.0 4.721 248 451 80 4.4 0.5 6.3 54.7 39. that is that is to say in other words namely viz.5 13. freq.7 40. freq.5 14.6 0.

5 1.2 2. comparison.6 35.52 22.110 Academic Vocabulary in Learner Writing Verbs are also common: cause.8 2.7 16.5 1.2 0.830 813 143 411 8.9 9. contrast.2 0.7 32.12).1 0.4 14.6 55.3 12.7 6.4 15.2 3.612 2.0 2. Table 4.8 1.3 6.9 0.06 17.0 24.7 2.175 500 183 1.5 0.5 0. lead to.6 0.4 259. contrast. difference and distinction) and verbs (e.0 2.0 0.3 14. freq.2 per cent of the lexical means used by expert writers (Table 4. differ.8 0.g.2 1.1 4.9 128.8 3.0 5. different. nouns cause factor source origin root reason consequence effect result outcome implication TOTAL NOUNS Verbs cause bring about contribute to generate give rise to induce lead to prompt provoke result in yield make sb/sth do sth arise from/out of derive emerge follow from trigger stem TOTAL VERBS 570 125 276 227 101 67 671 115 161 327 129 171 145 476 466 74 56 95 4.2 3.9 3.g.3 0. emerge. Patterns involving nouns (e.11 Ways of expressing cause and effect in the BNC-AC-HUM Abs.2 0. contribute to.8 0.g.8 3.4 1.5 4.8 8.6 1. freq.8 1. distinguish and differentiate) are often used to compare and contrast but adjectives (e. .25 % Rel.4 0.5 54.2 13. distinct.0 755 550 1. result in.4 16.252 2.802 450 1.2 4. bring about.4 0. differing and distinctive) play a more prominent role and account for 29.5 4.7 6.0 20. derive.6 1. and stem.9 5.

0 1.3 0.3 0.7 1.3 3. Adjectives consequent responsible (for) TOTAL ADJ.5 57.9 4. Adverbs therefore accordingly consequently thus hence so thereby as a result as a consequence in consequence by implication TOTAL ADVERBS Conjunctions because since As 5 for so that PRO is why that is why this is why which is why on the grounds that TOTAL CONJ. Prepositions because of due to as a result of as a consequence of in consequence of in view of owing to in (the) light of thanks to on the grounds of on account of TOTAL PREP.7 0.7 0.1 0.0 2.5 0.981 5.3 1.4 1. TOTAL 2.97 796.1 7.2 21.99 1.5 177.6 3.7 0.4 0.33 100 66.5 6.04 599 195 196 22 1 66 52 109 35 22 24 1.9 2.6 0.5 3.7 39.2 0.475 8.5 0.6 3.767 283 1.3 53.1 0.Rhetorical functions in expert academic writing Abs. freq.7 0.312 22.1 0.7 0. freq.6 31.894 182 101 20 14 35 5.3 3.6 10.1 4.321 2.7 0.2 0.912 26.4 12 % Rel.49 1.2 8.1 0.0 0.3 1.8 53 344 397 0.2 0.412 130 143 1.4 0.1 0.1 22.1 0.0 5.9 0.6 0.1 180.6 0.0 0.0 1.3 0.9 5.2 1.0 0.4 28.0 0.99 18.4 2.0 0.207 955 883 1.0 5.5 3.7 26.036 696 52 22 18 12 83 5.59 42. 111 .

8 2.3 0.1 2.1 0.1 29.7 4.3 0. Nouns resemblance similarity parallel parallelism analogy contrast comparison difference differentiation distinction distinctiveness (the) same (the) contrary (the) opposite (the) reverse TOTAL NOUNS Adjectives same similar analogous common comparable identical parallel alike contrasting different differing distinct distinctive distinguishable unlike contrary opposite reverse TOTAL ADJECTIVES Verbs resemble correspond look like compare parallel contrast 138 137 102 278 56 137 0.5 0.6 0.4 4.5 0.112 Academic Vocabulary in Learner Writing Table 4.46 3.5 0.7 30.6 0.4 0. .2 8.027 55 1055 223 137 52 98 63 2.9 0.7 0.6 2.8 3.1 4.9 3.3 % Rel.4 1.12 Ways of comparing and contrasting found in the BNC-AC-HUM Abs.9 1.9 75.8 6.0 1.1 4.2 0.7 257.3 2.4 0.9 0.9 0.1 3.6 1.3 0.2 14.2 8.4 0.552 0.496 72 278 163 33 43 27 127 23 8.3 16.580 1.1 0.3 0.9 0. freq.7 127.6 1.2 3.5 6.5 0.24 77.4 4.3 17.5 0.6 5.229 0.7 4.1 0.0 1.0 0.8 1.1 0.5 4.5 0.7 9.318 76 595 10 559 28 85 56 4.1 1.1 2.8 0.9 1.8 0.3 15.1 0.44 116 212 147 19 175 522 311 1. freq.7 2.7 31.4 39.1 8.3 0.2 0.5 0.8 0.9 1.

7.2 2.8 0.2 5.1 1.2 0.3 2.0 0.1 0.1 0.6 0.39 0.5 0.9 50.1 0.0 0. 2.6 242 404 74 1.9 5.4 0.2 4.4 0.2 113 0.9 0. differ distinguish differentiate TOTAL VERBS Adverbs similarly analogously identically correspondingly parallely likewise in the same way contrastingly differently by/in contrast by contrast in contrast by way of contrast by/in comparison by comparison in comparison comparatively contrariwise distinctively on the other hand (on the one hand) on the contrary quite the contrary conversely TOTAL ADVERBS Prepositions like6 unlike in parallel with as opposed to as against in contrast to/with in contrast to in contrast with versus contrary to by/in comparison with in comparison with in comparison to by comparison with in comparison with TOTAL PREP.6 1.3 0.4 2.7 0.9 0. freq.0 0.2 0.2 0.1 1.3 0.4 0.2 0.1 0.Rhetorical functions in expert academic writing Abs.6 0.72 0.8 1.674 1.7 11.3 5.3 84.2 0.5 1.88 (Continued) 394 2 2 29 0 118 56 3 97 185 116 69 0 23 14 9 69 4 25 372 136 95 2 62 1.2 47.0 0.00 1.2 0.0 0.6 11.1 0. freq.2 3.1 2.1 0.4 0.568 % 0.0 3.2 11.3 0.3 0.6 2.1 2.91 .3 12.36 Rel.0 0.5 2.3 0.4 0.812 244 8 121 46 82 73 9 53 66 52 14 4 21 14 3.6 0.6 7.4 104.9 0.0 0.3 1.1 0.484 9.0 0.7 0.0 0.

freq. 0. .51. This may be an effect of the uncertainty around television’s textuality. Conjunctions as while whereas TOTAL CONJ.751 17. This had important implications for the debate over access to birth control information and abortion – rarely were demands for freer access to birth control information devoid of maternalist rhetoric.23 % Rel. effect. Another direct result of conquest by force of arms was the development of slavery. Other expressions as .9 38. but it is now an extremely limiting effect for the development of theory.5 0.2 4. Most of the co-occurrents listed form quite flexible and compositional textual sentence stems with their nominal node.54. as illustrated in the following examples: 4.3 203.4 1. . However it is first necessary to consider another important consequence of the view of psychosis being presented here.55.12 Cont’d Abs.1 100 Table 4. 4.249 9.3 1. but actually merge with each other. freq.3 1.1 4.766 38 155 113 42 32 11 20 1 29. .3 1.5 23.53.114 Academic Vocabulary in Learner Writing Table 4. outcome.1 0.08 151. 4. 4.0 880.0 0. 4.3 0. which was widespread up to the beginning of the nineteenth century. Health for women was held to be synonymous with healthy motherhood.6 0. as in the same way as/that compared with/to compared with compared to CONJ compared to/with as compared to/with when compared to/with if compared to/with TOTAL 2.7 3.5 83.5 5.0 13.13 shows a co-occurrence analysis of several nouns that are used to express cause or effect in academic prose: reason.045 1264 442 6.52. The reason is that with Van Gogh art and life are not merely conditioned by each other to a greater degree than with any other artist. implication. result and consequence.

13b: implication Adjective + implication important practical political serious social Verb + implication have carry implication + verb be Auxiliary verb + implication be implication + preposition of for Preposition + implication with Determiner + implication this implication + conjunction that .Rhetorical functions in expert academic writing 115 Table 4. .13 Co-occurrents of nouns expressing cause or effect in the BNC-AC-HUM Table 4. supposing believing thinking accepting rejecting adopting There + verb + reason There is (no) reason to There seems no reason There are (DET/ADJ) reasons Table 4.13a: reason Adjective + reason good main sufficient obvious other different alleged simple tactical political major additional right valid similar fundamental real independent special possible historical particular Verb + reason have give see base on provide find examine Auxiliary verb + reason be seem reason + verb be justify reason + preposition for against Preposition (2L) + reason for reason + conjunction why which that Determiner + reason this another (no) reason to + verb believe suppose doubt prefer think fear accept reason(s) for . . .

13c: effect Adjective + effect adverse overall good profound knock-on indirect far-reaching damaging cumulative dramatic immediate excellent long-term practical particular powerful special full general important other Verb + effect have produce achieve create cause Auxiliary verb + effect be effect + verb be depend on occur effect + preposition of on upon Determiner + effect this effect + conjunction That Noun and effect cause Table 4.13d: outcome Adjective + outcome logical eventual likely different inevitable final outcome + preposition of Determiner + outcome this Verb + outcome influence determine represent affect outcome + verb be Auxiliary verb + outcome be .116 Academic Vocabulary in Learner Writing Table 4.

Rhetorical functions in expert academic writing Table 4.13f: consequence Adjective + consequence inevitable unintended unfortunate direct important necessary political natural bad practical social likely major possible Determiner + consequence this another consequence + conjunction that Verb + consequence have suffer (from) avoid consider outweigh discuss consequence + verb be follow ensue Auxiliary verb + consequence be consequence + preposition of for Preposition (3L) + consequence with of - .13e: result Adjective + result inevitable direct immediate beneficial eventual interesting practical main similar result + preposition of from Preposition (3L) + result with Determiner + result this Verb + result produce achieve yield give bring lead to show present interpret obtain have result + verb be Auxiliary verb + result be 117 Table 4.

major. important. tend or prove) is mainly used in passive constructions. . A few co-occurrences are collocations as illustrated by Example 4. “consciously intending” or being compassionate. as suggested above. appear. performing an attitudinal and pragmatic function in the discourse’ (Tognini-Bonelli. and in phrases introduced by the conjunction as (e. prove and tend typically used to express possibility or certainty. in the absence of speech. both of which carry implications of understanding to some degree.g. 1999: 16). inevitable. sufficient) and are used to express the ‘writer’s attitude or stance towards. outcome. 2002. serious. it has been suggested that. proved a complete failure). .g.56. effect. The variety of adjectives used with the nouns reason. the –ed form of suggest (unlike that of appear. Like nouns. verbs that serve specific rhetorical or organizational functions in academic prose generally enter compositional and flexible sequences. 4. as already suggested by). associating the formal patterning with a semantic field. . For example. Tutin. A large proportion of those are evaluative adjectives (e. Table 4.14 gives the most frequent lexical bundles containing one of the four verbs suggest. it has been suggested. . which basically means have implications. and an identifiable semantic prosody. ) or rhemes (e. sentence-initial adverbial clauses (e. We may certainly talk of animals. These extended units of meaning are categorized as textual phrasemes in Granger and Paquot’s (2008a) typology as they function as sentence stems to organize the propositional content at a metadiscoursal level. . implication. and thus different complete units of meaning. .51 to 4. it is sometimes suggested). 2000: 5). fundamental. result and consequence is also worthy of note and bears testimony to their prominent role in argumentation (Soler. 2002: 79). forthcoming). The verb carry is used in a delexical sense in the collocation carry implications. It is often used to report suggestions made by other people in impersonal structures introduced by it (e. it appears that). .g.g. or feelings about the entities or propositions that he or she is talking about’ (Hunston and Thompson. good.55 are good illustrations of what Sinclair and his followers have called ‘extended units of meaning’ where lexical and grammatical choices are ‘intertwined to build up a multi-word unit with a specific semantic preference. and that these constitute different form/meaning pairings.g.g. It is worth noting that each verb form has its own ‘distinctive collocational relationship’ (Sinclair. viewpoint on. Most clusters are lexico-grammatical patterns which function as textual sentence stems (e.56.118 Academic Vocabulary in Learner Writing The word combinations illustrated in Examples 4.

14 Co-occurrents of verbs expressing possibility and certainty in the BNC-AC-HUM Table 4. In summary.g. phrasal verbs. it may/might be suggested that) to make a tentative suggestion. idiomatic sentences. Suggested is also used in impersonal structures introduced by it followed by a modal verb (e. proverb fragments and the like (see also Pecman. Table 4. similes. the verb form suggests is typically used to make it clear that the suggestion offered is made on the basis of who/whatever is the subject of the sentence: 4. as I have suggested) to refer to a suggestion previously made. even) suggested that – it can / could / may be suggested that – this is suggested by – as (already) suggested by – as suggested above – (as) I (have) (already) suggested suggests – NP / it / this (ADV: strongly. results indicate that the phraseology of rhetorical or organizational functions in academic prose does not consist of idioms. More recent evidence suggests. as suggested above) and/or the first person pronoun I (e. (ADV: strongly) suggesting (that) – I am (not) suggesting that . 1998) conclusion that a rm large proportion of the lexical collocations found in academic discourse consist of a verb in a figurative sense and an abstract noun denoting a recurrent concept in academic discussion (e. adopt an approach/a method. which suggests (that) – as NP suggests suggest – NP / it / this might / may / would suggest (that) – NP does suggest (that) – there is evidence to suggest – I (would / want to) suggest – NP / it / this seems to suggest (that) suggesting – … . 4. By contrast..7 Referential phrasemes that serve to organize scientific discourse mainly consist of lexical and grammatical collocations. Sinclair Hood (1971) suggests that woollen cloth and timber were sent to Egypt in exchange for linen or papyrus.Rhetorical functions in expert academic writing 119 As-phrases are also used with an endophoric marker (e. however.57. . Results also confi Howarth’s (1996. also) suggests (that) – . .g. also.58. 2004 and Gledhill.g.g. 2000). commonly) suggested that – it was (first.14a: suggest suggested – it has been suggested that – it is (sometimes. that while it lives in woodland it actually hunts over nearby open areas.

may / might / would prove ADJ to – NP was to prove ADJ – attempt to prove – seek to prove proving Table 4. successful. as if. . look. develop an idea/a method/a model. by proving – . . possible – NP / it / proved to be (ADV) ADJ – NP proved NP proves – NP proves ADJ (impossible. possible) to prove . conclude / tend – NP tend to V (be.14d: tend tended – NP tended to V (be. . ignore. which tends to V – it tends to V V: be. take. necessary. The first is complex prepositions (e.14b: prove Academic Vocabulary in Learner Writing proved – NP / it / this proved to – NP / it / this proved (ADV) ADJ (to) with ADJ: difficult. . unable.g. . In academic prose. the category of textual phrasemes consists of three types of phraseme (cf. proving that – .g. inadequate. with respect to. even . reach a conclusion/a consensus/a point. difficult. impossible.6). abortive. . . Figure 4. easy. become. . . carry out a task/a test/a study). . favour. obscure. successful) – NP proves that – BE proving – . regard) tending draw an analogy/a comparison/a distinction. inadequate. confirm. of proving prove – ADJ (likely.14c: appear appeared – it appeared (ADJ) that – there appeared to be – this appeared to V – . see) tends – NP tends to V – . which appeared ADJ/ to V appears – NP / it / this appears to V – which appears to V – what appears to V – there appears to V – it appears that – as appears from/in appearing / appear NP would/might/may appear to be/V Table 4. so that. in addition to) and complex conjunctions (e..120 Table 4. . . support. see.

g. . Burger’s (1998) category of structural phrasemes). 2003: 269). which I refer to as ‘textual formulae’. . . . that is. . .’ to more inflexible phrasemes such as ‘to be a case in point’. which ‘form the springboard of utterances leading up to the communicatively most important — and lexically most variable — element’ (Altenberg. classic. possibly because rhemes are ‘usually tailored to expressing the particular new information the speakers want to convey to their listeners. good. in other words. as Altenberg (1998: 111) points out. and It is argued that. . as a result) and clausal linking adverbials (e. etc. Rhemes typically consist of a verb and its post-verbal elements (e.Rhetorical functions in expert academic writing Phrasemes 121 Referential function Referential phrasemes (Lexical) collocations Grammatical collocations Textual function Textual phrasemes Complex prepositions complex conjuctions Linking adverbials Textual formulae (including textual sentence stems and rhemes) Communicative function Communicative phrasemes Attitudinal formulae Figure 4. Examples include It has been suggested.6 The phraseology of rhetorical functions in academic prose though) which are used to establish grammatical relations (cf. is another issue). The second is multiword linking adverbials. Although the majority of linking adverbials are single adverbs. what is more. and are therefore not part of the phraseological spectrum. Another reason is . used to connect two stretches of discourse. another) ADJ (typical.g. 1998: 113). from flexible fragments such as ‘DET (a. They also sometimes function as textual phrasemes but are less frequent than sentence stems. . . “composed of variable items drawn from an open set”’ (De Cock.g. in addition. in conclusion. and are therefore. These first two categories of textual phraseme broadly correspond to Moon’s (1998) set of organizational fixed expressions and idioms. Textual formulae are particularly prominent in academic writing and display different degrees of flexibility. . Textual sentence stems and rhemes constitute the third type of textual phrasemes. for example. . to conclude) are also common in academic prose (Conrad.) example of [NP] is . 1999: 11–12). that is to say. prime. . Textual sentence stems are multiple clause elements involving a subject and a verb. prepositional phrases functioning as adverbs (e.

122 Academic Vocabulary in Learner Writing Attitudinal formulae make up a large proportion of communicative phrasemes in academic prose. together with information on the word’s frequency (see Coxhead et al. Such a contextual analysis will also make it possible to decide whether each word fits my definition of academic vocabulary and deserves to be retained in the Academic Keyword List. Summary and conclusion In this chapter. it seems that or it is noteworthy that. and outstanding example have traditionally been considered as peripheral or falling outside the limits of phraseology (Granger and Paquot. The list. .4. it should include the word combinations (frequent co-occurrences.. (forthcoming) for a similar project for Coxhead’s (2000) Academic Word List).g. They largely consist of sentence stems such as it is important/necessary that. collocations. This means that each AKL word has to be described in context. textual phrasemes. final outcome. example.) in which each AKL word is commonly found in academic prose. 2008a: 29) but results suggest that they are essential for effective communication and are also part of the preferred lexical devices used to organize scientific discourse. Co-occurrences such as direct result.. as was done above (Section 4. 2004: 389). To be useful to apprentice writers. e.’s (2004) category of stance bundles that ‘provide a frame for the interpretation of the following proposition. The AKL could be very useful for curriculum and materials design as it includes a high number of words that serve rhetorical functions in academic prose.2 has also validated the method used to select AKL words: the lexical items which were automatically extracted included the most frequent exemplifiers in academic writing (such as. notably). evidence suggests. The analysis of exemplifiers presented in Section 4. This group is similar to Biber et al. however. still needs to be refined in various ways. etc. I have shown that a high proportion of words in the Academic Keyword List (AKL) fit my definition of academic vocabulary and serve rhetorical or organizational functions in academic prose. 4. conveying two major kinds of meaning: epistemic and attitude/modality’ (Biber et al. exemplify. The frequency-based approach adopted to study the phraseology of rhetorical functions has also helped uncover a whole range of word combinations that do not fit traditional phraseological categories. for example and for instance) and lexical items which are not as frequent but which are more common in academic prose than in other genres (illustrate.2) for the function of exemplification.

(Gledhill. Phraseology is a dimension of language use in which patterns of wording (lexico-grammatical patterns) encode semantic views of the world. results have shown that textual phrasemes make up the lion’s share of multiword units that ensure textual cohesion in academic prose. Granger and Paquot. has often been neglected in theories of phraseology (cf.Rhetorical functions in expert academic writing 123 The type of data analysis presented in this chapter has also offered valuable insights into the distinctive nature of the phraseology of rhetorical functions in scientific discourse. the functions of all AKL words and their preferred phraseological and lexico-grammatical patterns should be identified by examining them in context. and at a higher level idioms and lexical phrases have rhetorical and textual roles within a specific discourse. Another objective of this chapter has also been to assess the adequacy of the treatment of rhetorical functions in EAP textbooks and investigate whether the AKL should be supplemented with additional academic words. namely collocation and the lexico-grammar. Attitudinal formulae serve a major role in a restricted number of functions such as ‘expressing personal opinion’ and ‘expressing possibility and certainty’. Most notably. Results have also pointed to the prominent role of free combinations to build the rhetoric of academic texts. The notion of phraseology implies much more than inventories of idioms and systems of lexical patterns. My findings thus support Gledhill’s call for a rhetorical or pragmatic definition of phraseology: Phraseology is the ‘preferred way of saying things within a particular discourse’. Some of these lexical items turned out not to be typical of academic prose or to be extremely rare (e. however. and a system of organization which encompasses more local lexical relationships.g. I claim that the phraseological analysis of a text should not only involve the identification of specific collocations and idioms. 2008a: 34–5). I listed the words and phrases given in academic writing textbooks as typical lexical devices to perform the five rhetorical functions analysed in detail in this book and compared them with the AKL. This type of phraseme. I identified the words that were not part of the AKL and examined their use in the BNC-AC-HUM. by way of illustration) and should therefore not deserve the attention they have been given in . Phraseology is at once a pragmatic dimension of linguistic analysis. 2000: 202) In line with this call. To do so. to name but a few. but must also take account of the correspondence between the expression and the discourse within which it has been produced.

do they use exemplifiers? Do they rely on words and phrasemes that are typical of academic prose? Do they use the expressions to name but a few and by way of illustration? If so. Such an analysis is presented in the next chapter.124 Academic Vocabulary in Learner Writing pedagogical materials. A pedagogically-oriented investigation of academic vocabulary cannot rest solely on native speaker data. do they use them correctly? And do they use them sparingly or do they make heavy use of these infrequent exemplifiers? These questions can only be answered by an analysis of learner corpus data. a large proportion of AKL words were not found in textbooks in spite of their relatively high frequency and major discourse functions in academic prose. It is essential to examine what learners actually do with lexical devices that serve rhetorical functions. . These findings show the power of a data-driven approach to the selection of academic vocabulary and clearly call for a revision of the treatment of rhetorical functions in academic writing textbooks. By contrast. For example.

html) was used to compute log-likelihood Instead.1 presents a detailed comparison of exemplificatory devices in native and learner writing. However these analyses are not presented in as much detail as for exemplification. This illustrates the type of results obtained when the range of lexical strategies available to EFL learners is compared to that of expert writers. The UCREL log-likelihood calculator website (http://ucrel.2 is on the general interlanguage features that emerge from these analyses. ‘comparing and contrasting’. The same methodology was used to examine learners’ use of words that serve the rhetorical functions of ‘expressing cause and effect’. lack of register awareness. These fall into six broad categories: limited lexical repertoire.1. However not all learner specific-features can be attributed to developmental factors.lancs. 6. both for reasons of space and because the presentation would soon become cumbersome. ‘expressing a concession’ and ‘reformulating: paraphrasing and clarifying’. The learner’s first language also plays a considerable part in his or her use of academic vocabulary. Section 5. A bird’s-eye view of exemplification in learner writing A general finding of the comparison between the International Corpus of Learner English (ICLE) and the British National Corpus – Academic Humanities . The whole learner corpus was compared to the BNC-AC-HUM but the results are only reported if they are common to learners from a majority of the mother tongue backgrounds considered. clusters of connectives and unmarked position of connectors. semantic misuse.64 (p < 0. the focus of Section 5. Differences between learner and native writing are highlighted by means of log-likelihood tests. In Section 5.Chapter 5 Academic vocabulary in the International Corpus of Learner English This chapter is devoted to academic vocabulary in learner writing. 5. I focus on transfer effects on French learners’ use of multiword sequences with rhetorical learner-specific phraseological patterns.01) was taken as the threshold value.3.

126 Academic Vocabulary in Learner Writing (BNC-AC-HUM) subcorpus is that exemplificatory lexical items are significantly more frequent in learner writing than in professional academic prose.g. The overuse of for instance has already been reported by Granger and Tyson (1996) for French learners and Altenberg and Tapper (1998) for Swedish learners. This shows that EFL learners’ overuse of the function of exemplification is largely explained by their massive overuse of the adverbials for example and for instance. Chen. The bar chart shows that EFL learners’ use of exemplifiers differs from that of professional writers in at least two ways. Overuse of for example has also been found in other learner populations such as Japanese and Taiwanese learners (Narita and Sugiura. which are underused in the ICLE. and does not apply to EFL learner writing in general: most L1 learner populations overuse exemplificatory discourse markers.. This explanation for German learners’ underuse of exemplifiers is not entirely satisfactory. By contrast.1. they do not choose the same exemplifiers. the nouns illustration and case in point and the expressions to name but a few and by way of illustration when comparisons are based on the total number of running words in each corpus. 2006). There is no significant difference in the use of the preposition such as. Figures and log-likelihood values for each corpus comparison are given in Table 5. .1 shows the frequencies per 100.000 words of exemplifiers in the ICLE and the BNC-AC-HUM. the abbreviation e. The lexical items are ordered by decreasing relative frequency in the ICLE. . The bar chart in Figure 5. whereas the most frequent one in the BNC-AC-HUM is such as. these lexical items are quite infrequent in both nativespeaker and learner writing. 2005: 255).g. Thus the most frequent exemplifier in the ICLE is the adverbial for example. learners tend to make little use of the verbs illustrate and exemplify and the adverb notably.) bespeak a general lack of concern for comprehensibility’ (Siepmann. the noun example 2 and the preposition like. This result highlights the importance of analysing several learner populations and comparing them so as to avoid faulty conclusions about EFL learner writing in general. The frequencies of individual items also differ widely.1. Siepmann (2005) finds that the adverbials for example and for instance are less frequent in German learner writing than in native and non-native professional writing and argues that ‘under-use of exemplification as a rhetorical strategy in student writing may (. Corpus comparisons based on the total number of running words have shown that exemplification is used significantly more in the ICLE than . 2006. Except for the preposition such as and the abbreviation e. the frequency of each exemplificatory lexical item can also be calculated as a proportion of the total number of exemplifiers. First.. As explained in Section 4.

80 70 60 50 40 30 20 10 0 Academic vocabulary in the ICLE fo xa re mp le exa mp le suc ha s like tance ins for e.1 Exemplifiers in the ICLE and the BNC-AC-HUM 127 . illu str ly ify ate ion point ion few tab empl rat rat ta no in x ust ust bu e e ill e f ill cas am yo on Ea wa t B by BNC-AC-HUM ICLE Figure 5.g .

128 Table 5.1 words

Academic Vocabulary in Learner Writing
A comparison of exemplifiers based on the total number of running

ICLE Abs. Nouns example example examples *exemple1 *exampl *examle illustration illustration illustrations (BE) a case in point TOTAL NOUNS verbs illustrate illustrate illustrates illustrated illustrating exemplify exemplify exemplifies exemplified exemplified *examplified exemplifying TOTAL VERBS prepositions such as like TOTAL PREP. Adverbs for example for example *for exemple for instance e.g. notably to name but a few by way of illustration TOTAL ADVERBS TOTAL 857 854 3 344 94 5 3 1 1304 3058 73.5 489 468 957 42 40.2 82.1 6 2 2 2 1 1 0 57 51 29 14 8 0 4.38 2.5 1.2 0.7 0 0.43 0.2 0.2 0.18 0.1 0.1 0 4.8 713 477 230 4 1 1 17 16 1 10 740 61.17 40.9 19.7 0.3 0.1 0.1 1.5 1.4 0.1 0.86 63.5 Rel.

BNC-AC-HUM Abs. Rel.


1285 665 620

38.68 20 18.7

91.6 (++) 134 (++) 0.5

77 63 14 18 1380

2.3 2 0.4 0.5 41.5


1.3 83.6 (++)

259 97 63 84 15 79 9 15 53

7.8 2.9 1.9 2.5 0.5 2.38 0.3 0.5 1.6

16.1 (− −) 0.6 2.6 17.7 (− −) 9 20.32 (− −) 0.4 2.1 20.09 (− −)

2 338 10.2


1.2 32.1 (− −)

1494 532 2026

45 16 61

1.8 199.6 (++) 55.3 (++)



209.9 (++)

29.5 8 0.4 0.3 0.1 111.9 262.4

609 259 77 4 3 2215 5959

18.3 7.8 2.3 0.1 0.1 66.7 179.4

47.3 (++) 0.1 22.1 (− −) 0.9 0 208.3 (++) 279.2 (++)

Legend: (++) significantly more frequent (p < 0.01) in ICLE than in BNC-AC-HUM; (− −) significantly less frequent (p < 0.01) in ICLE than in BNC-AC-HUM

Academic vocabulary in the ICLE
Table 5.2 A comparison of exemplifiers based on the total number of exemplifiers used
ICLE Abs. Nouns example illustration (BE) a case in point TOTAL NOUNS Verbs illustrate exemplify TOTAL VERBS Prepositions such as like TOTAL PREP. Adverbs for example for instance e.g. notably to name but a few by way of illustration TOTAL ADVERBS TOTAL 854 344 94 5 3 1 1301 3054 28 11.3 3 0.2 0.1 0 42.6 100 1263 609 259 77 4 3 2215 5959 21.2 10.2 4.3 1.3 0.1 0 37.2 100 39 (++) 2 8.1 (− −) 36.9 (− −) 0.2 0.2 15.3 (++) 489 468 957 16 15.3 31.3 1494 532 2026 25 8.9 34 80 (− −) 70.7 (++) 4.5 51 6 56 1.7 0.2 1.8 259 79 338 4.4 1.3 5.7 47.7 (− −) 35 (− −) 77.3 (− −) 713 17 10 740 23.3 0.6 0.3 24.2 1285 77 18 1380 21.6 1.3 0.3 23.2 2.8 11.7 (− −) 0 0.9 % BNC-AC-HUM Abs. % LogL


Legend: (++) significantly more frequent (p < 0.01) in ICLE than in BNC-AC-HUM; (− −) significantly less frequent (p < 0.01) in ICLE than in BNC-AC-HUM

in the BNC-AC-HUM, and that the four lexical items discussed above are largely responsible for this overuse (Table 5.1). Comparisons based on the total number of exemplifiers allow us to ask and answer different research questions. They give information about which lexical item(s) EFL learners prefer to use when they want to give an example, and in what proportions. Thus Table 5.2 shows that EFL learners select for example on 28 per cent of the occasions when they introduce an example, whereas native-speaker academics only use it to introduce 21 per cent of their examples. Both methods indicate that EFL learners overuse the preposition like and the adverbial for example. As shown in Table 5.3, however, the two methods may also give different results. The noun example appears to be overused in the ICLE when comparisons are based on the total number of running

130 Table 5.3
Lexical item

Academic Vocabulary in Learner Writing
Two methods of comparing the use of exemplifiers
Comparison based on total number of running words ++ // // ++ −− −− −− // ++ ++ ++ ++ // −− // // ++ Comparison based on total number of exemplifiers // −− // // −− −− −− −− ++ // ++ // // −− // // ++

example illustration (be) a case in point TOTAL NOUNS illustrate exemplify TOTAL VERBS Such as Like TOTAL PREPOSITIONS for example for instance e.g. notably to name but a few by way of illustration TOTAL ADVERBS

Legend: ++ significantly more frequent (p < 0.01) in ICLE than in BNC-AC-HUM; − − significantly less frequent (p < 0.01) in ICLE than in BNC-AC-HUM; // no significant difference between the frequencies in the two corpora

words in each corpus. However, a comparison based on the total number of exemplifiers suggests that the learners choose the noun example about as often as professional academics when they want to introduce an example (23.3% vs. 21.6%). More lexical items are significantly underused when figures are based on the total number of exemplifiers. In addition to illustrate, exemplify and notably, the noun illustration and the preposition such as are selected proportionally less often by EFL learners than by professionals to introduce an example. This first broad picture of the use of exemplifiers in the ICLE points to EFL learners’ limited repertoire of lexical items used to serve this specific EAP function. This characteristic of learner writing is discussed in more detail in Section 5.2.1. By comparison with academics, EFL learners overuse the preposition like and underuse such as. Figure 5.2 shows the relative frequencies per 1,000,000 words of like and such as in four sub-corpora of the British National Corpus representing different ‘super genres’ (see Section 3.3): academic writing, fiction, newspaper texts and speech (BNC-SP) as well as in the ICLE. The

Academic vocabulary in the ICLE
2500 2000 1500 1000 500 0 Academic writing Fiction News Speech Learner writing



such as

Figure 5.2 The use of the prepositions ‘like’ and ‘such as’ in different genres

50 45 40 35 30 25 20 15 10 5 0 speech Fiction Learner writing News Academic writing

Figure 5.3 The use of the adverb ‘notably’ in different genres

preposition like is much more frequent than such as in speech3, fiction, news and learner writing but is less frequent in academic prose. By contrast, such as is more frequently used in academic prose. Learners’ use of these exemplificatory prepositions thus differs from academic expert writing, but resembles more informal genres such as speech. Learners’ underuse of the adverb notably in their academic writing is another illustration of the same point (Figure 5.3).

freq. per million words


Academic Vocabulary in Learner Writing
for example 14% 2% 7% 10% 59% 11% 77% Academic writing News Fiction Speech 20% for instance

Figure 5.4 Distribution of the adverbials ‘for example’ and ‘for instance’ across genres in the BNC

A large proportion of EFL learner populations make repeated use of the word-like unit for instance. The use of this adverbial by native-speakers, however, differs significantly from that of for example, both in terms of frequency and register. Figure 5.4 shows that 77 per cent of all instances of for example in the BNC are found in the academic sub-corpus. However only 59 per cent of the occurrences of for instance appear in academic prose while 30 per cent are found in more informal genres such as speech and fiction. Lee and Swales (2006: 64) also showed that the use of these two adverbials differs across academic disciplines: for instance is more frequent in the social sciences and humanities while in natural sciences, technology and engineering, for example is strongly favoured to clarify a difficult or complex point through exemplification. Lack of register awareness manifests itself in a number of ways in learner academic writing. This will be the focus of Section 5.2.2. The phraseology of academic words is also a major source of difficulties to EFL learners. One of the main advantages of using a noun rather than the adverbials for example and for instance is that the use of a noun allows the writer to qualify the example with an adjective (see Section 4.2.2). However only 18 per cent of the adjective co-occurrents (types) of the noun example in the ICLE are significant co-occurrents in the BNC-AC-HUM (Table 5.4). A quarter of the adjective co-occurrents of example in the ICLE do not appear at all in the 100-million word British National Corpus (Table 5.5). A large proportion of these adjectives have been described by our

This mere example proves that the ideal union people dream of is not yet a total reality: national conflicts are still at work.1.Academic vocabulary in the ICLE Table 5.5 Adjectives co-occurrents of the noun ‘example’ in ICLE not found in the BNC Adjective big warning absolute bright cruel present day evident frightening impermissible freq. Which may most probably influence our feeling towards him. 1 1 1 1 1 1 1 1 1 native-speaker informant as forming awkward co-occurrences with example as illustrated in the following sentences: 5. We all know thousands of such manipulative examples.4 Significant adjective co-occurrents of the noun ‘example’ in the ICLE Adjective good extreme above clear striking simple Well-known freq. extreme. Of course. For example a disliked politician will be shot through such a zoom as to expose his ugly bits. Cinderella is a neglected child. every nation defends its own interests before fighting for those of “the group” they joined.4. where the union was imposed by a central power without real approbation of the states and against people’s will. (ICLE-RU) .5. 77 12 8 8 7 6 5 Adjective excellent typical classic interesting numerous outstanding freq. 4 3 2 2 2 1 133 Table 5. that was an overstated example. (ICLE-FR) 5. and once again the step-family is the guilty party. The opposite example is (the former?) USSR. (ICLE-DU) 5. 2 2 1 1 1 1 1 1 1 Adjective manipulative mere model opposite overstated polemic hair raising stirring upsetting freq. so to speak. (ICLE-PO) 5.3. (ICLE-FR) 5. The story of Cinderella is one more impermissible example.2.

6. 1 1 1 1 1 TOTAL 5 . less poverty and more social justice we would not find the same quantity of crime that we find in our society.7 Verb co-occurrent types of the noun ‘example’ in ICLE not found in BNC Left co-occurrents Verb culminate into glide into state plaster with derive write help as appear TOTAL freq. several of these verbs form awkward co-occurrences with the noun example: 5. In a new society made with less inequality. They are listed in Table 5. I can make the example of Naples: here there is everyday an incredible lot of crimes. only 23 per cent of the verb types that are used with example in the ICLE are significant co-occurrents of the noun in the BNC-AC-HUM (see Table 5.6 Significant verb co-occurrents of the noun ‘example’ in the ICLE Left co-occurrents Verb be take give find show serve illustrate provide cite consider TOTAL freq. 162 36 28 10 10 4 3 2 2 1 258 Right co-occurrents Verb be show illustrate concern suggest Suffice freq.134 Academic Vocabulary in Learner Writing Similarly. 119 31 15 2 1 1 TOTAL 169 Table 5. Some 27 per cent of the verb co-occurrents (types) of the noun example in the ICLE do not appear with example in the whole BNC. Like adjective co-occurrents. (ICLE-IT) Table 5.6). 1 1 1 1 1 1 1 1 8 Right co-occurrents Verb say reinforce criticize point out express freq.7.

unequality appears even between people living in towns and villages.Academic vocabulary in the ICLE Table 5. freq. 31 15 Rel. Italian and German learners were shown to underuse stems and rhemes with the verb be. social organization. 2.8 HUM 135 The distribution of ‘example’ and ‘be’ in the ICLE and the BNC-AC- be + example ICLE BNC-AC-HUM 162 (57.66 0.52 (++) 5.8). It appears in all 10 learner corpora (i. freq.45 LogL ICLE BNC-AC-HUM 34.7%) 139 (62.7%) TOTAL 281 223 Rel.1 6.9 The distribution of ‘there + BE + example’ in ICLE and the BNC-AC-HUM there + BE + example Abs.3%) 84 (37. the French army based on conscription. These results differ markedly from those reported in Paquot (2008a) in which French. political and economical stability of the country. Their understanding of the outside world differs. Spanish.3%) example + be 119 (42. for example. Textual sentence stems and rhemes with the verb be are significantly more frequent in learner writing than in professional academic writing (Table 5.8. running after a ball. There is the example of Great Britain where a professional army costs less than. irrespective of the learner’s mother tongue) as illustrated by the following sentence: 5. To glide into an extreme example. This difference may be explained by the fact that the reference corpus used for comparison in Paquot (2008a) was a collection of native-speaker student essays.9 shows that the structure there + be + example is more frequently used in learner writing than in professional academic writing.71 LogL 199.e. freq.76 (++) Table 5. The rules of the road you have to learn to pass your driving license are plastered with examples of children who cross the road unexpectedly.9. (ICLE-CZ) 5. (ICLE-GE) The copular be is the most frequent left and right co-occurrent of the noun example in learner writing. (ICLE-RU) . It originates in dissimilar climate.7. life-style. 24. Table 5.

13. so that I can write about anything I like. each consisting of about six people. (ICLE-CZ) 5.10.11) or in first person plural imperative sentences (Example 5. (ICLE-PO) Interestingly. at the beginning of the project seven committees were established. I can take the example of the ‘Société Générale de Belgique’ which is directed by ‘Suez’. Let’s take the example of a cooker. EFL learners prefer to use the verb take in active structures introduced by the personal pronoun I (Example 5. 5. 5. providing I will have it checked by two or three censors and an condition that I will not write against the government and religion. The verb have is often used in speech with an inclusive we as subject (Example 5. This pattern is very infrequent in ICLE.12). let’s is typically used with the verb take + example (Example 5. erm there we have an example of the attitude that the the council is taking towards the the re-use of employment sites. Er in relation to existing employment sites er and Mr Laycock referred to National Power. Let’s take the example of painting.13 and 5. (BNC-AC-HUM) 5. to investigate one of a range of competing architectural possibilities.17 and 5. (ICLE-FR) As illustrated by Examples 5.14.10). Let us have an example — an extract out of the famous Figaro’s soliloquy: There is a liberty of the press in Madrid now.16.18).16). (BNC-SP) The verb give is the most significant co-occurrent of the noun example in the BNC-SP. 5.14. the verb have and the first person plural imperative let’s are not significant left co-occurrents of example in the BNC-AC but they are in the BNC-SP corpus of spoken language.15.136 Academic Vocabulary in Learner Writing In professional academic writing. was judged to be awkward by our native-speaker informant.15). It is used in questions and first person plural imperative sentences (Examples 5. two patterns that are not found in the BNC-ACHUM despite the fact that the verb is also a significant co-occurrent of . (BNC-SP) 5. I have a good example in my family. learners often use the verb have in the same structures as take to introduce an example.11.12. To take one example. (ICLE-FR) 5. however. the verb take is mainly used in sentence-initial exemplificatory infinitive clauses with the noun example (Example 5. The imperative sentence.

is really a global assumption in search of some refined. the clerk Jankyn. Let me give you some examples. without argument. 2000). Let me give you one example – appaling shots from the war in ex-Yugoslavia that we can see nearly every day. (BNC-AC-HUM) 5.19).4 5. 2004. is superior to forms of knowledge that do not share its methodological characteristics.17. (BNC-AC-HUM) 5. the different forms of the verbs illustrate and exemplify are not all underused in learner writing.1 above shows that the two verbs are underused in their –ed form only. verb co-occurrents of the noun example provide further evidence for the genre-bound nature of phrasemes: the preferred phraseological environment of the noun differs in academic writing and speech (see Biber et al. (BNC-AC-HUM) . 1999. (BNC-SP) 5.23. who. Table 5. the past participle exemplified following a noun phrase (Example 5. Results suggest that EFL learners sometimes select co-occurrences that are more typical of speech.19. Differences in phraseological or lexico-grammatical preferences are often revealed by patterns of overuse and underuse of word forms. Piaget’s claim that thinking is a kind of internalized action. detailed and testable expression. in the Wife of Bath’s Prologue.Academic vocabulary in the ICLE 137 example in academic prose. By contrast. first person plural imperative sentences with the verb give do appear in the ICLE (Example 5. This underuse corresponds to an underuse of the passive constructions BE illustrated by/in (Example 5.5 miles) apart. Luzón Marco. that science.21.22.20) and BE exemplified by/in (Example 5.21). He assumed. (BNC-AC-HUM) 5. Can you give an example when you say that the law is designed? (BNC-SP) 5. The contrast between the conditions on the coast and in the interior is illustrated by the climatic statistics for two stations less than 30 km (18.22) and the patterns as illustrated/exemplified by/in (Example 5. as exemplified by physics. Thus.20. which can be interpreted as further indication of their lack of register awareness.23): 5. reads antifeminist material to her from his book Valerie and Theofraste. The association of this material with the clerk is clearly exemplified by Chaucer’s Wife of Bath’s fifth husband. (ICLE-CZ) In summary. exemplified in the assimilation-accommodation theory of infant learning mentioned above.18.

as you will see. did not start because the murder of archduke Frank Ferdinand.g. e. it sometimes appears in lexico-grammatical patterns that are not found in expert academic writing.’s. they want to be like them. the noun case in point is very rarely used in learner writing.3 below. Film stars are usually very attractive and it’s not a surprise that children want to follow them. I can illustrate that by a real example. in an infinitive clause with the verb take (Example 5. instead of e.27. to take a case in point. EFL learners may also experience difficulty with the meaning of single words and phrasemes. wars always break out for economical reasons.28.29). (ICLE-CZ) It is also frequently used in sentence-initial infinitive clauses (13. (ICLE-SP) 5. A great number of children spend more and more time watching television.26. the first world war. and more specifically with the personal pronoun I: 5.24.28) or determined by a definite article and followed by the verb be and a that-clause (Example 5.25. one has only to mention people’s disappointment when realizing how little value has the time spent at university. To illustrate this point. The case in point is that little children learn how to smoke how to drink how to be cunning and clever and get round the adults. . They take into consideration the behaviour patterns of film stars. . are very diverse. (ICLE-SP) 5.A. I would like to illustrate that by means of some examples which. heir of Autro-Hungary.g. To illustrate the truth of this. however. as an exemplificatory discourse marker (Examples 5. (ICLE-DU) 5. In the worst cases people decide to suicide. that was only the straw that broke the camel’s back.76%) in learner writing.72%): 5. However.2. (ICLE-RU) EFL learners’ phraseological and lexico-grammatical specificities will be discussed in detail in Section 5.S.30 . it would be interesting to compare our situation with the U. 5. For example.138 Academic Vocabulary in Learner Writing The verb illustrate is more often used with human subjects (11. they sometimes use the abbreviation i. (ICLE-FR) As in professional academic writing.29.e. . For example. Professional observers see some even deeper danger in the emerging situation. When used.

Thus soldiers learned mostly bad habits *as [such as] smoking. (ICLE-FR) .e. but many progressive social changes (*i. Another proof will be the role that imagination plays in all the Arts *as [such as] Literature.31. The states mostly tend to solve their politic problems in a peaceful way (*i.] video games).: [e.33. however. (ICLE-CZ) 5.38. (ICLE-RU) Learners also sometimes use as in lieu of the complex preposition such as (Examples 5.32. the adverb namely is also sometimes misused by EFL learners who use it instead of notably or another exemplifier. (ICLE-IT) 5. This new wave of revolting trivial events is all the more worrying since it is linked to a rise of the small delinquance. (ICLE-FR) 5.35.Academic vocabulary in the ICLE 139 to 5.33 to 5.] the split of Czech federation or the unification of Germany). in many cases. because nowadays children play with technological toys (*i.e. One of the examples that makes this point is related to children’s toys. time should be reserved for making children conscious of the fact that there is more to life than the things we see.30.32).e. (ICLE-CZ) 5.g. (ICLE-SP) As illustrated in Example 5. In this essay I would like to show how. both accused of all vices and *namely [(most) notably] of being too lax with those evils. implying a generalized climate of terror and a total mistrust of the citizens towards the police forces and the law. is a synonym of ‘that is’ used to reformulate by paraphrasing or clarifying.g. [e.37). It should be noted.34.37. 5. writing and mathematics. etc.] an increase of individual liberty) may lead to further increase of crime. It might seem absurd. they are so inactive that playing with these toys does not permit physical exercise. and not an exemplifier at all. The abbreviation i. 5. culture and politics can influence this natural inclination. Music and Painting. [e. that this erroneous use is more frequently found in learner populations with Romance mother tongue backgrounds. however. In addition to the familiar subjects *as [such as] reading.g. of course. (ICLE-SP) 5. drinking (if possible) and being lazy in their leisure time. 5.38. others for the young people. There should be particular institutions for those who are mentally alienated *as [such as] the rapists. crime is caused by a predisposition of the individuals and how. in my opinion.. and these toys do not let the children develop their imagination and.e. other factors *as [such as] society. (ICLE-DU) 5.36.

43. In Example 5. for instance.pdf Example: for example. to illustrate http://fr.2.5).paul/BACK itde Survie. built closely together. namely. for example. (ICLE-DU) 5. I described there only some examples from the great number of criminal offences. then. it is no more than a ten minute’s walk to get where you need to be for lectures and seminars. All the academic facilities are ?namely located on the main campus. just as. . like namely (c' est-à-dire) above all (surtout) http://page sperso-orange fr/ Figure 5.40. so many people object to gay marriages and. (ICLE-FR) 5.42. such as. yearn for equality? It is ?namely just equality what gay marriages are about. one example. namely is very often misused in learner writing and it is not always clear what learners mean when they use this adverb: 5. disbelieves. Another explanation for the general overuse of the function of exemplification in learner writing may be that exemplifiers are repeatedly used when they are superfluous. They ?namely create replacement products: they replace the gas in the aerosols and so we have ozone-friendly aerosols. short-sightedness or even nationalistic and xenophobic tendencies.41. such as. Why. st. .39. the logical relation between the two sentences is a causal link that is left implicit while an unnecessary exemplifier is used: 5.wikibooks. More generally. isn’t it? (ICLE-FI) 5.140 Academic Vocabulary in Learner Writing Pour donner des exemples for instance. . at the same time.5 The treatment of ‘namely’ on websites devoted to English connectors This confusion is relatively common. which is not surprising as it is even found on websites supposed to help learners master English connectors (Figure 5. Reluctance to eventually join The Common Market is ?namely caused by fear. redundant or even when other rhetorical functions should be made explicit. The efforts made by the firms are obvious. in particular. Because the campus consists of modern buildings.4. After some years many of those criminals will be set free because of their .43. (ICLE-PO) More examples of semantic misuse are illustrated and discussed in Section 5. inferiority complex.

2.14% of the occurrences of for example and 8. In Section 4. EFL learners’ use of exemplifiers also differs from that of expert writers with respect to positioning. and the noun example and the verbs illustrate and exemplify are used in learner-specific phraseological patterns. who got off with the light punishment. (ICLE-SW) 5.4% of the occurrences of for instance). But there are actually a number of things we all can do that make a difference. For instance. There were a lot of wars due to the religion.4. a Single Use Camera for instance.46. there ought to be information about different ways to save electricity. Let us have a good look at television for example. although this position is rare in academic professional writing (1. compared to the BNC-AC-HUM: 5. It was also argued that the pedagogical relevance of non-AKL items – the preposition .44. For example. They had for example youthful age.2. 1. (Youthful age – by the way in contrast to the punishment of 16 years old boys in our country.6% for for example. Learner corpus data support this claim as all the AKL words that are used to give examples in academic prose present one or more learner-specific diffi culties. in England were recently sentenced two 10 years old boys for murder of a 3 years old boy to the lifelong punishment!) (ICLE-CZ) Section 5. They only want an easy to operate camera.45.47. A sentence-initial position for the adverbials for example and for instance is clearly favoured in the ICLE.5. The adverb notably and the abbreviation e. are semantically misused.g. (ICLE-DU) Aspects of sentence position are dealt with in Section 5.5 will focus on the unnecessary use of lexical items that serve rhetorical or organizational functions as well as on learners’ tendency to clutter up their texts with too many logical devices. the adverbials for example and for instance are predominantly used in sentence-initial position.Academic vocabulary in the ICLE 141 relatively mild punishment.3% for for instance): 5. (ICLE-SP) The two adverbials are also repeatedly found at the end of a sentence in the learner subcorpora (7. England has always been divided according to the kind of religion in which a person believed. I argued that Academic Keyword List (AKL) lexical items and their phraseological patterns should be taught to EFL learners. (ICLE-PO) 5.

2.3 explores the type of phraseological and lexico-grammatical patterns that are found in most learner sub-corpora. ‘comparing and contrasting’.2. – The noun illustration should be specifically taught to upper-intermediate and advanced learners as it is underused in the ICLE. In Section 5.4 discusses patterns of semantic misuse of connectors and abstract nouns.6 illustrates their preference for placing connectors at the beginning of sentences. – The specific lexico-grammatical patterns of case in point should also be taught as this phraseme is repeatedly used in ‘unidiomatic’ patterns. Section An analysis of learners’ use of potential academic words from the . 5.1 focuses on learners’ limited lexical repertoire by examining aspects of over. 1999: 60). The analysis of the ICLE corpus suggests that: – A word of caution is needed against excessive reliance on the preposition like. Section 5.1.5 and Section 5. Limited lexical repertoire Several studies based on one or more ICLE subcorpora have argued that ‘these EFL writers are not equipped with the type of lexical knowledge necessary for the type of writing task they are undertaking’ (Petch-Tyson. Learners’ tendency to clutter their texts with unnecessary connectives is the focus of Section 5.142 Academic Vocabulary in Learner Writing like. Section 5.2. Academic vocabulary and general interlanguage features A comparison of words that serve the rhetorical functions of ‘giving examples’.and underuse. ‘expressing cause and effect’. The pedagogical implications of learner corpus-based findings will be further considered in Chapter 6.2. the characteristics of learner’s lack of register awareness are presented.2. the nouns illustration and case in point and the expressions to name but a few and by way of illustration – depended on whether learners already used these exemplifiers and how they used them. 5.2. ‘expressing a concession’ and ‘reformulating: paraphrasing and clarifying’ in learner and expert academic writing has made it possible to identify six specific areas of where learner English varies from native-speaker academic English.

Table 5. e.0%] 199 [21.9%] 87 [48. by.3%] 49 [56.g. the verbs argue and explain.2%] 34 [18. the verbs be and become and the adjectives difficult and important are very frequent words in general English (relative frequencies of more than 200 occurrences per million words in the whole BNC). 2000: 279). determiners. the proportion of words in the AKL that are overused in learner academic writing is only 21.0%] 277 [29.3%] 21 [28.8%] underused nouns verbs adjectives adverbs other TOTAL 86 [24.3 per cent for adverbs. these highly frequent prepositions are underused in the ICLE. the nouns argument. underused items such as the nouns hypothesis and validity. Meunier.2%] 40 [17. and of are quite representative of the nominal style of academic texts. appears to be more complex than Lorenz’s quote suggests. etc.3%] 33 [44. however. 1998. Many AKL words that appear with a relative frequency of more than 100 occurrences per million words in the whole BNC are underused in the ICLE. Table 5. By contrast.4 per cent . It could be argued that ‘learner usage tends to amplify the high frequencies and diminish the low ones’ (Lorenz 1999b: 59). overused items such as the nouns idea and problem. Not all high frequencies are amplified in EFL learner writing. The largest percentages of overused items are found in nouns and in the ‘other’ category which includes prepositions. a percentage that rises to 52. the verbs exemplify and advocate.8%] . in. the adverbs conversely and ultimately and the prepositions as opposed to and in the light of are much less frequent in English (relative frequencies of less than 30 occurrences per million words in the whole BNC). a fact that can be related to EFL learners’ tendency to avoid prepositional noun phrase postmodification (Aarts and Granger. The picture. However.9%] 16 [18. Key function words such as between. conjunctions.7%] 93 [39.4%] 21 [28.4%] 185 [52. the adjectives likely and significant and the adverbs generally and particularly (in bold in Table 5.1%] 100 [42.8%] 22 [25.1 per cent for nouns and 56. 2006). where 60 per cent of all noun phrases have a modifier (Biber. Table 5. For example.0%] 454 [48.9%] 59 [32.11). Conversely.10 The distribution of AKL words in the ICLE overused no statistical difference 84 [23.10 shows that almost 50 per cent of the words in the AKL are underused in the ICLE.Academic vocabulary in the ICLE 143 Academic Keyword List (AKL) supports this view. difference and effect.11 gives examples of overused and underused AKL words in the ICLE.

prove. several. less. extremely. more. practical. use common. become. The amplification of a restricted set of low frequency words in learner writing may be partly explained by teaching-induced factors. difficult. necessary. consequently. unlikely adequately. in response to. often. basis. assumption. cause. secondly. increasingly. propose. an. misleading. hence. reality. contrast. while its much less frequent synonym. position. special. view. upon. benefit. yield adequate. define. previously. depend. originally. important. emphasis. largely. possibility. subsequently. different. possible. readily. in terms of. reflect. allow. evidence. idea. by.6). assess. choice. due to. disadvantage. change. similarly. outcome. in the light of. obvious. the verbs participate and solve. participate. derive. the. in relation to. specify. appropriate. theory. latter. subsequent. particularly. extensive. detailed. perspective. solve. prior to. comprise. conduct. prime. highlight. as opposed to. to. sense. irrespective of genre. effect. treat. assert.11) are overused although they appear with frequencies of less than 50 per million words in the BNC. relative. each. study. describe. criterion. therefore other according to. some. successfully. advocate. suggest. this The preposition despite is underused. degree. particular. only. true. useful underused addition. from. reason. substantial. hypothesis. assume. validity adopt. specifically. emphasise. solution. generally. subject to. contribute. influence. between. avoid. parallel. In addition. concept. risk. explain. fact. major.11 the ICLE Academic Vocabulary in Learner Writing Examples of AKL words which are overused and underused in overused nouns advantage. or. Words such as consequently. is overused in learner writing (Figure 5. its. conversely. consider. indicate. comparison. moreover. issue. choose. extent. difference. moreover and secondly usually appear in the long and . given that. influence. form. consist. similar. summary. provided. relatively. enhance. be. aim. improve. cite. real. notably. exemplify. concern. however. theme. inherent. ultimately although. stress aim. representative. the complex preposition in spite of. essentially. problem. argue. bias. of. consequence. same. deal. many. primarily. likely. words such as the noun disadvantage. comprehensive. increase. develop. and the adverbs consequently and moreover (underlined in Table 5. conclusion. including. potentially. during. critical. note. unlike. main. effectively. interesting. ensure. explicit. for. mainly. create. which verbs adjectives adverbs also. contrast. example. rather than. examine. reveal. exist. scope. argument. than. in. despite. because.144 Table 5. significant. especially.

As argued by Baayen et al. Another tentative explanation may be that EFL learners do not amplify any high frequencies words except those that are common in speech. issue.1). general. indicate. the noun example and the prepositions like and such as. which is compared with that of expert writers in Appendix 1. however. This situation may be compounded by problems of semantic misuse as will be discussed in Section 5. assume. but semantically specialized. and vague words where more precise vocabulary should be used (Granger and Rayson. words probably stems from learners’ tendency to rely on all-purpose. ‘the complexity of the frequency variable has been underestimated’ and it may be that more emphasis should be placed on the explanatory potential of spoken frequency counts. The underuse of some frequent. it was shown that. (2006). and particularly are quite frequent in general English (as represented by the whole BNC). When corpus comparisons are based on the total .2. 1998. but their frequencies are significantly less when the conversation component is analysed separately. Broadly speaking.6 The use of ‘despite’ and ‘in spite of’ in different genres undifferentiated lists of connectors provided in EFL/EAP teaching materials (see Section 6.4. Underused words such as argument. Petch-Tyson. In Section 5. The same conclusion holds for learners’ use of cause and effect lexical items.Academic vocabulary in the ICLE 250 200 150 100 50 0 Academic writing News Fiction Speech Learner writing 145 despite in spite of Figure 5. learners overuse logical links signifying cause and effect in their argumentative essays. 1999). appropriate. They rely instead on a restricted lexical repertoire mainly composed of the adverbials for example and for instance. EFL learners make little use of a number of EAP-specific lexical devices such as the verbs illustrate and exemplify or the adverb notably. although they generally overuse exemplifiers.1. This overuse does not. affect all grammatical categories.

01) in ICLE than in BNC-AC-HUM. that usage is bound to include a number of instances of over-extension. Several of the overused lexical items are massively overused in learner writing.or underuse. prepositions and conjunctions. or even accounted for in the standard grammars.146 Academic Vocabulary in Learner Writing Table 5.12). In other words. Such “simplification” is one of the most frequently cited features of learner language’ (Lorenz. EFL learners prefer to use prepositions.12 Two ways of comparing the use of cause and effect markers in the ICLE and the BNC Absolute frequency / total number of words Absolute frequency / total number of ‘cause and effect’ markers −− −− // // ++ ++ nouns verbs adjectives adverbs prepositions conjunctions // // // ++ ++ ++ Legend: ++ significantly more frequent (p < 0. when frequencies are compared to the total number of cause and effect lexical items. He argued that ‘if a linguistic element is used as an all-purpose wild card.5 per cent of the ‘cause and effect’ . This means that. The overuse of conjunctions largely stems from learners’ marked preference for because. // no significant difference between the frequencies in the two corpora number of running words in each corpus. compared to expert writers. while nouns and verbs are underused (Table 5. The categories of nouns. adverbs to express a cause or an effect. and tend to avoid nouns and verbs. The adverb so represents 11. which represents 19. verbs and adjectives do not display significant patterns of over. 1999b: 60–1). but which are nevertheless observed by the native speakers.01) in ICLE than in BNC-AC-HUM.13 shows that. only prepositions and conjunctions are significantly overused. Lorenz (1999b) examined the use of causal links in essays written by 16-to-18-year-old German learners and described the marked overuse of the conjunction because as ‘wild-card use’. Table 5. even though EFL learners prefer to use prepositions. to a lesser extent. not all individual connectors are overused in learner writing. conjunctions and adverbs to express cause and effect. conjunctions and. it can be expected that learners may disregard target-language restrictions which are not that obvious. the overuse seems to be generally attributable to adverbs.9 per cent of all cause and effect markers in the ICLE. − − significantly less frequent (p < 0. By contrast.

yield. hence. Other examples of ‘lexical teddy bears’ (Hasselgren. give rise to. provoke. on account of 0 10 [100%] 11 [100%] 2 [40%] conjunctions TOTAL because. actually. Hyland and Milton (1997) reported similar findings: Cantonese learners used a more limited range of epistemic modifiers. this/that is why 12 [21%] 3 [60%] for. due to. derive. induce.13 The over. with the ten most frequently used items (will. In their study of expressions of doubt and certainty. so 3 [27%] prepositions because of. implication 13 [76%] generate. origin. in (the) light of 11 [100%] 17 [100%] adjectives 0 1 [50%] responsible (for) 2 [100%] 4 [40%] adverbs consequently. on the grounds of. stem. result in.5 .Academic vocabulary in the ICLE 147 Table 5. trigger 1 [50%] consequent 4 [40%] accordingly. may. lead to underuse TOTAL 2 [19%] nouns root. owing to. usually. emerge. as a result. as a consequence.and underuse by EFL learners of specific devices to express cause and effect (based on Appendix 1) overuse no statistical difference 4 [37%] cause. thus.2 per cent of those in expert writing. result 3 [18%] bring about. know. so that. effect. in fact. follow. 1994) or ‘pet’ discourse markers (Tankó. factor. consequence 1 [6%] cause verbs 5 [45%] source. arise. thereby 2 [18%] in view of. contribute to. as a consequence of. outcome. in consequence of. thanks to 2 [20%] therefore. always. prompt. 2004) are the prepositions because of and due to. reason. on the grounds that 5 [100%] 56 [100%] 16 [29%] 28 [50%] lexical items used by learners while it only accounts for 7. in consequence 6 [54%] as a result of. think. and probably) accounting for 75 per cent of the total. would.

the degree of underuse varies significantly. the concept of ‘labelling’ explained in Section 1. and unlike) account for 59 per cent of all underused lexical items in the comparison and contrast category. arise. EFL learners tend to rely heavily on a restricted set of greatly overused adverbs. Nouns and adjectives (e.2) as well as commonly misused expressions such as on the contrary (see Section 5. These findings are not restricted to EFL learners: although they may become fluent in English conversational discourse. resemblance. in the same way) (see Section 5. Logical links can also be provided by nouns (cf. similarity. Comparisons and contrasts are generally underused in learner writing.4).3). Unlike the cause and effect lexical items. but 64. contrast. emerge and stem from).148 Academic Vocabulary in Learner Writing On the other hand. An analysis of the lexical items which serve to express a comparison or a contrast in academic prose shows that the rate of underuse is also quite high in this function.g. this may be explained by teachinginduced factors. the verbs induce. however. The rate of overuse is relatively low.2. While underuse was found in all grammatical categories. English as a Second Language (ESL) speakers have also been reported to ‘continue to have a restricted repertoire of syntactic and lexical features common in the written academic genre’ (Hinkel. Table 5. but once again overused items include words and phrasemes that are more frequent in speech (e.14 provide useful . effect and implication. As will be discussed in Section 6.14 shows that almost half of all comparison and contrast markers are underused. Tables 5. prepositions or conjunctions to establish textual cohesion. which often account for a large proportion of the lexical strategies used to serve a specific rhetorical or organizational function in expert academic prose.3 per cent of them are underused in the ICLE (e. As with cause and effect lexical items. as lexical cohesion has been largely neglected in teaching materials (textbooks and especially grammars).g.2. 2003: 1066). verbs and adjectives. Nouns and verbs constitute a large proportion of the possible ways of expressing a cause or an effect in academic prose.g.13 and 5. the nouns source. where the focus has generally been on adverbial connectors. These cohesive devices.1. yield. result in. do not seem to be readily accessible to upper-intermediate/advanced EFL learners. 50 per cent of the lexical devices which serve to express cause or effect in expert writing are underused by learner writers. similar. as lexical cohesion has generally been neglected in teaching materials. distinct. the proportions varied significantly. look like. In summary. overused comparison and contrast word do not compensate for the underused ones. This is not particularly surprising.

reversely. differentiate 4 [22%] alike. on the one hand. opposite. differ. by/in contrast. differentiation. while6 0 4 [100%] 9 [100%] 21 [100%] 18 [100%] underuse 10 [67%] resemblance. CONJ compared with/to conjunctions 3 [100%] TOTAL 11 [13. in contrast to/with. difference. identical. comparatively 3 [33%] in parallel with. by/in comparison with + erroneous expressions 0 1 [25%] other expressions as … as. distinctiveness. + erroneous expressions 2 [22%] prepositions like. contrary to 1 [33. likewise. distinguish. the contrary.9%] 31 [39. the same.Table 5.14 The over. analogy. analogous. parallel. comparable. differently.67%] as. 10 [48%] analogously. correspondingly. quite the contrary. correspond. the reverse 2 [22%] 9 [100%] 15 [100%] TOTAL nouns verbs look like. distinguishable. identically. distinct. by/in comparison. contrast. contrariwise. contrast 12 [67%] similar.8%] 79 . as opposed to. reverse parallel. compared with/to. conversely. common. on the other hand. different 4 [19%] adverbs in the same way. differing. unlike 7 [33%] similarly. versus 2 [66. contrasting. parallel. distinctively 4 [44%] unlike. contrary. comparison. the opposite 2 [22%] 5 [56%] resemble. on the contrary.33%] whereas 3 [75%] in the same way as/ that. as against.and underuse by EFL learners of specific devices to express comparison and contrast (based on Appendix 2) overuse 0 no statistical difference 5 [33%] parallelism. distinctive. contrastingly. similarity. by way of contrast.2%] 37 [46. distinction. compare 2 [11%] adjectives same. parallely.

2. 5.51. So they want to get rid of the military service. (ICLE-RU) 5. According to Crystal it has little further potential ouside Spain. it will be shown that the limited nature of EFL learners’ lexical repertoire also stems from a restricted use of the phrasemes and lexico-grammatical patterns typically found in expert academic prose. Lack of register awareness Many learner corpus-based studies have reported on EFL learners’ lack of register awareness (e. have often focused on learners with the same mother tongue background. In the ICLE. the stem I am going to talk about to introduce a new topic. different thoughts and emotions. Meunier. (ICLE-DU) 5.and underused AKL single words and mono-lexemic units used to perform specific rhetorical functions. Granger and Rayson.2. But practically everybody is able to dream. Of course. what I want to demostrate is that a good way of making politics can cut the roots to crime. (ICLE-IT) . too. 1998. Many people who are in this situation think that this is a waste of time: you lose an entire year.150 Academic Vocabulary in Learner Writing information about learners’ particular needs. Lorenz. the adverb though to introduce a concession.49. 2000.50. though.3. These studies.48 to 5. there are different people with different concepts of happiness. In this essay I am going to talk about the link between crime and politics. Spanish holds an important position in South America and increasingly so in the United States. 2006). 5. In this section.48.g. Section 6. Examples 5. and the adverbial all in all which is used to ‘show that you are considering every part of a situation’ (Longman Dictionary of Contemporary English (LDOCE4)).2. however. (ICLE-FI) 5. Ädel. most rhetorical functions are characterized by the overuse of at least one lexical item that is more typical of speech than of expert writing (Table 5. The large-scale study undertaken here allows for a more systematic description of register awareness.52 illustrate overused lexical items that are more frequent in the BNC spoken component than in the BNC-ACHUM: the adverb so to express an effect. 1998. the adverbial of course to express certainty. 1999b.15).3 will discuss how they can be used to inform pedagogical material. the breadth of EFL learners’ lexical repertoire has been examined in terms of the proportion of over. In Section 5. Altenberg and Tapper. by exploring the way EFL learners with different mother tongue backgrounds use academic vocabulary.

. by no means. Thanks to them anyone willing to broaden his/her general knowledge of the world has an easy access to useful information.52. and Turkish) from the second version of the International Corpus of Learner English (ICLEv2) (Granger et al.15 Speech-like overused lexical items per rhetorical function Rhetorical function Exemplification Cause and effect Speech-like overused lexical item like thanks to so because that/this is why look like like the (sentence-final) adverb though sentence-initial and the adverb besides I think to my mind from my point of view it seems to me really of course absolutely maybe I would like to/want/am going to talk about thing by the way first of all 151 Comparison and contrast Concession Adding information Expressing personal opinion Expressing possibility and certainty Introducing topics and ideas Listing items Reformulation: paraphrasing and clarifying Quoting and reporting Summarizing and drawing conclusions say all in all 5. all positive or good for us. The corpus totalled around 1.15 in the ten learner corpora used here as well as in four L1 sub-corpora (Norwegian. All in all.5 million words. (ICLE-PO) Gilquin and Paquot (2008) examined the use of some of the lexical items listed in Table 5. Chinese. there are many ways in which mass media affect our approach to reality and they are. Japanese.Academic vocabulary in the ICLE Table 5.. 2009).

. For example. Spanish and Swedish learners’ heavy reliance on I think to express their personal opinion is reported by Granger (1998b). As shown in Table 5. Our findings support Lorenz’s (1999b: 64) statement that there is ‘mounting evidence that text-type sensitivity does indeed lie at the heart of the NS/NNS numerical contrast. of course and certainly) are even more frequent in learner writing than in speech. Different EFL learner populations. Neff et al. Studies in contrastive rhetoric (e. however. relative frequencies range from 17. However some of these items (so expressing effect. Connor. This huge difference may be partly explained by L1 influence. (2007) and Aijmer (2002). I would like/want/am going to talk about. Using the ICLE. It remains to be seen. Although all L1 learner populations overuse the adverb maybe when compared to the BNC-AC-HUM. whether lack of register awareness is a typical feature of EFL learner writing or whether it is a more general characteristic of novice writing. French and Swedish learners’ overuse of of course is highlighted by Narita and Sugiura (2006). really. The overuse of several of these speech-like lexical items has been highlighted in a number of studies focusing on specific L1 learner populations. and are therefore likely to be developmental or teaching-induced. Table 5. Lorenz (1999b) discusses the marked overuse of the conjunction because and the adverb so in German learner writing.57 occurrences per 100. 1998) have shown that features of writer visibility in academic prose may differ markedly across languages. Granger and Tyson (1996) and Altenberg and Tapper (1998).4.152 Academic Vocabulary in Learner Writing We compared the frequencies of speech-like lexical items in learner writing with their frequencies in the 10-million word spoken component (BNC-SP) and the 15-million word academic sub-corpus of the British National Corpus. it seems to me.16 shows that relative frequencies differ widely across L1 populations. This issue will be touched upon in Section 5. by the way and though in Figure 5. Another example is EFL learners’ use of I think.g. absolutely. Chen (2006) reports on the overuse of besides in Taiwanese student writing. Vassileva. do not use speech-like lexical items similarly. however. my results suggest that these features are often shared by a large proportion of the learners investigated. French.000 words in the Swedish one (ICLE-SW).7). Japanese. 1996.29 occurrences per 100.000 words in the Polish learner sub-corpus (ICLE-PO) to 143.17. irrespective of their mother tongues. definitely.’ They show that the relative frequency of these speech-like lexical items in learner writing is often situated between their frequency in academic prose and in speech (see the bar charts for maybe. which is overused by all L1 learner populations while showing marked differences across learner L1 sub-corpora.

5million words) Speech: British National Corpus. spoken component (10million words) Figure 5. learner writing and speech (based on Gilquin and Paquot. academic component (15million words) Learner writing: ICLEv2 (14 L1s.7 The frequency of speech-like lexical items in expert academic writing.Academic vocabulary in the ICLE 350 300 250 200 150 100 50 0 1200 1000 800 600 400 200 0 153 Frequency of maybe (pmw) 40 35 30 25 20 15 10 5 0 Frequency of so expressing effect (pmw) 20 15 10 5 0 Frequency of it seems to me (pmw) Frequency of I would like/want/am going to talk about (pmw) 2000 1500 1000 500 0 really of course certainly absolutely definitely Frequency of amplifying adverbs (pmw) 120 40 30 20 10 0 100 80 60 40 20 0 Frequency of by the way (pmw) Frequency of through at the end of a sentence (pmw) Academic writing: British National Corpus. 2008) . 1.

16 The frequency of ‘maybe’ in learner corpora relative freq.61 72. The phraseology of academic vocabulary in learner writing In this section.2.37 13.000 words ICLE-IT ICLE-GE ICLE-DU ICLE-CZ ICLE-SP ICLE-SW ICLE-FI ICLE-FR ICLE-PO ICLE-RU BNC-AC-HUM 48.01) than in the BNC-AC-HUM 5. per 100.13 101.17 The frequency of ‘I think’ in learner corpora relative freq.14 ++ ++ ++ ++ ++ ++ ++ ++ ++ ++ Legend: ++ frequency significantly higher (p < 0.000 words ICLE-SW ICLE-IT ICLE-RU ICLE-CZ ICLE-FR ICLE-GE ICLE-SP ICLE-FI ICLE-DU ICLE-PO BNC-AC-HUM 143.88 32.154 Academic Vocabulary in Learner Writing Table 5.28 31.74 20.3.59 55.13 32.93 ++ ++ ++ ++ ++ ++ ++ ++ ++ ++ Legend: ++ frequency significantly higher (p < 0.77 17. I first present the major results of an analysis of recurrent word sequences in EFL learner writing.01) than in the BNC-AC-HUM Table 5.34 16.79 6. per 100.26 1. I focus on aspects of overand underuse of word sequences that include AKL words before discussing learner-specific clusters that are not found in professional academic prose.34 35.57 134.7 94.06 121.87 51.11 66.21 24. Learner writing is also typically recognizable by a whole range of cooccurrences that differ from academic prose in quantitative and qualitative .18 38.

quite (e. 1998b: 155). significant number. Granger (1998b) suggests that the use of these sequences ‘could be viewed as instances of what Dechert (1984: 227) calls “islands of reliability” or “fixed anchorage points”. different points.g. crucial importance. .or underused in learner writing. I will discuss.18 shows that EFL learners overuse adjective + noun sequences with ‘nuclear’ adjectives (see Section 1. real (e. real problem. the problem is that). different reasons) and big (e.g. The comparison between the ICLE and the BNC-AC-HUM was performed with the Keywords option of the software tool WST4.g.g. quite clear) and very (e. I illustrate this with a comparison of the co-occurrents of the noun conclusion in academic and learner writing and examine EFL learners’ phraseological infelicities and lexico-grammatical errors.g. more generally. EFL learners rely instead on a restricted set of clusters which they massively overuse (e. relatively few. different problems. 1998b: 156). because of the fact7). important (e. big problem) to the detriment of more EAP-like phrasemes such as extensive use. Table 5. they overuse adverb + adjective/adverb /conjunction sequences with highly frequent adverbs such as mainly (e. more and more.Academic vocabulary in the ICLE 155 terms. prefabricated formulaic stretches of verbal behaviour whose linguistic and paralinguistic form and function need not be “worked upon”’ (Granger. almost entirely. main reason. lesser extent and wide variety. i. important factor). real value). great importance). in order to. it depends. significantly different.g.g. Similarly. main problem). The results show that learner writing is characterized by a marked underuse of a large proportion of the 2-to-5 word sequences that include AKL words and that are typically used to serve specific rhetorical and/or organizational functions in academic prose. important role. from my point of view. different (e.g. main cause.g. important question. integral part. The foreign-soundingness of EFL learner writing also stems from learners’ overuse of AKL words in clusters that are not typical of the particular genre of academic prose but are more frequently used in speech or more informal types of writing (e. great (e. very important) but make little use of phrasemes such as readily available. for example. people claim that. particularly interesting.e.g. main reason. highly significant and precisely because. it can also be due to an excessive use of them’ (Granger. This is also consistent with the author’s statement that ‘while the foreignsoundingness of learners’ productions has generally been related to a lack of prefabs. central issue. closely associated.1) such as main (e. mainly because). great number.1.g. An analysis of word sequences in EFL learner writing The results presented in this section are based on an analysis of 2-to-5 word sequences that are over.

listed above. with the exception of. of great importance.18 Examples of overused and underused clusters with AKL words Overused clusters for example. mainly because. on average. it is very important to. it is very difficult. similar to that of. in certain respects. to the advantage of. they suggest. the absence of. at any rate. in so far as they. major source. on the basis of. in his view. can be related to. with the result that. somewhat different. he remarks. in particular. in the presence of. more and more. as a consequence. described as. it seems likely that. main problem. it is also true. little evidence. negative consequences. certain respects. to a great extent. good idea. take into consideration. in view of. I consider. be defined in terms of. it is more likely that. in relation to. by showing that. the fact is that. provide evidence. it is possible that. this need not. it has been suggested that. were subject to. good use. was by no means. it means that. extensive use. on the assumption that. affect our approach. to the extent that. a wide variety of. it was claimed. considerable degree. as a conclusion. extent to which. still further. inferred from. there are more and more. when compared with. main reason. precisely because. to answer this question. by comparison. great part. perhaps because. is ultimately. it was difficult to. the hypothesis that. therefore I. radically different. relatively few. it seems likely. the implications of. important factor. crucial importance. it is obvious that. in my view. provides us with. high proportion of. from my point of view. it depends. as a matter of fact. great importance. general principles. but it is true that. inherent in. for instance. this is not the case. good example. despite the fact. an account of. he concludes. pay attention to. I can. wide variety. more significantly. reports that. it follows that. prevents us from. real value. suggestion that. the total number of. central figure. discussed in. he cites. quite clear. it is assumed that. another important. reported by. take into account. crucial role. conclusion I. may well have been. the extent to which. because of the fact. be explained in terms of 2-word clusters 3-word clusters 5-word 4-word clusters as a result. important question.Table 5. one of the most important . may suggest. they argued. in order to achieve. a great number of. what appears. a theory of. have problems. is described in it may be that. may say that. are likely to be. closely associated with. highly significant. it is very difficult to. important part. an attempt to. allows us. are concerned. can choose. opportunity of. significantly different. integral part. a similar. as noted above. great amount. particular attention. in the absence of. it would appear that. it could be argued that. it is a fact that. was effectively. almost entirely. much emphasis. in practice. in the belief that. is the fact that. different points. different problems Underused clusters by contrast. in the presence of. real problem. be ascribed to. no reason to suppose. we can. aim of this. as much as possible. best solution. to the effect that. it is true that. allowing for. more difficult. final analysis. readily available. central issue. it does not follow. it is necessary for. by reference to. it is high time. and therefore. might have been expected. the edge of the. the problem is that. because I. would seem to be. big problem. people claim that. suggested above. was probably. there are also people. his method in terms of. as an attempt to. very important. the view. as distinct from. partly because. it is hardly surprising that. various forms of. high degree of. important to. totally different. the view that. a concern with. the issue of. different reasons. this suggests that. a considerable degree. may have been used as in the case of. it is unlikely that. more generally. far as I am concerned. consistent with. with the exception of. absolutely necessary. I will discuss. main cause. only because. advantages and disadvantages. the immediate aftermath of. to this extent. it is worth noting that. different way. more or less.

be ascribed to. it was claimed. 1998. but still display over. it is a fact that. but are highly underused in learner writing. Word sequences used as self mentions are also much more frequent in learner writing than in academic prose (Aijmer. By contrast. Verbs may have similar frequencies as lemmas in learner writing and academic prose. their past participle form. For example. Lorenz. e. they suggest. it is very important to) and boosters (e. a difference which can be related to the more intertextual nature of professional academic texts. might have been expected. it is possible that. inferred from. may have been used. 2003. Neff et al. This is consistent with Granger and Paquot’s (2009b) finding that past participles are the most frequent verb forms in academic prose.Academic vocabulary in the ICLE 157 The results also seem to support the widely held view that EFL learners’ academic writing is characterized by ‘firmer assertions.. Ädel. but it is true that. when compared with. it is very difficult to. I consider.g. However. EFL learners state propositions more forcefully and make a more overt persuasive effort: they overuse communicative phrasemes that serve as attitude markers (e. it may be that. listed above. discussed in and reported by.g. suggested above. he remarks. and be explained in terms of. it is obvious that). be defined in terms of. the 4-word clusters can be related to. 2009b). . the lemma provide is underused in learner writing compared to expert writing. we can. 1998. 1997: 193) (see also Petch-Tyson. the 3-word clusters closely associated with. use of other forms of the verb does not differ significantly in the two corpora. I can. Similarly. as noted above. some verbs are under. academic writers use more clusters with third person pronouns with an evidential function. EFL learners also underuse a whole set of word sequences involving the –ed form of verbs. and more precisely. it is unlikely that. in my view. and it would appear that. Examples include therefore I. and the 5-word clusters it has been suggested that.19 shows that the picture can even be more complex: verb forms may be overused in some specific lexical bundles. it seems likely that. it could be argued that. 2004a). they underuse the 2-word clusters described as. they underuse hedges such as it is (more) likely that. it is assumed that.g. For example.or overused as lemmas without this affecting all forms of the verb.or underuse of some forms (Granger and Paquot. and from my point of view. De Cock. more authoritative tone and stronger writer commitments when compared with native speaker discourse’ (Hyland and Milton. his method. but an analysis of word forms indicates that this only applies to provided. provides us with. The lemmas do not differ significantly in their use. Conversely. Table 5. she cites. is described in. Examples of AKL verbs following this pattern are differ and discuss. differ is underused in its –ing form while discuss is overused in its unmarked form (discuss) and underused in its –ed form. 2006). I will discuss. because I. 2002.

by allowing. to provide a depend (++) depends (++) depending (++) depended (− −) differ (//) differing (− −) discuss (//) discuss (++) discussed (− −) tend (//) tend (++) tended (− −) provide (− −-) provided (− −) tend to. be allowed to. to discuss. allow it. not allowed to. affect us. far as I am concerned depends on. depend upon the. allowed him. to provide an. affect our approach. in discussing. people tend.and underused in learners’ writing. are not allowed. allow them to. can provide. depending on. as I am concerned. concerning the. to provide. concerned about. been concerned. affects the. people tend to. allow them. depended on. discussed in chapter they tended to. allow that. we tend. is concerned with the depending upon.01) in ICLE than in BNC-AC-HUM. which allowed. allows us to. differs from the was discussed. concerned with the. media affect our approach to allowed to. provide an. will depend. to allow for concern (++) concerning (++) was concerned. depending on the will discuss. been concerned with. are allowed. provides that. not affect the allow (++) allowed (++) allow them. and discussed. already discussed. allows for. provide them with Legend: ++ significantly more frequent (p < 0. provides us with.158 Academic Vocabulary in Learner Writing Table 5. not allow. provide evidence. be allowed. depends upon the. provide the. it allows. have tended to. I am concerned. they tend to provides us. by comparison with expert academic writing Lemmas and their word forms affect (++) affect (++) affects (++) Overused clusters Underused clusters affect our. not allowed. provide a. affect our approach. was concerned with. are allowed to.19 Clusters of words including AKL verbs which are over. allowing for. allowed him. − − significantly less frequent (p < 0. mass media affect our. much depends. mass media affect. provide them.01) in ICLE than in BNC-AC-HUM. provides a. concerned to. media affect our approach. depended upon. am concerned. depends on the. was to provide. it depends. we are concerned. // no significant difference between the frequencies in the two corpora . provides an. are concerned. media affect our. provide us. discussed in. to allow. I will discuss was affected. it concerns. are not allowed to is concerned. and tended to. will depend on differed from. they tend. it depends on. discussed below. tended to be might provide. affect our approach to reality. provide us with. allows us. concerned with. has tended to. we tend to. media affect.

is why. in spite of. far as I am concerned. my opinion. some people say that. of course. of course. I believe that. I want. I don’t agree with.4 on semantic misuse). I would like. from my point of view. I do not agree with. I believe. to sum up. it is easy to I do not think that. means that. I think. I would like to say. and of course. people say.g. by the way. thanks to. we look. there are a lot of (see Section 5. a look at.20. are more and more. people feel. look at. everybody knows that. I guess. it seems to me that. I am afraid that. if we. people often.2. due to the fact that.g. we think. said that. They include: – word sequences that are more frequently used in speech. it is impossible to. as a matter of fact. the sequence as far as I am concerned which is repeatedly used to express a personal opinion in the ICLE. I think that. in order to. to my mind. people believe that. just imagine. quite sure. no matter.2. on the other side (see Section 5. even worse. I agree. . people think. all kinds of. we want. let them. first of all. last but not least. we must. e. e. there are. we have to. at all. we can say that. when I. I would say. in fact. really think. I am sure that. the verb form concerned is overused in as I am concerned and concerned about but underused in been concerned withor we are concerned. I am convinced. It makes it possible to uncover a whole range of words and word sequences that are not typical of academic prose but which are nevertheless used by EFL learners to organize scientific discourse and build the argument of academic texts. I must. I want to say. or maybe. that is why. why we. I would like to. we get. think twice. on the contrary. I hope. helps us in my opinion.g. but I. I do not think so 3-word clusters 4-word clusters 5-word clusters while being underused in others. if you. – sequences that exist in English but are very rare in all types of discourse. sure that. but if we look. many things. I agree that. – sequences that are not used in English to establish the logical link intended by the EFL learner. A keyword analysis of recurrent word sequences is indispensable if we want to build up a full picture of all the possible lexical realizations of rhetorical functions in learner writing. it means that.20 Examples of overused clusters in learner writing Examples 2-word clusters 159 in sum. I think that. so why. many people think on the one hand. opinion is.2). in this essay I. EFL learners overuse the sequences it allows and allows us to and underuse the EAP sequence allows for. everybody knows. e. Similarly. on one hand. I will try. I really. if we want. we all know. let us. there is. there are a lot. people believe.Academic vocabulary in the ICLE Table 5. but at the same time. For example. people think that. very serious. instead of. Examples of learnerspecific sequences that do not include an AKL word are given in Table 5.

This includes as a conclusion but not in conclusion.160 Academic Vocabulary in Learner Writing – ‘unidiomatic’ sequences such as as a conclusion used as a textual phraseme to introduce a conclusion (see below for a co-occurrence analysis of the noun conclusion in the ICLE). as discussed above. in contrary to that are used to express a contrast in EFL learner writing.2. Les fiches essentielles du Baccalauréat en anglais (published in 2008 by Clairefontaine) give a list of linking words that French students are encouraged to use in the English test of the ‘Baccalauréat’ (the final secondary school examination which gives successful students the right to enter university) to ‘enrich their essay and give more clarity to their argumentation’. For example. (Mukherjee and Rohrback. the latter being notoriously overused by German learners of English at university level as well.1.2 per cent of the concluding textual phrasemes involving the noun conclusion in the ICLE. the most frequent phrase is no longer in conclusion. in the contrary. but as a conclusion. EFL learners’ overuse of sequences that are rarely used by native-speakers (such as as far as I am concerned or last but not least) or ‘unidiomatic’ sequences (such as as a conclusion) may be partly explained by poor teaching materials and/or the influence of their mother tongue. it was shown that EFL learners manifest a marked preference for a restricted set of single words and mono-lexemic phrasemes to express logical links. In a longitudinal study of German learners’ use of the noun conclusion. They also use learner-specific functional equivalents of these markers such as the sequence as a conclusion instead of in conclusion.8 The rare expression as far as I am concerned is also given as a key expression for voicing one’s own opinion. This certainly is a problematical development because in conclusion is much more frequent and idiomatic than as a conclusion. Preferred co-occurrences in EFL learner writing In Section 5. 2006: 224) This development may be related to the increasing use of the internet for study purposes and of the type of teaching materials available on this channel. This learner-specific word combination represents 39. Mukherjee and Rohrback (2006) commented that the sequence as a conclusion is gaining ground in learner writing to the extent that it is even more frequent than in conclusion in the more recent corpus they use: Interestingly. Another example of a learner-specific logical . by the contrary. – erroneous sequences such as in contrary.

I is generally followed by the modal would to produce the word sequence in conclusion. Another example is learners’ use of the noun conclusion.8 shows that the textual phraseme in conclusion (or one of its learner-specific functional equivalents) is very often directly followed by the personal pronoun I in the ICLE. When tokens are analysed.8 Phraseological cascades with ‘in conclusion’ and learner-specific equivalent sequences marker is on the other side which they use instead of on the other hand to compare and contrast (see Section 5. However almost half of the verb co-occurrent types (46. The sequence in conclusion.8 per cent of verb co-occurrent types are significant cooccurrents of the noun conclusion in the BNC-AC .Academic vocabulary in the ICLE 161 In conclusion 59 As a conclusion 40 As conclusion 3 I 37 (36%) would 21 (20. it was shown that mono-lexemic phrasemes such as for example have their own phraseological patterns in academic prose. and concluding) in Swedish learner writing.9 Figure 5. Some 30. chains of shared collocates)’ (Gledhill.21 lists the verb co-occurrents of the noun conclusion in the ICLE. The sequence in conclusion.7%) mention speak about reiterate quote 6 2 1 1 1 1 1 Figure 5. I would. metadiscourse items that refer explicitly to the writer and/or reader.2. EFL learners use AKL nouns and verbs in different lexico-grammatical or phraseological patterns than professional writers.e. these do not seem to be readily available to EFL learners. In Section 4. anticipating the reader’s reaction. This is consistent with Ädel’s (2006) finding that personal metadiscourse. in turn.1.6%) like to say emphasize tell 13 (12. ‘collocational patterns which extend from a node to a collocate and on again to another node (in other words. which.1.4 below for more details of learners’ use of on the other side). very often introduces the sequence like to. who tend to produce their own phraseological ‘cascades’.2. I would like to either introduces the verb say or another verb of saying such as tell or mention. However. This has already been illustrated by learners’ use of the noun example and the verbs illustrate and exemplify in Section 5. arguing. serves a wide range of rhetorical functions (including exemplifying. the percentage of verb co-occurrents . Table 5.2%) used in the ICLE do not appear in the BNC-AC. 2000: 212). i.

in ICLE Statistically significant co-occurrent in BNC-AC − − − ** − − ** − ** − ** − − ** ** − − − − − − ** ** Appearance in the BNC-AC conclusion as subject + verb Freq.21 Verb co-occurrents of the noun conclusion in the ICLE Freq. in ICLE Significant co-occurrent in BNC-AC ** − − ** − − − Appearance in the BNC-AC √ √ X √ X √ X Verb + conclusion as object add up to apply approach arrive at bring bring sb to come to *come into confirm contain draw *draw up end with escape express find gather get give have influence jump to lead to 1 1 1 5 1 2 52 1 1 1 25 1 1 1 1 2 1 1 1 1 1 2 4 x x x √ x √ √ x √ x √ x x √ √ x x √ √ √ √ √ √ emerge arise contain be come need bring sb to 1 1 1 23 1 1 1 Academic Vocabulary in Learner Writing .162 Table 5.

x the co-occurrent is not found in the BNC-AC 163 .leave sb with look for make point to put put forward reach write as TOTAL 1 1 11 1 1 1 3 1 − − − * − − ** − 128 tokens (32 types) x x √ √ x x √ x TOTAL 29 tokens (7 types) Academic vocabulary in the ICLE Legend: ** significant co-occurrent in the BNC-AC (p < 0. − not significant co-occurrents in the BNC-AC.01). √ the co-occurrent appears in the BNC-AC.

To sums up. However. which is always used in the BNC-AC when the conclusion (underlined in the examples) is introduced by a that-clause. In Example 5. these findings support Nesselhauf’s (2005: 25) argument that collocations should not be viewed as involving only two lexemes. come to + conclusion. other elements closely associated with them should also be taught.55. (ICLE-SP) The collocation escape + conclusion appears in two phraseological patterns in academic prose: ‘it is difficult to escape the conclusion that’ and ‘we cannot escape the conclusion that’. EFL learners use the collocations arrive at + conclusion. Conversely.g. The single occurrence of the collocation that appears in the ICLE is used in the native-like lexico-grammatical pattern ‘cannot escape the conclusion that’ but its subject is a nominal phrase headed by the noun evaluation: 5. (ICLE-PO) 5. we could enjoy doing things as dream and imagine more frequently.54.164 Academic Vocabulary in Learner Writing that are significant co-occurrents of the noun conclusion in the BNC-AC rises to 75. And taking into consideration that Marx was a materialist we can come to a conclusion that he himself would be attracted by the advantages of television. lead + conclusion and reach + conclusion.54. the frequent phraseme lead to the conclusion that is used with the personal pronoun us. However. (ICLE-DU) . and religion for him would remain the opium of the masses. 5. (ICLE-RU) 5.56 However.55. In the context of EFL teaching/learning. when we consider all the pros and cons of fast food we will certainly arrive at a conclusion that it is not an ideal way of eating. and that it is the criminalization of drugs that has created a much heavier burden on society.53 and 5.53. the indefinite article a is used instead of the definite article the. they do not always use them in native-like lexico-grammatical patterns.8 per cent as several of the verbs are repeatedly used in learner writing (e. draw + conclusion. a pattern which is very rarely found in academic prose. a more objective evaluation of the problem cannot escape the conclusion that. drug use and abuse have occurred in all civilizations all over the world. the percentage of verb co-occurrents that are not found in the BNC-AC falls to 12 per cent as ‘non-native’ co-occurrences are rarely repeated. due to new techniques and industrialization. come to and draw). all I have mentioned before lead us to the conclusion that if our lifes were a little ‘easier’ and we wouldn’t be dominated by a world that is constantly changing. In Examples 5.

59. two nouns that. like conclusion. In Example 5. which result in crude approximations. Howarth (1996. i.57. Finally. (ICLE-SP) There are many other examples of EFL learners’ attempts at using nativelike collocations. the verb put forward is used with the noun conclusion. 5.58. However. the phrasal verb draw up is used in place of draw and in Example 5. The woman started to think about the price of progress and came into conclusion that automation causes more problems than it solves. plan draw proposal put forward conclusion Figure 5. 1998) refers to this phenomenon as a collocational overlap.9).57. the verb express has acquired a semi-technical sense and means ‘make something public’. (ICLE-SP) 5. It may be hypothesized that the learner who wrote this sentence has been influenced by the nativelike co-occurrence ‘express one’s opinion/view’. Its single occurrence in the ICLE can be qualified as ‘nonnative like’ as it appears with the first person singular pronoun I as subject and the possessive determiner my (Example 5. . This verb is commonly used with the abstract nouns plan and proposal. Finally. a set of nouns which have partially shared collocates (see also Lennon. (ICLE-PO) In Example 5.59. 1996). that is: technology.60.9 Collocational overlap . It is mainly used in legal discourse and thus conveys a rather formal tone as illustrated in Example 5. the preposition into replaces to. science and industrialization have not killed dreams and imagination.58).e. (BNC-AC-HUM) 5. the verb put forward is not used with the noun conclusion in English (see Figure 5. .60. a conclusion can be drawn up emphasizing our first statement. 5. The Divisional Court expressed its conclusion in the following terms: . I wanted to express my conclusions.61.Academic vocabulary in the ICLE 165 In the collocation express + conclusion. combine with the verb draw to form collocations. and no article is used in an attempt to produce the native-speaker sequence ‘came to the conclusion that’.

similar. (ICLE-RU) 5. consequently it needs to take steps in order to prevent combativity which will endangered their interests. (ICLE-PO) The phraseology of EFL learner writing is also characterized by a number of lexico-grammatical infelicities and errors.g. 5. tentative. foregone. As already pointed out by Nesselhauf (2005). however.62. (ICLE-SP) The same remark can be made about several adjective + conclusion co-occurrences (Example 5. Looking at this idea from the Polish point of view. The verb put forward means ‘to suggest an idea. main. The first ten most significant adjective co-occurrents of the noun conclusion in the BNC-AC are general. logical.166 Academic Vocabulary in Learner Writing 5. we can nevertheless notice that a certain importance is granted to them. and definite. also brings double standard conclusions. different.g.61. an account *about a murder (ICLE-RU)) or the preposition of instead of for after the noun demand (e. the demand *of raw material (ICLE-GE)). a conclusion can hardly be put forward as it is supposed to be more than a suggestion and the result of serious consideration and discussion.64). explanation etc. adjective co-occurrents of the noun conclusion in learner writing are not the most typical ones in academic prose even though a large proportion of them occur in the BNC-AC (see Table 5. firm. The noun conclusion enters into combinations that are not found in academic prose and which are semantically awkward: 5. This reveals learners’ weak sense of native speakers’ ‘preferred ways of saying things’. Without putting forward premature conclusions. especially one that other people later consider and discuss’ (LDOCE4) while a conclusion is ‘something you decide after considering all the information you have’ (LDOCE4).22). Looking for the conclusion I would like to say that every person is individual and each has his or her own character. More importantly.64. None of these appear in learner writing except for logical.63. opposite. Learners sometimes use the preposition about after the abstract noun account (e. Having considered the various aspects of capitalism a conclusion must be gathered: the system cannot provide for the basic needs of the population. (ICLE-FR) The semantic incongruity of the co-occurrence ‘put forward a conclusion’ is made apparent by contrasting the definitions of put forward and conclusion. EFL learners also produce deviant verb + noun free combinations. Thus. They also use .

Academic vocabulary in the ICLE Table 5.22 Adjectives 167 Adjective co-occurrents of the noun conclusion in the ICLE Frequency Significant co-occurrents of conclusion in the BNC-AC − − ** ** − − ** − − − ** − − − ** − ** ** − ** − − − − − − ** ** − − ** − − − − Appearance in the BNC-AC √ x √ √ x √ √ x √ √ √ x √ x √ x √ √ √ √ √ √ x √ √ x √ √ x √ √ x x √ √ absolute awful certain clear clever concrete depressing double standard fair false final frightening interesting liberal logical long-searched for obvious overall only own personal premature private radical right sad same satisfactory satisfying sensible successful terrifying understated unequivocal wrong TOTAL 1 1 3 1 1 1 1 1 1 1 5 1 1 1 4 1 1 1 1 4 1 1 1 1 2 1 2 2 1 2 1 1 1 1 1 51 tokens (35 types) Legend: ** significant co-occurrent in the BNC-AC (p < . − not significant co-occurrents in the BNC-AC.01). √ the co-occurrent appears in the BNC-AC. x the co-occurrent is not found in the BNC-AC. .

The negative image of feminism makes it twice as hard for women to rise above it than it would be if men were facing *a [the] same kind of dilemma. the possibility *to learn a good job (ICLE-FR)). (ICLE-GE) 5. and discuss *about. (ICLE-CZ) It should be noted that very few of these errors are widespread in learner writing and that some of them may be partly L1-induced. 5. their rivalry made them hold continuous battles. Luckily. the function of comparing and contrasting was shown to be generally underused in learner writing. 1989): 5.2. attempt *of.g. as *it was [as was] the case of Catholics and Protestants. I would say because otherwise only the rich would be able to posses them as *it is [as is] unfortunately the case with many products in other areas of living. For example. 2009b). It is a matter of fact that these ‘things’ cannot be bought and sold like shares on the stockmarket. Because of the ambition for the power. An analysis of individual lexical .4.69. Example 5.65 illustrates learners’ confusion between the prepositions despite and in spite of. French learners use the erroneous colligation discuss *about as a translation of the French discuter de (Granger and Paquot. which results in the blend *despite of (cf. related *with.168 Academic Vocabulary in Learner Writing a to-infinitive structure after the noun possibility instead of an –ing form (e. (ICLE-RU) Learners also have a tendency to use the impersonal pronoun it in the subject position after as: 5. When different people read *a [the] same book they have probably various imaginations while reading. Other examples of colligational errors include suggest *to. (ICLE-SP) Another source of error is the adjective same which is sometimes preceded by the indefinite article in the ICLE: 5.1. Despite *of [Despite] the absence of such professionalism our nation overcame fascists. Dechert and Lennon. Semantic misuse In Section 5. (ICLE-FI) 5.67.

1999: 534-535). most Americans have moved to the USA from different countries as immigrants. (Crewe. (ICLE-PO) 5.70. we are all familiar with students who use ‘on the contrary’ for ‘however/on the other hand’. in Example 5. (ICLE-RU) 5. thus adding an unintended ‘corrective’ force to the merely ‘contrastive’ function sought. as an equivalent alternative to on the other hand. Raskolnikov differs from Onasis.70 to 5. etc. EFL learners’ semantic misuse of the phraseme on the contrary has already been reported for different learner populations: In Hong Kong. Crewe. Lake (2004) . 5.70. (cf.Academic vocabulary in the ICLE 169 items. EFL learners typically use on the contrary erroneously (instead of a contrastive discourse marker such as on the other hand or by contrast) to contrast the qualities of two different subjects (underlined in Examples 5. the fact that Onasis had everything is contrasted with the fact that Raskolnikov had nothing and the phraseme by contrast would have been more appropriate. *on the contrary [by contrast]. The main users of this kind of vehicles are families.72.72). The young like crazy driving. reveals that the adverbials on the contrary and on the other hand are overused in the ICLE. As Lorenz (1999b: 72) has demonstrated. *On the contrary [By contrast]. that is. Europeans have lived in their countries for hundreds of years. overtaking and leading on the roads. For instance. This is confirmed by our corpus-based analysis of EFL learners from different mother tongue backgrounds (see also Celce-Murcia and Larsen-Freeman. had nothing. Sports cars are created for this use and this may be the reason why their price is so high and use is expensive. of course. station wagons are not expensive in maintenance. 1990). 1990: 317) Granger and Tyson (1996) report the same conceptual problems for French learners. Thus. Raskolnikov. by contrast. Onasis had everything but he wanted to have more. (ICLE-FI) The semantic inappropriacy of on the contrary in EFL learner writing has been attributed to teaching practices. overuse is often accompanied by patterns of non-native usage. *On the contrary [By contrast].71. Teaching materials often provide lists of connectors in which the adverbial on the contrary is described as a phrase of contrast. Lake (2004) states that a large proportion of EAP non-native speakers who use on the contrary do so inappropriately. however. For pedagogical purposes.

2004: 137). such functional phrases [connectives] are usually familiar to learners from an early stage. argued that French learners’ overuse and misuse of on the contrary is probably due to an over-extension of the semantic properties of the French ‘au contraire’. for example. an argument.1. however. This view.170 Academic Vocabulary in Learner Writing proposes a checklist of contextual features that should be present when on the contrary is employed: As for the implications for learners. (in lieu of e. it now becomes possible to consult a checklist of contextual features that should be present in order for on the contrary to be appropriate: one subject. it ought to prove a useful starting point from which EAP teachers can devise their own practice materials. adjacent to the phrase both form a refutation. either genuinely present or implied. Granger and Tyson (1996). (Lake. 2004: 142) Lake (2004) rules out the possibility of an L1 influence on EFL learners’ semantic misuse of on the contrary on the basis that over 70 per cent of international students from widely different mother tongue backgrounds produced two distinctly separate L1-equivalent items in a cloze test in which they were required to insert on the contrary or on the other hand and provide an equivalent phrase for both adverbials. the preposition as (instead of such as) and the . which can be used to express both a concessive and an antithetic link. Such a checklist may be simplistic in that it does not cover all the possible lexico-syntactical environments in which the phrase might be encountered. but as a guideline for production. is over-optimistic and is clearly not reflected in our corpus-based learner data. one positive statement and one negative statement open to similar interpretations. and do not pose great problems of usage’ (Lake. to which the two statements.). EFL learners’ inappropriate use of the abbreviation i. The L1 equivalent forms to on the contrary and on the other hand may be characterized by different patterns of usage and thus be the source of negative transfer. Lake (2004) considers EFL learners’ misuse of on the contrary to be ‘something of an exception’ and writes that ‘in the EAP context. probable that misguided teaching practices and L1 interact here.g.e. two contrasting qualities. It is. however. The potential influence of the first language on French learners’ use of on the contrary is discussed in Section 5.3 below. In Section 5.

all countries should understand . I strongly believe that there is still a place for dreaming and imagination in our modern society. Everybody knows that children like inventing funny stories and amusing plays by using their wide fantasy. *On the other hand.Academic vocabulary in the ICLE 171 adverb namely was discussed. on the other side. This is one reason why children always bring happiness and awake the adults’ childish part. [P] This might be an illusion. So. who use it to mean ‘another side or aspect’. They suggested that this semantic misuse might be L1 induced: the Chinese equivalent of on the other hand is often misused by novice L1 writers.73. which suggests that there are other contributing factors to this learner difficulty.76. The function of punishment is to show that crimes are not acceptable or that they can solve any problems. where there is a child. because *on the other hand [ø] the death penalty develops violence and is incompatible with the basic laws of humanity. This does not occur in academic prose. The following extracts are examples of the use of on the other hand in the ICLE where it would have been more appropriate to use no connector or an additive marker: 5. because criminals would be afraid of the severe punishment.74. it is children who keep dreams and imagination alive! (ICLE-IT) 5.75. fantasy is [also] a useful mean used by teachers in primary schools to teach school subjects to their little students. Poland cannot reply with isolation as the unification still remains the best solution to its problems. Criminality would be limited. with no implied contrast. [P]10 Firstly. The re-introduction of the death penalty may have positive sides. erroneous uses of on the other hand are found in most ICLE sub-corpora. Although L1 influence may play a part in Hong Kong Chinese students’ inappropriate use of the adverbial. Other examples of semantically misused lexical items in learner writing include on the other hand. moreover. there are always dreams and imagination. It is illustrated in the following examples: 5. (ICLE-GE) 5. and even if. besides. too. *On the other hand the aim of punishments is [also] to make the criminals obey the laws and show example to other’s so that they will not follow the bad example and commit the same crime. On the other side. Field and Yip (1992: 25) reported that on the other hand is frequently used by Cantonese speakers to make an additional point. (ICLE-FI) The word combination on the other side sometimes appears in the ICLE in places where a contrast seems to be the logical link intended by EFL learners.

In the second sentence. not a concession. EFL learners’ use of labels is . Even if these descriptions are valid they still leave open a number of questions. (ICLE-FR) There is also some confusion between the conjunctions even if and even though in EFL learner writing. (ICLE-PO) 5. *even if [even though] we are not in the European Union. EFL learners also experience difficulty with the semantic properties of other types of cohesive devices.80. But on the other side it is sometimes hard to live without car or aerosols. This hole is caused by technical improvements in the last decades. But on the other side we will form a new nation with new hopes. Another big problem is our environment. Semantic misuse has often been discussed in the literature in relation to logical connectives.81.172 Academic Vocabulary in Learner Writing that history and its consequences cannot divide the continent.*even if [even though] I agree that the American public school system is defective. as I feel that parents are not the best teachers for their own children. he or she does not. and claim that are inherently unspecific and require lexical realization in their co-text. Compare: 5. Europe 92 means well a loss of identity since we’ll be no longer Belgians. There is pollution wherever you look. English . home schooling to me is no real alternative. In the first sentence. abstract nouns such as issue.. (ICLE-SW) Even if should be used to introduce a condition. . Even though these descriptions are valid they still leave open a number of questions. (ICLE-GE) 5. 2006).83. 5. The successful process of unification should be carried out with respect to nations’ rights and without special privileges given to the powerful.. *even if [even though] they are cheaper.e.79. labels. i. argument. Italians. particularly why the same mechanisms do not operate with girls. (ICLE-PO) 5.77. new ideas . However. and more specifically. In addition to phraseological and lexico-grammatical infelicities. However. We are as much a part of Europe as any other country here. Learners often use even if in lieu of even though to introduce a concession: 5. We must forget about refrigerators containing CFC-11 and CFC-12.82. particularly why the same mechanisms do not operate with girls. . We can no longer enjoy the sun in summer because of the hole in the ozone layer. either beforehand or afterwards (Flowerdew. the writer knows and accepts that the descriptions are valid. (ICLE-GE) 5.78. but Europeans.

The most important question concerning genetic engineering is the problem [that] of gen manipulation with humans. [P] First. However it will not be possible until males re-think and.89.86.85. (ICLE-FR) The noun argument also seems to cause difficulty to EFL learners. the companies feared the consequences that would follow a confession. hopefully. This short discussion of the main points linked to the problem [issue] of capital punishment leads to the final question. (ICLE-PO) 5. the rather unidiomatic expression ‘familiar arguments about’ should be rephrased as ‘widespread or popular beliefs about’.88. economy and environmental protection would be to the benefit of all. In Example 5. Industrialisation has transformed dreaming into a waste of time which is now “cleverly” linked to money. Female participation in making decisions concerning war and peace.87. irrational and dependent on instincts. Learners. reject familiar arguments [widespread/popular beliefs] about women being unreliable.87.88. another aspect introduces a second example (about the unemployed and housewives) of the fact that you are judged by what you do rather than by what you are. In Example 5. (ICLE-GE) 5. If we are aware of the fact that such time-tables are very common for people living in a modern society like ours. 1999b) in lieu of more specific nouns such as issue or question as illustrated in the following sentences: 5. (ICLE-PO) 5. aspect and issue.84. .90. In Example 5. for example. . the sentences that follow the label argument would be better described as ‘reasons’ why Big Tobacco did not depart from prepared statements. There are two main arguments [?reasons] that help us understand why Big Tobacco stuck to their statements for so long. Lorenz. in certain aspects stands for in some respects and the aspect of money probably refers to the ‘money issue’ or the ‘money question’ in Example 5. 5. . . In Example 5.91. the problem [question] of the place of imagination and dreaming is not even worth examining. They feared that there was going to be even more legislation and regulation if they would ever admit to lying.Academic vocabulary in the ICLE 173 characterized by semantic infelicity or lack of semantic precision. among many others. use the noun problem as an ‘all purpose wild card’ (cf. (ICLE-DU) Other problematic labels include. contrasting it with the first example (about physicists and mathematicians). .

it is still a matter of distant future. you are judged by what you do rather than by what you are.93. And another aspect is that [?by contrast.94. There were and there are different ideas about making all people equal. philosophy and first of all. Your job is your “trademark”. but true. in other words. but only *in certain aspects [in some respects].] the unemployed or housewives are sometimes treated as social outcasts. Chains of connective devices EFL learners’ texts are sometimes characterized by the use of too many connective devices (Crewe.92. a hyperbolic issue [product] of my vivid imagination. for as far as real good relations among countries are concerned. Narita and Sugiura. issue most probably stands for ‘product’: 5. [3] So one can wonder if a university degree really prepare students for real world and what his value is nowadays. Our modern western society puts a lot of pressure on people as far as work is concerned.2. Her issues [?] lies on the verge of theology.174 Academic Vocabulary in Learner Writing 5. (ICLE-PO) 5. For example. The aspect [?issue/question] of money includes the problem of equality. The picture I draw from my dear old houseman admittedly is nothing but a mere cliché. 1990. 2006). She is employed in defining the relation between faith and the mind. Human beings can eventually feel as one great family.5. according to popular opinion you must be very intelligent if you are a physicist or a mathematician. Actually. religion. [4] I think it is true that lectures in themselves are theoretical. because it was considered that this would lead to common happiness. tries to provide answers to them. (ICLE-GE) 5.92. 2006.89. Chen. Uta Ranke-Heinemann. Sad. The following text is an excerpt from an essay written by a French-speaking EFL learner.91. Each sentence contains at least one connective device – typically an adverbial connector or a sentence stem – which is often found in sentence-initial position (see Section 5. (ICLE-PO) 5. [1] But what about these prestigious institutions today? [2] To caricature them rapidly one could say that universities consist of courses given by professors (competent in their fields) in front of a silent audience who is conscientiously taking notes. A legend exists that money was invented by the devil to tempt the mankind.2. [5] Firstly because students spend . (ICLE-RU) In Example 5. the most famous woman in the field of Catholic theology.93. bits of information from the remotest parts of the globe reach us in an instant.6 below). 5. (ICLE-GE) 5. Or. it is not quite clear what her issues refer to and in Example 5.90.

). [6] Secondly because the subjects of the lectures are theoretical. to confront them and then to be able to pass judgment on them. which means that they have to dissect them. which is not often the case at university. firstly.g. [22] I think that this is a fully justified criticism against this institution. [7] For example: during a general methodology course (which. Lado’s are studied in detail but practical points are hardly ever considered. moreover. I think that academic studies develop a critical mind. thanks to the theoretical background they have learned. Crewe (1990) attributed EFL learners’ massive overuse of connective devices to their attempt at imposing ‘surface logicality on a piece of writing where no deep logicality exists’ (Crewe. I could say that a teacher in front of a classroom does not think about particular methodological theories again but that he has created his own methodology. secondly. and at worst causes the thread of the argument to zigzag about. [15] The students are indeed trained to analyse pieces of information coming from different horizons from a critical point of view. university students are able to build up their own way to achieve their aim. for example. I want to show that theory must always be accompagnied by practical applications. [11] The reason is that.g. Several of these connectors are superfluous and sometimes wrongly used (e. [19] Let’s take the example of a student in economics who has his certificate in his pocket and proudly goes working in a big firm for the first time. could be more practical) different theories as Krashen’s. [13] To take the example of a teacher again. [10] First I think that university degrees are theoretical on purpose (as opposed to high schools which are more practical. [17] Nevertheless. [14] Secondly. [8] However is it true that this formation does not prepare students for real world? [9] I am of the opinion that the answer is no. Some EFL learners use many logical connectives between sentences simply to indicate to the reader that they are adding another point (e.Academic vocabulary in the ICLE 175 most of their time sitting in big classrooms which do not allow practical exercises but only ex cathedra lectures. 1990: 320). indeed in sentence [15]). [20] I would compare this business man to a gentleman who perfectly knows the highway code and who knows how to start and how to run through the gears but who finds himself in the center of Paris at the peak hours the first time he really drives! [21] By this example. we think. to take the example of). He added that ‘over-use at best clutters up the text unnecessarily. first. [12] Moreover they are also able to adapt or to modify their method according to the situation. I do not want to go too far. The following excerpt from an EFL learners’ essay is a good example . [16] That is the way they should create a personal opinion for themselves. [18] I really think that theory is essential but I am convinced that practice should also be present. as each connective points it in a different direction’ (ibid: 324). moreover in sentence [12].

5. As a consequence. because they are not able to live together in harmony. these ideas were as much an insult to man’s estimation of himself as Darwin’s allegation. (ICLE-SW) 5.96. two hundred years later. (ICLE-DU) As Aijmer (2001) showed in a study of Swedish EFL student writing. one for almost every sentence. learning when not to use them is as important as learning when to do so. They often use I think or an equivalent expression (e. I think and as far as I am concerned. can lead to prose that sounds both artificial and mechanical’ (Zamel. Of course. Sentence [18] in Example 5. Furthermore. Hobbes is a stern determinist. ‘words or expressions that may be sprinkled over a text in order to give it an “educated” or “academic” look’ (Crewe.97 respectively are two more instances of rhetorical overstatement.g.97. (ICLE-IT) The pedagogical implication of these findings is that. that our ancestors used to live in trees. students need to be taught that excessive use of linking devices. Hobbes even considers people as artificial creatures. 5. For example. Hobbes was accused of being an atheist and forbidden to publish any more books. . which the author regards as typical of non-native-speaker argumentative essays. I think that in Examples 5. The clusters To me.e. learners use I think to make their claims more persuasive rather than to express a tentative degree of commitment. He regards man. The sequence I think it is true in Sentence [4] corresponds to what Aijmer (2001) described as a ‘rhetorical overstatement’. I agree with George Orwell. not belonging to nature. I am convinced that) when it is communicatively unnecessary in the flow of argumentation.176 Academic Vocabulary in Learner Writing of EFL learners’ use of logical connectors as ‘stylistic enhancers’. In other words. because as far as I am concerned I think that in every country there are few people which are rich and many people which are poor. 1983: 27). like nature. something which animals like bees and ants are capable of. Therefore a concept like ‘free will’ is impossible. I am of the opinion that. and then on the other hand I understand that they also can be seen as separate. ‘important as these links are. i. 1990: 316) but whose presence will not make the text coherent.94 could be rephrased as ‘Theory is essential but practice should also be present’.96 and 5. because they are natural. To me I think technology and imagination are very much interrelated. as subject to the chain of cause and effect.95.

which meant that a greater number of people could be fed.99.5 per cent in the BNC-AC-HUM (see Example 5.102. 5. and sentence-initial due to which is repeatedly used in learner writing but hardly ever occurs in academic prose (Example 5. Our analysis of connectors in the ICLE supports this hypothesis. (BNCAC-HUM) 5. Despite its commercial character Christmas still means a lot to me. Table 5. 1992. often immediately after the subject. In practice. (ICLE-DU) Another example is the adverb therefore which often appears in sentenceinitial position in the ICLE but is not often used in that position in the BNC-AC-HUM: . as does however in Example 5.e. Due to these developments the production expanded enormously. Narita and Sugiura.6. 2006).100. (BNC-SP) EFL learners’ marked preference for the sentence-initial position has been reported in various studies focusing on one L1 learner populations (Field and Yip. The final position is also possible. Agriculture was brutally collectivized and no concessions were made in the use of the Ukrainian language and culture.98.101). They can also occur in a medial position. within the sentence. however. 5. It’d be worth asking him first. Examples include the preposition despite which appears in sentence-initial position in 52 per cent of its occurrences in the ICLE but only in 34. Granger and Tyson (1996: 24) commented that ‘it is likely that this tendency for learners to place connectors in initial position is not languagespecific’. Zhang.99. Sentence position 177 Linking adverbials occur in different sentence positions. though.23 shows that the total proportion of sentence-initial connectors in learner writing is much higher than that found in academic prose (13. 1999b. Denikin’s White armies counter-attacked and after seven months the Red Army was obliged to withdraw. Coysevox’s bust of Lebrun repeats – again with a certain restraint – the general outlines of Bernini’s bust of Louis XIV. (BNC-AC-HUM) 5.2. but is more typical of speech as illustrated in Example 5.100. the Red Army units did nothing to conciliate the Ukrainian Left or the peasants.102).98. as shown in Example 5. (ICLE-FI) 5.101. Lorenz. shows a realism and subtlety of characterization that are Coysevox’s own. However. i. They often occur initially. 2000.Academic vocabulary in the ICLE 5.2% compared to 6%). The face.

1 26.5 20.8 145.4 70 25. He showed that.8 53.730 50.2 676 1374 65 22 31 151 46 60 235 3 94 28 233 86 176 882 42 365 392 48 155 675 5 75 756 6.436 199 689 446 43.3 18.9 53 265.6 1249 60. (ICLE-PO) These findings provide evidence for EFL learners’ lack of knowledge of the preferred syntactic positioning of connectors in English. pmw 203.6 690 58.3 52.5 12. on rest.4 46.2 39.6 19.9 20.2 88.5 52 11.2 27.7 44.5 68.4 40 218.5 6.9 24.6 2009 although and as a result as a result of as far as X is concerned because because of consequently despite due to even if even though for example for instance furthermore however in spite of moreover nevertheless on the contrary on the other hand so thanks to therefore thus TOTAL 263 1456 71 24 96 107 62 103 50 29 83 46 235 93 113 673 47 255 170 92 228 805 68 340 221 5.353 159 495 676 95 372 1.5 63. 522 32.6 79.67 42 34.3 49.5 22.3 5.4 42.9 30.505 % Rel.675 29.4 57.7 11.5 13.128 106 292 250 164 418 1.236 103 79 167 2.5 56 34.4 14.2 1.3 87. Scientific research as well as individual observations prove that eating habits have a great impact on the condition of the human body and soul and.207 599 143 681 195 451 248 1263 609 217 3. pmw 225.5 201.6 4.8 18 70.412 1.7 35.4 73. Flowerdew (1993) argued that teaching materials do not provide students with authentic descriptions of syntactic patterns of words.5 41.5 4.6 82.306 102 194 59 2.79 7.5 227.24 S-I BNC-AC-HUM Total freq.9 118 14.894 35 1.276 91.8 96.4 13.767 110.5 1. 2.5 27 96.9 71.4 91.1 54. Therefore people should pay more attention to what they consume.7 203.5 413.8 11.493 530 179 96 246 274 127 854 344 127 1.916.23 The frequency of sentence-initial position of connectors in the BNC-AC-HUM and the ICLE ICLE S-I Total freq.178 Academic Vocabulary in Learner Writing Table 5.7 57. contrary to what is often .7 58 50.8 6 5.1 81.3 291. freq.2 8.6 59.808 % Rel.3 26.2 49.11 This lack has often been attributed to L2 writing instruction. consequently.3 42.7 189.7 0.3 68 56.8 30.3 36.6 9 45.9 577. sleeping and even dreams.7 1.4 4.3 11.9 195.6 6.6 14.9 28.103. freq.6 109.8 78.

Europe’s history is inseparable from world history between 1880 and 1945. they introduce the cause of something that is described in the main clause: 5. 2004: 135). Because the death-rate was high. 1999:13) (see also Biber et al. EFL learners’ tendency to place connectors in unmarked sentence-initial position seems to be reinforced by teaching (see Granger. EFL learners sometimes use sentence-initial because to introduce new information in independent segments and give the cause of something that was referred to in the previous sentence: 5. It is also used to serve different functions in learner writing.106. The crime rate would also strongly reduce and this is of course the main objective of all this measures. Similarly. only be to win Pyrrhic-victories.Academic vocabulary in the ICLE 179 taught in course books.. the adverbial connector then rarely occurs in sentence-initial position. Unmarkedness provides another possible explanation for EFL learners’ massive overuse of sentence-initial connectors. sentence-initial becauseclauses are attached to a main clause. (ICLE-SW) . 1985). (BNC-AC-HUM) 5. Thus. In academic prose. marriages were usually short-term. sentenceinitial because is significantly more frequent (relative frequencies of 9. She showed that. Milton (1999: 225) discussed the problematic aspects of teaching connectors by means of lists of undifferentiated items.54 in academic prose). (ICLE-DU) 5. EFL learners seem to use the unmarked sentence-initial position as a safe bet. (BNCAC-HUM) Unlike expert writers. the proportion of sentence-initial because is lower in learner writing than in professional writing. To directly try to change people with ‘experience of life’ would. at best. compared to this effective investment.104. Contrary to our expectations. and suggested that one way in which instruction may skew EFL learners’ style is ‘by the presentation of these expressions as if they occurred in only sentence-initial position’ (see also Narita and Sugiura. Because deep inside every man’s heart lies the ‘Indian’-insight that we are only borrowing the earth from our children.. Because these changes were worldwide. As shown in the following examples. the highest percentage of linking adverbials appear in sentence-initial position and concluded that ‘initial position seems the unmarked position for linking adverbials’ (Conrad. 2006).107.105. in both conversation and academic prose. but is more usually found in a medial position. However. 1999 and Quirk et al. Conrad (1999) studied variation in the use of linking adverbials across registers. Because everybody wants to live in a safe society.18 in learner writing and 4.

Linking adverbials which occur between an auxiliary and the main verb. 2.: It is difficult to believe therefore that one of these mosaics was not influenced by the other. anyway and though — account for the relatively high proportion of sentence-final linking adverbials in native-speaker’s conversation. transferred conjunction strategies from speech to essay writing. in addition to showing a link with previous discourse. teaching materials tend to focus on sentence-initial position. (BNC-AC-HUM) 3. and EFL learners probably feel unsafe about other syntactic positionings for connectors.g. a fact stressed by all reputable modern historians who have worked on this intractable subject. Three types of medial position are particularly frequent (Conrad. (BNC-AC-HUM) A medial position for connectors is quite typical of academic prose. Linking adverbials which occur immediately after the subject as illustrated in Example 5. most linking adverbials are placed in sentence-initial or medial position.108. The final position is frequent in conversation. can also play important roles in the interpersonal interaction .180 Academic Vocabulary in Learner Writing 5. Conrad (1999) reported that. e. (ICLE-CZ) EFL learners share this characteristic with ESL writers.99 above. However. it is clearly less favoured by EFL learners.24 shows that several connectors are also repeatedly used in sentence-final position in the ICLE. In a comparison of strategies for conjunction in spoken English and English as a Second Language (ESL) writing. In my opinion it is useful only for them. Linking adverbials which occur between the main verb and its complement. Schleppegrell (1996) found that students who had spent most of their lives in the US and learnt English primarily through oral interaction. Because their sorrow is found as the extenuating circumstance. She argued that these linking adverbials are commonly found in sentence-final position as they serve important interpersonal functions: Adverbials in conversation. They employed both an ‘afterthought’ because (Altenberg. Conrad (1999) found that three highly frequent items – then. which is quite uncommon in the BNC-AC-HUM. in academic prose. As indicated above. for their trial. 1984) to add information in independent segments. but rare in academic prose. and other types of speech-like clause-combining strategies. such as: All estimates of population size must therefore allow for a large measure of conjecture. 1999: 14–15): 1. Table 5.

freq. . final anyway is often associated with expressions of doubt or confusion.9 S-F 20 20 8 18 14 17 7 BNC-AC-HUM Tot. and (. It has.2 1. been suggested that learners’ preference for the sentence-initial . It may be.3 4. freq.2 was on interlanguage features that are shared by most learner populations when compared to expert academic writing. then.3 2.8 4.5 0.6 0. may combine to influence learners’ use of academic vocabulary. Transfer-related effects on French learners’ use of academic vocabulary The focus of Section 5.5 0.1 5.3 1.4 9.7 1. and the textual and/or interpersonal functions they serve. ) then typically indicates that a speaker is making an inteference (sic) based on another speaker’s utterance.Academic vocabulary in the ICLE Table 5. 71 1263 609 1413 863 3062 178 % 28. for example.6 0.4 0. 132 854 344 257 750 1054 256 % 18. Thus. (. .2 that takes place. The placement of these adverbials in final position is consistent with previous corpus analysis of conversation that has found that elements with particular interpersonal importance are often placed at the end of a clause (.3 1.3.0 5. freq.6 0. none of the linking adverbials commonly associated with the final position in conversation are common in formal writing. however. . and which are therefore likely to be developmental. .24 HUM 181 Sentence-final position of connectors in the ICLE and the BNC-AC- ICLE S-F anyway for example for instance indeed of course then though 25 63 31 15 34 35 11 Tot.5 3. These findings suggest that the positioning of linking adverbials in native discourse is directly influenced by the register in which they appear.9 3. (Conrad 1999:14) The type of interpersonal interaction that takes place in conversation is not typical of academic prose. ) [A] final though often occurs when speakers are disagreeing or giving negative responses. 5.6 1. 2. .4 2.2 0. that in some cases in conversation there is a tension between placing the linking adverbial at the beginning of the clause. and at the end of the clause.9 Rel. .3 Rel. Multiple factors. freq. due to its interactional function. 0.0 0. These roles are often particularly noticeable for the common adverbials in final position.9 7.5 0. ). due to its linking function.

He argued that transfer studies should minimally consider at least three potential effects of L1 influence when presenting a case for or against L1 influence: 1.12 Intra-L1-group homogeneity is verified by comparing the interlanguage of learners sharing the same mother tongue background. (Jarvis. Jarvis referred to a number of studies reported by Ringbom (1987) that . To illustrate this first L1 effect. As Granger (1998b: 158) put it. 2000: 252) Jarvis translated his working definition of L1 influence into a list of specific types of L1 observable effects that should be examined when investigating transfer. 2000: 254). 1972).g. Claims made about the nature of L1 influence and its interaction with other factors. which is intended as a methodological heuristic to be used by transfer researchers: L1 influence refers to any instance of learner data where a statistically significant correlation (or probability-based relation) is shown to exist between some features of learners’ IL performance and their L1 background. Intra-L1-group homogeneity in learners’ IL performance is found when learners who speak the same first language behave as a group with respect to a specific L2 feature. Jarvis (2000) incorporated three types of L1 observable effects into a unified framework for the study of L1 influence and proposed the following working definition of L1 influence. have often been built on shaky methodological foundations and suffer from what Jarvis (2000: 246) referred to as a ‘you-know-it-when-you-see-it’ syndrome. ‘learners clearly cannot be regarded as “phraseologically virgin territory”: they have a whole stock of prefabs in their mother tongue which will inevitably play a role – both positive and negative – in the acquisition of prefabs in the L2’.182 Academic Vocabulary in Learner Writing position for connectors may be attributed to the influence of instruction or ‘transfer of training’ (Selinker. 2. I like very much movies). Jarvis used Selinker’s (1992) finding according to which Hebrew-speaking learners of English as a group tend to produce sentences in which adverbs are placed before the object (e. however. To remedy this situation. Inter-L1-group heterogeneity in learners’ IL performance is found when ‘comparable learners of a common L2 who speak different L1s diverge in their IL performance’ (Jarvis. The marked difference in frequency of I think across the learner sub-corpora may be partly explained by different academic writing conventions in the different mother tongues. To illustrate this effect.

2000: 254–5). 2000: 255). The identification of two simultaneous L1 effects is necessary to present a convincing case for L1 influence. Jarvis concluded that. Jarvis argued that ‘this type of evidence strengthens the argument for L1 influence because it essentially rules out developmental and universal factors as the cause of the observed IL behaviour. 2000: 255). Table 5.Academic vocabulary in the ICLE 183 have shown that Finnish-speaking learners are more likely than Swedishspeaking learners to omit English articles and prepositions.25 L1 effect Jarvis’s (2000) three effects of potential L1 influence reliability poor strong strongest sufficient criterion no no no Intra-L1-group homogeneity in learners’ IL performance Inter-L1-group heterogeneity in learners’ IL performance Intra-L1-group congruity between learners’ L1 and IL performance . Identifying the three L1 effects would be even more convincing if it were not that ‘the ubiquity of conditions that can obscure L1 effects renders the three-effect requirement unrealistic in many cases’ ( Jarvis. 3. The added value of this third L1 effect is that it also has explanatory power by showing ‘what it is in the L1 that motivates the IL behavior’ (Jarvis.25. These three effects can emerge in circumstances in which transfer is not at play and can thus be misleading when considered in isolation. As shown in Table 5. Intra-L1-group congruity between learners’ L1 and IL performance is found where ‘learners’ use of some L2 feature can be shown to parallel their use of a corresponding L1 feature’ (Jarvis. 2000: 255). it shows that the IL behaviour in question (omission of function words) is not something that every learner does (to the same degree or in the same way) regardless of L1 background’ (Jarvis. Inter-L1-group heterogeneity is identified by comparing the interlanguage of learners from different mother tongue backgrounds. none of the three effects is sufficient by itself to verify or characterize L1 influence. Intra-L1-group congruity is confirmed by an IL/L1 comparison. Selinker (1992) uses this type of evidence to show that Hebrew-speaking learners’ positioning of English adverbs parallels their use of adverbs in the L1. In other words. despite differences in the degree of reliability.

174-word comparable corpus of essays written by French-speaking students collected at the University of Louvain. Unlike Jarvis (2000). the three transfer effects are found in French learners’ use of on the contrary. i. The International Corpus of Learner English appears to be ideally suited to analysing the three potential effects of L1 influence described by Jarvis (2000).26 L1 effect Academic Vocabulary in Learner Writing Jarvis’s (2000) unified framework applied to the ICLE-FR Corpus comparisons A comparison of the use of a specific lexical item in all the essays written by French learners A comparison of the use of a specific lexical item in the ICLE-FR against other L1 subcorpora A comparison of a specific lexical item in the ICLE-FR to the use of its equivalent form in a comparable corpus of French native student writing Intra-L1-group homogeneity in learners’ performance Inter-L1-group heterogeneity in learners’ IL performance Intra-L1-group congruity between learners’ L1 and IL performance I made use of Jarvis’s (2000) unified framework to investigate the potential influence of the first language on multiword sequences that serve rhetorical functions in French learners’ argumentative writing. This strongly supports Granger and Tyson’s (1996) suggestion that French learners’ overuse and misuse of the connector is probably due to an over-extension .e. Intra-L1-group homogeneity in learners’ performance is investigated by comparing all the essays written by French learners to verify whether they behave as a group with respect to a specific L2 feature.184 Table 5.2. Inter-L1-group heterogeneity in learners’ IL performance is verified by a comparison of the number of texts in which a specific lexical item is used in the ICLE-FR and in other L1 sub-corpora. Table 5. I made use of comparison of means tests and post hoc tests such as Ryan’s procedure and Dunnett’s test to confirm this second L1 effect. French EFL learners’ use of a specific lexical item is compared to the use of its equivalent form in a 225. the Corpus de Dissertations Françaises (CODIF).4 as potential explanations for the frequent misuse of the adverbial. indicating that L1 influence most probably reinforces the conceptual problems and misguided teaching practices that were identified in Section 5.13 To establish intra-L1-group congruity between learners’ L1 and IL performance. Applying Jarvis’s (2000) framework to the ICLE texts reveals the potential influence of transfer on French learners’ use of multiword sequences that serve specific rhetorical functions in English.26 lists the three steps needed to investigate the influence of French on recurrent word sequences in the ICLE-FR. For example.

semantic extension) or ‘transfer of form/meaning mapping’ (e. Biskup. transfer of the phraseological environment. cognates) (see Jarvis and Pavlenko. However. They are illustrated in the remaining of this section. there is congruity between French learners’ use of according to me in English and selon moi in French. in fact.g. ‘transfer of meaning’ (cf. This may explain why French EFL learners are keen to use what they regard as a direct translation of a common French expression. and transfer of L1 frequency. and influential sculptor since Bernini”. however. 2007 for excellent syntheses on lexical and semantic transfer). Odlin. . Granger. By contrast. borrowing). 1989. Granger and Swallow 1988. where and how to use it (Nation. semantic transfer.g. Granger 1996b). 1998b. According to George Heard Hamilton. 2003 and Ringbom. Most transfer studies have focused on what we can call ‘transfer of form’ (e. These other types of knowledge can also give rise to transfer. which can be used to express both a concessive and an antithetic link.Academic vocabulary in the ICLE 185 of the semantic properties of the French au contraire. Van Roey 1990.g. the most admired. Nesselhauf. This sequence is repeatedly used in the ICLE-FR. Multiword sequences with a pragmatic anchor seem to be quite easily transferred. It helps to identify a number of transfer effects that remain largely undocumented in the SLA literature: transfer of function. Applying Jarvis’s (2000) unified framework on learner corpus data brings to light interesting findings relating to L1 influence on word use. Moreover. Next to knowledge of form and meaning. BNC-AC-HUM). 2001: 27). 1992. where it is extremely rare. transfer of style and register. knowing a word also involves knowing in what patterns. research into learners’ use of cognates has highlighted transfer effects on style and register (cf. when. they differ in one significant way: according to me is usually not accepted as a correct English phraseme. For example. prolific. French learners’ use of the idiosyncratic expression *according to me is a good example of transfer of function. Rodin became “a figure of international significance. much remains to be done regarding ‘transfer of use’. selon moi is perfectly fine in French and is. it does not appear in other learner sub-corpora except for the ICLE-DU and the ICLE-SW. 2008. with what words. Studies focusing on learners’ use of phrasemes have brought to light transfer effects on collocational restrictions and lexico-grammatical patterns (e.g. which are probably regarded as translation equivalents by French EFL learners. 2003). However. These four transfer effects often accompany transfer of form and meaning and may also reinforce each other. The English preposition according to and the French selon both mean ‘as shown by something or stated by someone’ (e. quite frequent in French native-speaker students’ writing.

115. Selon moi. Selon moi.A. the real problem now is not that man refuses to pay heed but that man refuses to make some sacrifices for the sake of ecology and to understand that the values that we have chosen are the wrong ones. la chanson est un vecteur de culture parce qu’elle est un art qui impose l’engagement des différents acteurs. a pattern that is also the preferred lexico-grammatical environment of illustrer in the corpus of French essays (Example 5. To illustrate this point. (ICLE-FR) Figure 5.27 shows that French EFL learners frequently use the verb illustrate in its infinitive form. we can mention the notion of culture and language in the north of Belgium. Although it is not found in many texts written by French learners. tout le monde pense ce qu’il veut et comme il veut. (CODIF) French learners’ knowledge of the verb illustrer in their mother tongue probably influences the type of word combinations and lexico-grammatical patterns in which they use the English verb illustrate.111. agit comme il l’entend en respectant la loi et les codes établis.186 Academic Vocabulary in Learner Writing The following examples illustrate French students’ use of selon moi and French EFL learners’ use of according to me: 5.S. According to me. The percentage of use of this form (40%) is quite similar to that of the infinitive form of the French cognate verb illustrer in CODIF. A closer look at the occurrences of the infinitive form of illustrate in ICLE-FR reveals that it is repeatedly used in sentence-initial to-infinitive structures (Examples 5. it is much more frequent in ICLE-FR overall than in any other learner sub-corpus. Table 4. To illustrate this.6%) (cf.2.113 and 5. but differs significantly from the proportion of infinitive forms of the English verb illustrate that were found in the BNC-ACHUM (23. Transfer effects are also detectable in French learners’ use of lexicogrammatical and phraseological patterns. (ICLE-FR) 5. According to me. 5. it would be interesting to compare our situation with the U.115).113. (ICLE-FR) 5. prenons l’exemple des pâtes alimentaires italiennes.2). (ICLE-FR) 5.114. The English verb illustrate is a case in point.112. (CODIF) 5.110. Pour illustrer cela. the prison system is not outdated: it has never been a solution per se.6 in Section 4.’s. Table 5.109. (CODIF) 5.10 represents graphically how the misleading translation equivalent may be created by French EFL learners.114). .

etc. sentenceinitial Pour conclure. etc. FRENCH LEARNERS' INTERLANGUAGE 'according to' 'according to' + [+HUM] 'according to' + [+HUM] 'according to X' *'according to me' e. viz. 'according to' + [+HUM] e. etc.g. lui. argument. idea. theory. 'selon moi' 'selon' + [-HUM] e. Hugo. théorie. idée. monsieur Bernanos.g. Civil Liberty Members. situation. article. 'according to' + [-HUM] 'according to' ENGLISH Figure 5.g. Xavier Flores.10 A possible rationale for the use of ‘according to me’ in French learners’ interlanguage Similarly. norme. principe. argument. supporters. certains. philosophie.Academic vocabulary in the ICLE FRENCH 187 'selon' 'selon' + [+HUM] 'selon X' e. French EFL learners almost always use the verb conclude in the sentence-initial discourse marker To conclude followed by an active structure introduced by a first person pronoun + modal verb. Judge Kamins. loi. etc.g. The following examples show that longer sequences . This pattern is less frequent in the writing of EFL learners with other mother tongue backgrounds and parallels a very frequent way of concluding in French.

(ICLE-FR) My findings also point to a transfer of style and register. it is used as a code gloss to . Pour conclure. In Section 5. Pour conclure.188 Academic Vocabulary in Learner Writing Table 5. This difference in use between ICLE-FR and the other ICLE sub-corpora proved to be statistically significant.119.116. freq. je dirais que chaque individu est unique. (CODIF) 5. To conclude. the two-word sequence occurs in 25. To conclude.55 and phraseological cascades (see Section 5. ‘illustrer’ in CODIF 8 13 3 1 1 26 31% 50% 12% 4% 4% 100% 11.2.28. we can say that many people are today addicted to television. technology and industrialisation certainly stand in the way of human relationships but not in people’s dreams and imagination.27 A comparison of the use of the English verb ‘illustrate’ and the French verb ‘illustrer’ En. For example.117. An analysis of concordance lines for let us shows that this sequence is repeatedly used by French speaking students to serve a number of rhetorical and organisational functions.2.3) may also be transferrelated.118. 5.9 per cent of the texts produced by French learners and is much more frequent in the ICLE-FR than in any other learner sub-corpus. différent et qu’il est facile de vouloir ressembler aux autres plutôt que de s’accepter tel qu’on est.000 words 10 8 2 0 0 20 50% 40% 10% 0% 0% 100% 14. per 100. the first person plural imperative form let us was shown to be overused by all L1 learner populations when compared to expert academic writing.67 Fr. nous pouvons dire que les deux stades sont aussi importants l’ un que l’ autre : il est nécessaire que l’ homme soit membre d’ un groupe mais il est tout aussi primordial qu’ il s’ en détache pour construire son identité propre. I would say that science. As shown in Table 5.3. (ICLE-FR) 5. ‘illustrate’ in ICLE-FR simple present infinitive past participle imperative past TOTAL Rel. (CODIF) 5.

24 12.85 Number of texts including ‘let us’ or ‘let’s’ Number of texts % 189 59 19 19 10 14 10 23 47 14 9 224 228 147 196 167 179 79 221 194 149 81 1641 25. (ICLE-FR) 5. (ICLE-FR) As explained in Section 4.65 introduce an example (Example 5. let us in French. Let us now turn our attention to the students who want to apply for a job in the private sector. of ‘let us’ and ‘let’s’ per 100.120.23 18.46 occurrences per 100.120).122). our behaviour would be cowardly.33 8. Equivalence is however found at the morphological level as French makes use of an inflectional suffix to mark the first imperative plural form.4 11.3.9 9.88 25.123). suppose. (ICLE-FR) 5. In the BNC-ACHUM.95 19. To illustrate the truth of this. (ICLE-FR) 5.21 38. a transition marker to change topic (Examples 5.2. return. 5. 1998.122. let us take the example of Britain which was already fighting its corner alone after Mrs Thatcher found herself totally isolated over the decision that Europe would have a single currency.78 13.7 10.000 words).9% 12. i.e. look.73 26. freq.Academic vocabulary in the ICLE Table 5. Thus. to investigate the third L1 effect.57 26. begin.121 and 5. There is no lexically equivalent form to En. take and have.8 12.. say.11 13. Let us then focus on the new Europe as a giant whose parts are striving for unity. and an attitude marker (Example 5.2 9. but it is not frequent (relative frequency of 5.123. Let us be clear that we cannot let countries tear one another to pieces and if we closed our eyes to such an atrocity. the first person plural imperative form let us is found in professional academic writing.7 6 7. Hyland. 2002). there are only eight significant verb co-occurrents of let us: consider.28 ‘let us’ in learner texts Rel.121. It is also restricted to a limited set of verbs (see Swales et al.69 20. intra-L1 group congruity between learners’ L1 and IL .000 words French Czech Dutch Finnish German Italian Polish Russian Spanish Swedish TOTAL 71.4 24.

let us take ‘prenons’.3 per 100. This generalized overuse of the first person plural imperative in EFL French learner writing as a rhetorical strategy does not conform to English academic writing conventions but rather to French academic style. Prenons l’exemple des sorciers ou des magiciens au Moyen Age.128. let us (not/never) forget ‘oublions/n’oublions pas que’. Considérons un instant le cinéma actuel. (CODIF) 5. (CODIF) 5.000 words in the BNC-SP but only 5. Ajoutons qu’une partie plus spécifique de la population est touchée. As a result. than in English expert or novice writing. (CODIF) 5. .g. and more specifically in academic writing. Examinons successivement le problème de l’abolition des frontières d’un point de vue économique. (CODIF) 5. 2008a): 5. 1996) and make use of imperatives in English academic writing in the same way as in French academic writing. Imaginons un monde ou règne une pensée unique. The rhetorical and organisational functions fulfilled by let us in French EFL learner writing can be paralleled with the very frequent use of first person plural imperative verbs in French student writing to organize discourse and interact with the reader (Paquot. juridique et enfin culturel. let us examine ‘examinons’. the frequency of let us in the ICLE-FR is much closer to the frequency of first person plural imperative verbs in student writing in French.127. This example points to yet another type of transfer effect. to illustrate this. the speech-like nature of let us in French EFL learner writing leads to an overall impression of stylistic inappropriateness. let us think ‘pensons’). to conclude and *according to me. let us (and more precisely its contracted form let’s) is much more typical of speech (relative frequency of 42. French EFL learners seem to transfer their knowledge of French academic writing conventions (Connor.5 occurrences per 100.125. Imperative forms that are repeated in the ICLE-FR often have formal equivalents that are found in CODIF (e. As shown in Table 5.000 in the BNC-AC).190 Academic Vocabulary in Learner Writing performance. Comparons cela à la visite de la cathédrale d’Amiens.29. Other examples of sequences that have French-like frequencies in the ICLE-FR include on the contrary. Envisageons tout d’abord la question économique. let us take the example.130. In English. on the other hand. I compared the use of let us in ICLE-FR with that of first person plural imperative verbs in CODIF.124.129. (CODIF) 5. let us hope ‘espérons’. (CODIF) 5. let us take the example of ‘prenons l’exemple de’. (CODIF) First person plural imperative verbs serve specific discourse strategies in French formal types of writing.126. let us consider ‘considérons’. namely transfer of L1 frequency.

.29 The transfer of frequency of the first person plural imperative between French and English writing Corpus Relative frequency per 100. ‘let us take the example of’ En ‘let us not forget’ En.7 3 191 French L1 students (CODIF) French EFL learners (ICLE-FR) English expert writers (BNC-AC-HUM) English novice writers (LOCNESS) FRENCH Fr.Academic vocabulary in the ICLE Table 5. FREQUENCYFR REGISTERFR FUNCTIONFR PHRASEOLOGYFR Figure 5. ‘let us take the example of’ En ‘let us not forget’ En. ‘examinons’ FREQUENCYFR REGISTERFR FUNCTIONFR PHRASEOLOGYFR ENGLISH En.000 words of first person plural imperative verbs 95. 1st plural imperative En.. ‘let us examine’ . ‘prenons’ example de’ Fr.9 5. ‘n’ oublions pas’ Fr. 1st plural imperative En. 1st plural imperative Fr.11 A possible rationale for the use of ‘let us’ in French learners’ interlanguage .5 71. ‘let us examine’ FREQUENCYEN REGISTEREN FUNCTIONEN PHRASEOLOGYEN FRENCH EFL LEARNERS' INTERLANGUAGE En.

192 Academic Vocabulary in Learner Writing Transfer effects often interact in learners’ use of English lexical devices. My results show that the expression of rhetorical and organizational functions in EFL writing is characterized by: A limited lexical repertoire: EFL learners tend to massively overuse a restricted set of words and phrasemes to serve a particular rhetorical .. 5. Summary and conclusion The data presented in this chapter support the idea that the ‘English of advanced learners from different countries with a relatively limited variation of cultural and educational background factors share a number of features which make it differ from NS language’ (Ringbom. EFL learners’ knowledge of words and word combinations in their mother tongue includes a whole range of information about their preferred co-occurrences and sentence position.4. and more precisely in argumentative essays. transfer of style and register. As illustrated in Figure 5. Thus. 1998: 49). They are the result of many encounters with these lexical items in L1 speech and writing. Mental primings in the L1 lexicon probably influence EFL learners’ knowledge of English words and word sequences by priming the lexico-grammatical preferences of an L1 lexical item to its English counterpart. and transfer of frequency – make up what. The transfer effects identified in this section – transfer of function. stylistic or register features. transfer of lexico-grammatical and phraseological patterns. I refer to as ‘transfer of primings’. following Hoey (2005: 183). in French-like phraseological patterns and to serve the same organizational and interactional functions. French EFL learners’ use of textual phrasemes such as let’s take the example of. let’s examine or let us not forget mirror the stylistic profile of the French sequences prenons l’exemple de .11. Primings for collocational and contextual use of (at least a restricted set of frequent or core) L1 lexical devices are particularly strong in the mental lexicon of adult EFL learners. examinons et n’oublions pas in French academic writing. This textual dimension is particularly difficult to master and has been described by Perdue (1993) as the last developmental stage before bilingualism in second language acquisition.. The focus of the analysis has been on the lexical means available to learners to perform specific rhetorical and organizational functions in academic writing. discourse functions and frequency. French EFL learners use English first person plural imperatives in academic writing with the frequency of French imperative verbs in the corresponding register.

with labels. Learners’ attempts at using collocations are not always successful and sometimes result in crude approximations and lexico-grammatical infelicities. i. although it is typical of academic prose. A medial position is not favoured by EFL learners. adverbs and prepositions rather than phraseological patterns with nouns. A lack of register awareness: texts produced by EFL learners often ‘give confusing signals of register’ (Field and Yip. Most . verbs and adjectives. Lexico-grammatical and phraseological specificities: EFL learners’ writing is distinguishable by a whole range of lexico-grammatical patterns and co-occurrences that differ from academic prose in both quantitative and qualitative terms. 1999b: 56).e. however. ‘the misuse of logical connectives is an almost universal feature of ESL students’ writing’. The methodology used in the first part of this chapter has made it possible to draw a general picture of the writing of upper-intermediate to advanced EFL learners from different mother tongue backgrounds. where lexical items are employed to signal grammatical and textual relations’ and that ‘a lack of coherence in advanced learners’ writing must at least partly be attributable to lexico-grammatical deficits’ (Lorenz. Preferred co-occurrences in the ICLE are often not the same as in academic prose. either beforehand or afterwards. My results also support Lorenz’s (1999b) remark that ‘advanced learners’ deficits are most resilient in the area of lexico-grammar. They also seem to prefer to use conjunctions. 1992: 26) as they display mixed patterns of formality and informality. is that EFL learners also experience difficulty with the semantics of other types of cohesive devices. The frequency of informal words and phrases in learner writing is often closer to their frequency in native-speakers’ speech than in their academic prose. Semantic misuse: As Crewe (1990: 317) commented. abstract nouns that are inherently unspecific and require lexical realization in their co-text. Chains of connective devices: EFL learners’ texts are sometimes characterized by the use of superfluous (and sometimes semantically inconsistent) connective devices. and specifically. A marked preference for sentence-initial position of connectors: connectors are often used in the unmarked sentence-initial position in learner writing.Academic vocabulary in the ICLE 193 function and to underuse a large proportion of the lexical means available to expert writers. What is less well-documented in the literature. which reveals learners’ weak sense of native speakers’ ‘preferred ways of saying things’.

2009). The reason for the initial positioning of conjunctions was again due to the transfer of the Chinese language where conjunction devices with similar meaning are mostly used at the beginning of a sentence. although it is indeed very significant in the Chinese component of the second edition of the ICLE (Granger et al. 2000: 77). As explained above. Consider the following quotations by Zhang (2000). teaching-induced factors have been identified as a possible cause for learners’ preference for sentence-initial position. Tripartite comparisons between professional writing. in this case Chinese: The overuse of this expression [more and more] was most probably due to language transfer since a familiar expression in the Chinese language ye lai yue was popularly used. we can check to what extent they are specific to EFL learners or just typical of novice writing. as developmental. . This suggests that. who attributes a number of features to the influence of the learners’ mother tongue. Neff et al. 2004: 135–6).6.1) has a role to play. Syntactic positioning of connectors is rarely taught and EFL learners often consider the sentenceinitial position to be a safe strategy. My methodology makes it possible to avoid hasty interpretations in terms of L1 influence. 2000: 83). teaching-induced and transfer-related effects can reinforce each other (Granger. This is precisely where a corpus of essays written by English native university students such as LOCNESS (see Section 2.. (Zhang. sentence-initial positioning of conjunctions is common to most learner populations. once linguistic features of upper-intermediate to advanced EFL learner writing have been identified. but they have always been reported on the basis of only one or two L1 learner populations. The mother tongue may reinforce learners’ preference for sentence-initial position but cannot be regarded as a complete explanation for this learner-specific feature. It is not always possible to attribute learner-specific features to a single factor. (Zhang.194 Academic Vocabulary in Learner Writing of these features have already been mentioned in the literature.g. Another advantage of the method I used is that.. 2008).2. As for the overuse of the expression more and more. In Section 5. it is probably not the only explanation. while transfer may be at work in Chinese learners’ use of more and more. this feature is actually common to all learner populations represented in the corpus. foreign learner writing and native student writing make it possible to distinguish between learner-specific and developmental features (e.

but as a general rule. absolutely.12 Features of novice writing – Frequency in expert academic writing.12 shows that a whole range of lexical items that Gilquin and Paquot (2008) found to be overused in EFL learners’ writing – maybe. Figure 5. spoken component (10m words) Figure 5. native-speaker and EFL novices’ writing and native speech (per million words of running text) . the lexical items of course.12 also shows that not all learner-specific speech-like lexical items are overused in the writing of native-speaking students. it seems to me. Thus. so expressing effect. certainly. around 1. which) is why (pmw) 200 150 100 50 0 Freq. which suggests that native novice writers do not transfer all spoken features to their 3000 200 150 100 50 0 Freq. of PRO (this. academic component (15m words) Native-speaker student writing: Sub-corpus of LOCNESS (100.5m. really. the findings suggest that the main feature shared by native and non-native novice writers is a lack of registerawareness. of I think (pmw) Expert academic writing: British National Corpus. of first of all (pmw) 2500 2000 1500 1000 500 0 Freq. 2004a: 152). Figure 5. I think and first of all –are also more frequently used by native-speaker student writers than in expert academic prose. (2004a) who described it as a general ‘novice-writer characteristic of excessive visibility’ (Neff et al. They are even less frequent in native-speaker students’ writing than in academic prose. that.Academic vocabulary in the ICLE 195 Whether a feature is learner-specific or developmental varies from lexical item to lexical item. this/that is why. sentence-final though. by the way and I would like/want/am going to talk about are quite rare in LOCNESS. words) Native speech: British National corpus.14 The overuse of I think in both EFL learner and native-speaker student writing has already been reported by Neff et al.702 words) EFL learners' writing: ICLEv2 (14L1s.

196 350 300 250 200 150 100 50 0 Academic Vocabulary in Learner Writing 1200 1000 800 600 400 200 0 Freq. of I would like / want / am going to talk about (pmw) 2000 1500 1000 500 0 really of course certainly absolutely Freq. of it seems to me (pmw) Freq. of by the way (pmw) 0 Freq. of amplifying adverbs (pmw) 120 100 80 60 40 20 Freq.12 Continued . of so expressing effect (pmw) 18 16 14 12 10 8 6 4 2 0 Freq. of sentence-final though (pmw) definitely 45 40 35 30 25 20 15 10 5 0 Figure 5. of maybe (pmw) 40 35 30 25 20 15 10 5 0 Freq.

and of L1 frequency – that I referred to as ‘transfer of primings’.g. . It seems that lexical items which are not particularly frequent in speech and are rare in academic prose (e.g. I focused on the potential influence of the first language on multiword sequences that serve rhetorical functions in French learner writing.Academic vocabulary in the ICLE 197 academic writing. Developmental factors in L1 and L2 acquisition cannot. As Gilquin. In the last part of this chapter. so expressing effect). of the phraseological environment. I made use of Jarvis’s (2000) framework for assessing transfer and identified a number of transfer effects – transfer of function. However. possibility *to. however. according to me and as a conclusion). despite *of.g. lexical items that are very frequent in speech. especially at higher levels of proficiency. 1984: 121). but more subtle transfer effects. maybe. Other linguistic features are limited to EFL learners. the issue of the degree of overlap between novice native writers and non-native writers has far-reaching methodological and pedagogical implications and is clearly in need of further empirical study. These include lexico-grammatical errors (*a same. These results support Kellerman’s claim that the ‘hoary old chestnut’ according to which transfer does not afflict the more advanced learner ‘should finally be squashed underfoot as an unwarranted overgeneralization based on very limited evidence’ (Kellerman. be held responsible for all learner specific-features. the first language also plays a part in EFL learners’ use of academic vocabulary. the use of non-native-like sequences (e. Granger and Paquot (2007a: 323) have argued. discuss *about). are very likely to be overused (e. By contrast. of style and register. and acceptable in academic prose. I want/would like/ am going to talk about) are less likely to be overused by native novice writers. they also suggest that the main effect of the students’ mother tongue on higher-intermediate to advanced learner writing is not errors. and the overuse of relatively rare expressions such as in a nutshell. In addition to teachinginduced factors and proficiency.

This page intentionally left blank .

built a list of academic keywords from corpora of expert writing. I discuss some of the important pedagogical implications of this research. and the use of corpora. There are three key aspects: the influence of teaching on learners’ writing. in the development of EAP teaching materials. and more specifically. the role of the first language in EFL learning and teaching. Chapter 7 then briefly summarizes the major results. and suggests several remaining issues and avenues for future research. . discusses some of their implications.Part III Pedagogical implications and conclusions In the first two sections I defined the concept of ‘academic vocabulary’. and analysed their use in ten sub-corpora of the International Corpus of Learner English. In Chapter 6. learner corpora.

This page intentionally left blank .

and the role of corpora in EAP material design.Chapter 6 Pedagogical implications This chapter considers three areas where my findings have major pedagogical implications: teaching-induced factors. in particular data from learner corpora. However this adverb should only be used to indicate that one situation is the exact opposite of another: 6. in a nutshell. 1983. 2007) are also discussed. conversely. Hyland and Milton. Overuse of connectors such as nevertheless. Lake. Flowerdew. The ways in which corpus data. Teaching-induced factors Factors linked to teaching have repeatedly been denounced in the literature as being responsible for a number of learners’ inappropriate uses of connectors (see Zamel. British buyers like brown eggs. and the misleading presentation of the conjunctions even if and even though as synonyms. 1997. Milton. 2004).1). 6. 1990. 1998. 1999). Jordan (1999) describes the adverbial on the contrary as a phrase of contrast equivalent to on the other hand and by contrast (see Figure 6.1. as far as I am concerned. (LDOCE4) Also problematic are the categorization of besides as a marker of concession. This can cause semantic misuse (Crewe. For example. Connectors are often presented in long lists of undifferentiated and supposedly equivalent items. American consumers prefer white eggs. have been used to inform the academic-writing sections of the second edition of the Macmillan English Dictionary for Advanced Learners (MED2) (Rundell. The same is said about conversely. on the one hand.1. the role of the first language in EFL learning and teaching. and on the other hand can also be attributed to the long . classified in broad functional categories.

Contrast. Milton (1999) has shown that there is a strong correlation between the words and phrases overused by Hong Kong students and the functional lists of expressions distributed by tutorial schools (private institutions which prepare most high school students in Hong Kong for English examinations).g. the connectors most frequently used to serve rhetorical functions are sometimes missing from these lists.1 Connectives: contrast and concession (Jordan. .202 Academic Vocabulary in Learner Writing A.g. Concession indicates the unexpected. on the other hand . B. surprising nature of what is being said in view of what was said before: besides (or) else however nevertheless nonetheless notwithstanding only still while (al)though yet in any case at any rate for all that in spite of/despite that after all at the same time on the other hand all the same even if/though Figure 6. .3 that sequences that are rarely used by native speakers (e. .2. as far as I am concerned or last but not least) or ‘unidiomatic’ sequences (e. with what has preceded: instead conversely then on the contrary by (way of ) contrast in comparison (on the one hand) . .1 By contrast. as a conclusion) are sometimes found in teaching materials. Another direct consequence of these lists is EFL learners’ stylistic inappropriateness. 2004: 135) as no information is given about their frequency or semantic properties. especially in the lists of connectors freely available on the Internet. 1999: 136) lists of connectors found in most textbooks (Granger. The selection of connectors to be taught may also lend itself to criticism. It was shown in Section 5. as Milton explains: Students are drilled in the categorical use of a short list of expressions – often those functioning as connectives or alternatively those which are .

Pedagogical implications 203 colourful and complicated (and therefore impressive) – regardless of whether they are used primarily in spoken or written language (if indeed at all). It is most probable that lexical cohesion has been neglected in EFL teaching because ‘there have been no good descriptions of the forms and functions of this phenomenon’ (Flowerdew. Learners’ marked preference for the sentence-initial positioning of connectors has also been related to L2 instruction (see Flowerdew. This example also illustrates the fact that no information about the connectors’ grammatical category or syntactic properties is made available to the learners. register and frequency. The preposition notwithstanding is listed together with adverbs and adverbial phrases (e. that is. and learners use the sentence-initial position as a safe bet.2 Another problem of teaching practices (which has not often been documented) is that too much emphasis tends to be placed on connectors.g. Transfer of primings means that words or word sequences in the foreign language may be primed for L1 use in terms of discourse function. These activities should not be restricted to ‘helping learners focus on errors typically committed by learners from a particular L1’ (Hegelheimer and . Narita and Sugiura. Labels. Thus. Awareness-raising activities focusing on similarities and differences between the mother tongue and the foreign language are clearly needed to achieve this. although. the spoken-like expression all the same is given as an equivalent alternative to more formal connectors such as on the other hand or notwithstanding in Jordan (1999) (see Figure 6. 1998. 1976). to the detriment of lexical cohesion. One of the many roles of teaching should thus be to counter these ‘default’ and sometimes misleading primings in EFL learners’ mental lexicons. collocational and lexico-grammatical preferences.3 However I have shown in this book that nouns.2. Positional variation of connectors is usually not taught. however. The role of the first language in EFL learning and teaching My findings have at least two important pedagogical implications relating to the role of the first language in EFL learning and teaching. 1999. on grammatical cohesion (see Halliday and Hasan.1). verbs and adjectives all have prominent rhetorical functions in academic prose. yet) as well as conjunctions (e. 2006). Milton. 2006: 345). have also been found to fulfil a prominent cohesive role in this particular genre.g. while). 6. or to which text types they are appropriate (1998: 190).

when an ESL class consists of speakers of Chinese. a web-page devoted to linking words and hosted by the ‘Académie de Lille (Anglais BTS Informatique)’ lists according to me as a direct translation equivalent of the French ‘à mon avis’.1 lists examples of infelicitous translation equivalents. This recommendation stands in sharp contrast to Bahns’s (1993: 56) claim that collocations that are direct translation equivalents do not need to be taught. and Yoruba. Persian. However. it is questionable whether the type of contrastive information they provide is fully adequate. Similarly. Yet even in such classes. teachers can help individual students in using any contrastive information that their dictionaries provide. However. They should also raise learners’ awareness of more subtle differences such as the register differences and collocational preferences of similar words in the two languages.1) includes an essay-writing section in which first person plural imperatives in French are systematically translated by structures employing let us in English (Granger and Paquot. moreover. Table 6. Spanish.204 Academic Vocabulary in Learner Writing Fisher. as Odlin commented. I showed that first person plural imperatives are not the best way of organizing discourse and interacting with the reader in English academic writing.4 . one type of contrastive information is frequently available: bilingual dictionaries. there is not likely to be any textbook that contrasts English verb phrases with verb phrases in all of those languages – and even if there were. If the class size allows it. it is not always possible to make use of the first language in the classroom and to rely on contrastive data: Whatever the merits of contrastive materials in some contexts. and as a conclusion as a possible equivalent of the French ‘pour conclure / pour résumer ’. For example. however. For example. the most carefully prepared dictionaries often provide some comparisons of pronunciation and grammar as well. Tamil.3. Although the comparisons are sometimes restricted to words in the native and target languages. (1989: 162) Bilingual dictionaries should ideally facilitate the teacher’s task in multilingual as well as monolingual classrooms. 2008b). it is clear that such materials are not always feasible. the Robert & Collins CD-Rom (Version 1. teachers could not profitably spend the class time necessary to illuminate so many contrasts. In Section 5. Learners have no way of knowing which collocations are congruent in the mother tongue and the foreign language. the differences between the collocations in L1 and L2 may lie in aspects of use rather than form or meaning. 2006: 259).

.. let us note.. let us mention briefly’ = ‘let us add to this or added to this’ Introducing an example Stating facts = ‘(let us) take the case of’ = ‘let’s recall the facts’ Emphasizing particular points = ‘let us not forget that’ . sur Terre.. notons toutefois que le rôle du Conseil de l’ordre a été déterminant Nous reviendrons plus loin sur cette question... as a starting point’ = ‘firstly. that’ = ‘we shall come back to this question later. as well as’ = ‘without dwelling on the details. let us examine’ = ‘after studying .1 Le Robert & Collins CD-Rom (2003–2004): Essay writing French sentence Prenons comme point de départ le rôle que le gouvernement a joué dans l’élaboration de ces programmes En premier lieu. mais signalons déjà l’absence totale d’émotion dans ce passage Avant d’aborder la question du style.. however. Victoria l’Américaine débarque à Londres en 1970 et réussit rapidement à s’imposer sur la scène musicale N’oublions pas que. let us now consider’ = ‘now let us come to’ = ‘let us examine . considérons maintenant le style Venons-en maintenant à l’analyse des retombées politiques Assessing an idea Examinons les origines du problème ainsi que certaines des solutions suggérées Sans nous appesantir or nous attarder sur les détails. but let us point out at this stage’ = ‘before tackling . mentionnons brièvement le choix des métaphores Adding or detailing Ajoutons à cela or Il faut ajouter à cela or À cela s’ajoute un sens remarquable du détail Prenons le cas de Louis dans «le Nœud de vipères» Rappelons les faits. examinons ce qui fait obstacle à la paix The other side of the argument Après avoir étudié la progression de l’action. la gravité pilote absolument tous les phénomènes 205 Essay writing: function Developing the argument Proposed English equivalence = ‘let us take .Pedagogical implications Table 6..

learner corpora are the most valuable resources for designing EAP materials which address the specific problems that EFL learners encounter (see also Flowerdew. Even when they are corpusinformed. 1998 and Tseng and Liou. Hamp-Lyons and Heasley 2006). it is arguably less useful for non-native learners. few make use of authentic texts and very few are informed by the use of corpora. Granger. 2003). as well as the items they tend to under. despite Thurstun and Candlin’s (1998) claim that it is equally appropriate for native and non-native writers. EFL writing is characterized by a number of linguistic features that differ from novice native-speakers’ writing. relies exclusively on data from a native-speakers’ academic corpus. Thus. Multilingual corpora clearly have an important role to play here by providing an empirically-based source of translation equivalents (Bowker. such corpora have very rarely been used systematically to inform EAP materials (see Milton. EAP resources tend to be based on data from native-speakers only. The value of pedagogical tools for non-native speakers of English would be greatly increased if findings from learner corpus data were also used to select what to teach and how to teach it. If not. 2003. there is a danger that the emphasis on teaching the most frequent markers may focus on ones already familiar to and correctly used by students. 2006 for two exceptions in 5 Computer-Assisted Language Learning). decisions made should also be based on findings from a parallel student corpus to ascertain where students’ main deficiencies lie. Bailey 2006. Yet. As Flowerdew (1998) put it. the Longman Dictionary of Contemporary . 2001.3. By showing. Although this is one of the most innovative EAP textbooks to date. 1998: 338). or in this case.g. The only type of resource in which learner corpus data have been relatively successfully implemented up to now is the monolingual learners’ dictionary (MLD). 2009). King. For example. Thurstun and Candlin’s (1997) Exploring Academic English. 6. which uses concordance lines to introduce new words in context and familiarize learners with phraseological patterns. The role of learner corpora in EAP materials design While teaching materials designed to help undergraduate students improve their academic writing skills are legion (e.206 Academic Vocabulary in Learner Writing These findings are quite representative of a general lack of good contrastive studies on which pedagogical materials can be based. in context. As shown in Section 5. ‘when choosing which markers to teach.or overuse.2. the types of infelicities EFL learners produce and the types of errors they make. exacerbate the problem with their overuse’ (Flowerdew.

the countable use of the noun information). 1999: 47). 2002. the sections provide information about how to use these words appropriately by focusing on their: – – – – – semantic properties. 2003). The method used in Chapter 5 has made it possible to identify a number of common features of EFL learners’ expression of rhetorical and organizational functions. Widdowson. but not slaves to. Yet. learner corpora should not only be exploited to compile error notes but also to improve other aspects of the dictionary.2 and 6. (4) expressing cause and effect. ‘materials should be influenced by. (8) introducing topics and related ideas. (6) expressing possibility and certainty. These were identified in Section 4. syntactic positioning. (3) exemplification: introducing examples. A selected list of features were used to inform a 30-page writing section which I and two other members of the Centre for English Corpus Linguistics (CECL). (9) listing items. the confusion between the adjectives actual and current. (5) expressing personal opinions. 1996: 262) of the many lexical means that are available to expert writers to perform a specific function. 2007b: IW1–IW29). Special emphasis is placed on AKL nouns. The writing section includes 12 functions that EFL learners need to master in order to write well-structured academic texts. All the examples come from the academic component of the British National Corpus. designed for the second edition of the Macmillan English Dictionary for Advanced Learners (Gilquin et al. however. . Gaëtanelle Gilquin and Sylviane Granger. style and register differences. frequency. Each writing section includes a detailed ‘corpus-based rather than corpus-bound’ description (Summers. As shown in Figures 6. collocations. adjectives and verbs and their phraseological patterns. (2) comparing and contrasting: describing similarities and differences. (11) reporting and quoting. As put by Cook (1998: 57) referring to Carter’s (1998b) standpoint. (10) reformulation: paraphrasing or clarifying.Pedagogical implications 207 English and the Cambridge Advanced Learner’s Dictionary include a number of learner corpus-informed usage notes which warn against common learner errors (e. (7) introducing a concession. corpus findings’ (see also Swales.3.1 as typically appearing in EAP textbooks which adopt a functional approach to academic writing: (1) adding information.g. if MLDs are to take further ‘proactive steps to help learners negotiate known areas of difficulty’ (Rundell. (12) summarizing and drawing conclusions.

obvious A close analogy can be drawn between cancer of the cell and a society hooked on drugs Figure 6. of clothes and of weapons. ideas. The sections specifically address the types of problems discussed in Chapter 5 — limited lexical repertoire. The orang-utan is the primate most closely related to man. processes. superficial The distribution of votes across the three parties in 1983 bears a close resemblance to the elections of 1923 and of 1929. etc which are similar in some ways. Certain. interesting. There are close parallels here with anti-racist work in education.208 Academic Vocabulary in Learner Writing You can use the nouns resemblance. There is a remarkable similarity of techniques. situations. lack of register awareness. ideas. parallel. phraseological infelicities. Collocation Adjectives frequently used with resemblance and similarity. He would have recognized her from her strong resemblance to her brother. or people. You can also use the noun parallel to refer to the way in which points. and analogy to show that two points. ideas. overuse of connective . they share some characteristics but are not exactly the same: There is a striking resemblance between them. 2007b: IW5) Evidence from learner corpora was used in several ways to inform the writing sections. or situations are similar in certain ways: If there is a resemblance or similarity between two or more points. or people: These theories share certain similarities with biological explanations. semantic misuse. its lively facial expressions show striking similarities to those of humans. striking.2 Comparing and contrasting: using nouns such as ‘resemblance’ and ‘similarity’ (Gilquin et al. situations. or people. remarkable. The noun similarity also refers to a particular characteristic or aspect that is shared by two or more points. Collocation Adjectives frequently used with analogy and parallel close. strong. are similar to each other: Scientists themselves have often drawn parallels between the experience of a scientific vocation and certain forms of religious experience. similarity. ideas. close. usually made in order to explain something or make it easier to understand: A usefull analogy for understanding Piaget's theory is to view the child as a scientists who is seeking a 'theory' to explain complex phenomena. An analogy is a comparison between two situations. situations..

e. i. network failures) should be reported immediately. questions about what we can know and how we can know it. which is relatively rare in academic prose and much more typical of speech (see Figure 6.e. The abbreviation i.e. a graph is used to show that learners have a strong tendency to use the adverb so. ‘that is’ and ‘that is to say’ (Gilquin et al. Note that. That is and that is to say are usually enclosed by commas. Academic writing Freq. per million words 160 140 120 100 80 60 40 20 0 i. the nationalized industries.e.e. in the section on ‘Expressing cause and effect’. Numerous authentic examples are provided to illustrate . that is that is to say Figure 6. four days and nights. Thus.and underuse).. i. There are also ‘Get it right’ boxes which are intended to give guidance on how to avoid common errors. to detain people without charge. that to say. Descartes was obsessed by epistemological questions. that is.’.e.e. (short for 'id est'. and that is are much more frequent than that is to say.Pedagogical implications 209 When you want to explain or define exactly what you mean by something. register confusion and atypical positioning. you can use the abbreviation i. follows a comma or is used between brackets: Network emergencies (i. in the form of graphs which help the reader visualize the differences between learners’ language and that of native writers. Our treatment of these problems is mainly explicit. 2007b: IW9) devices and syntactic positioning.4). These notes are typically supported by frequency data.3 Reformulation: explaining and defining: using ‘i. in academic writing and professional reports. the Latin equivalent of 'that is') or the expressions that is and that is to say: The police now have up to ninety-six hours. it excludes the public sector. First. in that we draw learners’ attention to error-prone items and we provide negative feedback in the form of ‘Be careful!’ notes which focus on problems of frequency (over.

The reader is referred to Gilquin et al. (2007a.210 Be careful! Academic Vocabulary in Learner Writing Learners often use so to express an effect. While Gilquin et al. My investigation of academic vocabulary has shown that the use of learner corpus data. other writing resources. so expressing effect Freq. but it is more typical of speech and should therefore not be used too often in academic writing and professional reports. and their systematic comparison with native corpora. This use is correct. such as textbooks or electronic writing aids. 2007b: IW13) all the points we make. (2007a) for more detailed information on the principles that guided the design of these writing sections. can bring to light a wide range of learner-specific features. per million words 1200 1000 800 600 400 200 0 Academic writing Learner writing Speech Figure 6. 2007b) have shown how these findings can be integrated into a learner’s dictionary.6 could equally benefit from the use of learner corpus data. not limited to grammatical or lexical errors. ..4 Expressing cause and effect: ‘Be careful’ note on ‘so’ (Gilquin et al. but also including over-reliance on a limited set of lexical devices and under-representation of a wide range of typical academic words and phraseological patterns.

from topic introduction to concluding statements. I have therefore argued in favour of a functional . of scientific knowledge. 1997: 5). 2000: 145). These lexical items also contribute to discourse organization and cohesion. however.Chapter 7 General conclusion This book lies at the intersection of three areas of research: English for academic purposes.. incorporating a formal. The chapter concludes with some avenues for future research.1. and as a teaching practice that deals with ‘the teaching of the skills and language that are common to all disciplines’ (Dudley-Evans and St Johns. Academic vocabulary: a chimera? The status and usefulness of EAP has been questioned by Hyland who believes that ‘academic literacy is unlikely to be achieved through an orientation to some general set of trans-disciplinary academic conventions and practices’ (Hyland. I take stock of the main findings of the present study and bring out its major contributions to these three research areas. supports and substantiates the concept of ‘English for (General) Academic Purposes’ both as a macro-genre which subsumes a wide range of text types in academic settings (Biber et al. with proficiency in the language use’ (Jordan. In this final chapter. 1998: 41) and focuses on ‘a general academic English register. My own contribution to legitimizing EAP has been to demonstrate – on the basis of corpus data – that ‘it is possible to delimit a procedural vocabulary of such words that would be useful for readers/writers over a wide range of academic disciplines involving varied textual subject matters and genres’ (McCarthy. and more generally. academic style. Academic texts are characterized by a wide range of words and phrasemes that refer to activities which are typical of academic discourse. learner corpus research and second language acquisition. This book. 1991: 78). 1999). 7.

and have developed a rigorous and empirically-based procedure to select potential academic words for this list. be regarded as an end product. it is not a list of academic vocabulary in a functional sense. As such. range and evenness of distribution. My findings call into question the systematic use of Coxhead’s Academic Word List as the exclusive vocabulary syllabus in a number of recent productivityoriented vocabulary textbooks. the list is the raw result of the application of purely quantitative criteria to native-speaker corpus data. a category which has so far largely been neglected in EAP courses. Teachers should not assume that EAP students know the first 2. lexicogrammatical patterning and phraseology in expert academic prose needs to be carefully described and learner corpus data should be used to . 2009) and proposed the following definition: academic vocabulary consists of a set of options to refer to those activities that characterize academic work.000 most frequent words in English. Following researchers such as Hanciog lu et al. (2008). Each word still needs ‘pedagogic mediation’ (Widdowson. and provides a good illustration of the usefulness of POS-tagged corpora for applied purposes. these words serve important discourse-organizing functions in academic writing. however. The methodology makes use of the criteria of keyness. However. I have therefore ˘ questioned the fuzzy but well-established frequency-based distinction between general service words and academic words. organize scientific discourse and build the rhetoric of academic texts. One important feature of the methodology adopted here is that it includes the 2.. As a result. a large proportion of what has been referred to as academic vocabulary in this book consists of core words. In its current form (see Table 2. Another fact that stands out is that a clear distinction should be made between vocabulary needs for academic reading and writing. Numerous so-called general service words are not mastered productively by L2 learners. I have derived a productive counterpart to the Academic Word List. particularly teaching aimed at productive activities. 2003): its different meanings. Unlike Coxhead’s (2000) definition of the term.17). thus making it possible to appreciate the paramount importance of core English words in academic prose. even at upper-intermediate to advanced levels of proficiency.212 Academic Vocabulary in Learner Writing definition of ‘academic vocabulary’ (Martínez et al. The outcome of this procedure is the Academic Keyword List.000 words of English. This list should not. this suggests that they should be the target of teaching.

a role which is hardly ever mentioned in EFL/EAP teaching. . Biber et al. the method has helped to demonstrate that an essential set of phrasemes in academic prose consists of ‘lexical extensions’ (Curado Fuentes. These partof-speech categories. genres. This procedure has already been applied to the study of words that serve discourse functions (such as exemplifying. the aim of this study. . issue. and registers’. serve organizational functions as diverse as exemplification. 2009a). Milton (1999: 223) commented that ‘a great deal of research [was] still necessary to describe with any empirical rigour the lexis that is characteristic of particular purposes. These words acquire their organizational or rhetorical function in specific word combinations that are essentially semantically and syntactically compositional (e. irrespective of discipline. however. Different disciplines may also have their preferred ways of performing rhetorical or organizational functions. . 2004) and contribute to push ‘the boundary that roughly demarcates the “phraseological” more and more into the zone previously thought of as free’ (Cowie. the next section aims at . As a result. EAP tutors are left wondering how they can possibly meet the needs of all their . Adverbs do not have a monopoly on lexical cohesion and discourse organization in academic writing. Second. an example of .General conclusion 213 complement these descriptions. there has been a huge increase in the number of corpus-based studies highlighting the specificity of vocabulary and phraseology in different academic disciplines and genres. however. A decade ago. 2001: 115) of academic words (e. .g. conclusion. comparing and contrasting) in academic prose. argue). I have shown that a phraseological approach to the description of academic vocabulary provides a mine of valuable information for pedagogical tools. is .. . and expressing cause and effect. claim. As well as their common core features. expressing cause and effect.. verbs and adjectives and their phraseological patterns. it has been suggested) (Oakey. their findings do not easily lend themselves to being used in general EAP courses and it is now essential to find ways of reconciling research findings and the reality of EAP teaching practice. . comparing and contrasting. 1998: 20).. The focus has been on words that are reasonably frequent in a wide range of academic texts and their preferred lexicogrammatical and phraseological patterns. Since then. My results have provided ample evidence for the prominent discursive role of nouns. these words may also have a discipline-specific phraseology (Granger and Paquot. The first result of this method has been to dethrone adverbs from their dominant position as default cohesive markers. 2002. The primary motivation of these studies.g. as discussed below. has not been pedagogical.

1994). history. They do not know either what should be taught. (2007: 216) In a heterogeneous EAP class. for example. We have shown that it is possible to identify both the common core features of an academic word and its discipline-specific characteristics in terms of meaning. With the emergence of a wide range of interdisciplinary curricula. students are better equipped to examine the ways in which grammatical patterns and lexical choices combine to perform rhetorical functions within their own disciplines and hence to apply this knowledge to their own academic writing. this approach allows teachers to emphasize general academic words and phrasemes which ‘are not likely to be glossed by the content teacher’ (Flowerdew 1993: 236). which consists of making use of corpus data as a source of learning materials for language students (Johns. etc. Faced with this difficulty. 2009a) a balanced approach which concurs with Hyland’s (2002b) plea for more specificity in EAP teaching while also subscribing to Eldridge’s view that an essential function of research is to identify ‘similarities and generalities that will facilitate instruction in an imperfect world’ (Eldridge. where disciplinary variability constitutes a serious problem. 2003: 6). especially in mixed classes. 2008: 111). the process of investigation is itself of great value in raising students’ awareness of the patterned nature of academic discourse. The study of ‘individualized’ examples derived from specialized corpora can be of considerable benefit in helping learners to appreciate the possible linguistic realisations of rhetorical and organizational functions in their own disciplines. As Charles put it. phraseological patterns. we have advocated elsewhere (Granger and Paquot. while also empowering learners by giving them the tools to investigate authentic texts and practices . 2003: 8). the problem is likely to become even more acute in the future. that the LSP [Language for Specific Purposes] teacher has the disciplinary knowledge needed to provide reliably accurate instruction in technical varieties of language’ (Huckin. sociology or psychology.214 Academic Vocabulary in Learner Writing students in classes which are ‘often composed of students from different disciplines and/or language backgrounds with different purposes for taking the class’ (Huckin. lexico-grammar. With this understanding. not only for students but also for their teachers as ‘it seldom happens. One way of implementing this ‘happy medium’ approach in the classroom is to apply a data-driven learning methodology. to law students who also have to take courses in economics. although it may not be possible in all teaching situations to provide materials that are specifically tailored to the disciplines of the students taught.

The other type of analysis compares two (or more) interlanguages. ‘thereby allowing considerations of subject specificity and disciplinary variation to inform classroom discussion’ (Groom. to the pedagogical implications that can be drawn from the results – has contributed to fleshing out this concept and has convincingly demonstrated that academic vocabulary is anything but a chimera. and more specifically. 1996) involves two types of comparison. and more precisely argumentative. most studies using the method have been of the first type. . Several of these linguistic features. essays written by upper-intermediate to advanced EFL learners share a number of linguistic features irrespective of the learners’ mother tongue backgrounds or language families. for example native English and the English produced by French-speaking learners. the semantic misuse of connectors and labels. the extensive use of chains of connective devices and a marked preference for placing connectors in the sentence-initial position. I have tried to make the most of CIA by systematically exploiting the two types of comparison it allows to examine EFL learners’ use of academic vocabulary. interlanguage and second language acquisition Contrastive Interlanguage Analysis (CIA) (Granger. 2005: 273). In this book. Although the CIA method has become quite popular. The results show that academic. 7. Learner corpora. other features such as lexico-grammatical errors. or are L1-specific (and so possibly transfer-related). One compares native with non-native (or inter-) language.2. My journey into academic vocabulary – from the extraction of potential academic words through their linguistic analysis in expert and learner corpus data. the lack of register awareness.General conclusion 215 in their own disciplines. The common core of interlanguage features that characterize the expression of rhetorical and organizational functions in EFL writing includes a limited lexical repertoire and a lack of register awareness as well as lexico-grammatical and phraseological specificities. may also be found in novice native-speaker writing. However. for example the English produced by French-speaking learners and the English produced by Italian-speaking learners. Studies comparing more than one IL usually focus on learners from one mother tongue background and use data from one or two other learner populations only to check whether the features they have highlighted in one corpus are common to other learners. the use of non-native-like sequences and the overuse of relatively rare expressions seem to be largely learner-specific.

and should feature prominently in the battery of data types used by all SLA specialists.e. There are many other variables that interact in learners’ interlanguage which are also in need of careful operationalization. Avenues for future research A promising area of research which has only been touched upon in this book lies in the investigation of patterns of difficulty shared by . they arguably provide a good account of the complexity and versatility of L1 influence. 7. I have also shown that it is not always possible to attribute learner-specific features to a single factor. corpus linguistics clearly has numerous resources and specific tools to offer SLA researchers who wish to further investigate the manifestations of L1 influence on learners’ interlanguage. 2004:135–6). Transfer of primings includes L1 influence on collocational use. By focusing on shared features across L1 learner populations. and frequency of use. lexico-grammatical and phraseological patterns. My study has helped to identify a number of transfer effects relating to word use that make up what. because developmental. following Hoey (2005). register differences and phraseology.216 Academic Vocabulary in Learner Writing A systematic analysis of several interlanguages is necessary to analyse the potential influence of developmental. I have highlighted the important role played by developmental and teaching-induced factors in learners’ written production. They are not the exclusive preserve of learner corpus researchers. has rarely been investigated.3. Applying Jarvis’s (2000) methodological framework to learner corpus data has helped identify a number of transfer effects that until now have been largely undocumented in the SLA literature. teaching-induced and transferrelated factors on EFL learner writing. discourse function. i. The valuable theoretical insights provided by a learner-corpus based approach to the study of L1 influence bring to the fore the potential contribution of learner corpora for SLA studies. Learner corpora are probably the best – if not the sole – type of learner interlanguage samples which can be used to investigate these transfer effects. style and register preferences. teaching-induced and transfer-related effects can reinforce each other (Granger. Lexical transfer has too often been narrowed down to transfer of form/meaning mappings and the third aspect of word knowledge. Learner corpora can clearly act as a test bed for studies that aim to provide empirical evidence for theories of second language acquisition. I refer to as ‘transfer of primings’. With its focus on frequency. use. In addition.

and some overlap between native and non-native writers’ (Howarth. There is also an urgent need for learner corpora which represent academic text types other than argumentative essays. All in all. A new corpus currently under development at Louvain. 1999: 151). with NS and NNS students of varying levels of proficiency in between. has been designed as the ESP . Hoey (2005) insisted that primings are constrained by register and genre. and more particularly with its highly conventionalized phraseology.General conclusion 217 mother-tongue English-speaking students and EFL learners. and produce coherent and cohesive texts in a foreign language. longitudinal corpora of learner language are sorely lacking. He gave the example of the word research which is primed in the mind of academic language users to occur with recent in academic discourse and news reports on research. New corpora such as the British Academic Written Corpus and the Michigan Corpus of Upper-level Student Papers are thus particularly welcome. The collocation is not primed to occur in other text types or other contexts. Further research is clearly needed to shed more light on the similarities and differences between EFL learners’ use of academic words and phrasemes and that of novice native-speaker writers. in a variety of disciplines. Novice native-speaker writers have been shown to have difficulty with academic language.1 L1 writing skills also need to figure more prominently in future research. A direct implication of Hoey’s theory of lexical priming is that academic phraseology cannot be assumed to be primed in the mental lexicon of novice native-speaker writers who have had little contact with academic disciplines. much more could be achieved in the field if other types of corpora were collected. both native and non-native speakers. It does not make sense to expect learners to write properly in English. I have shown that the research paradigm of corpus linguistics is ideally suited to studying the lexical specificities of academic discourse in native-speaker and learner writing. Howarth postulated the existence of a continuum of phraseological competence that would ‘encompass mature NS writers at one extreme and weak NNS writers at the other. However. Learner corpus research would greatly benefit from the design of comparable corpora of L1 and L2 writing produced by the same learners. In particular. Such research would enable linguistic features that are characteristic of novice writing to be separated from those features that have commonly been attributed to EFL writing. the Varieties of English for Specific Purposes dAtabase (VESPA) learner corpus. as they consist of ESP texts produced by writers at different stages of undergraduate and graduate level study. The many corpora already available make it possible to examine a wide range of genres and text types. if they cannot already perform this task in their mother tongue.

and the methodological aspects of interlanguage studies. The volume largely focuses on input frequency. I have sought to unify several aspects of English for academic purposes. and implicit vs. intake3. New avenues of research can now be explored by SLA specialists.2 Not a single article in the special issue of Studies in Second Language Acquisition (2002. The role of frequency is a key issue in second language acquisition. Navigating my way through the complexity of each of these research areas. Gregg (2003) only addresses the issue of frequency in relation to the role of input. Similarly. the relative influence of developmental features and transfer effects. The role of L1 frequency is particularly interesting. and experiment with a wide range of tools and methods. Volume 24/2) is devoted to L1 frequency effects and their implications for second language acquisition. and can be expected to be the object of much attention in the next few years. and its relation with language processing. thus restricting his discussion to the question of ‘how often does input of X need to be provided in order for X to be acquired?’ (Gregg.218 Academic Vocabulary in Learner Writing counterpart of the International Corpus of Learner English. and second language acquisition into a coherent whole. learner corpus research. explicit learning. However. There is still so much to explore. It includes English for specific purposes texts written by L2 writers from various mother tongue backgrounds. . in a state-of-the-art article on SLA theory. corpus learner researchers and teachers alike. 2003: 846). it has generally been conceived of in terms of L2 frequency. My journey into academic vocabulary has led me to explore a large number of fascinating fi elds of research. I hope that this book will serve as a starting block for further research into the many issues raised. but the potential influence of L1 frequency on learner interlanguage has also been highlighted. Not only have a number of largely unrecognized transfer effects been brought to light. The challenges presented by such a cross-disciplinary position have quickly been proved worthwhile by the fresh light the approach has shed on key issues such as the nature of academic vocabulary.

4 13.7 16.3 (++) 27. nouns cause cause causes *causae factor factor factors source source sources *sourse origin origin origins *origine root root roots reason reason reasons *reaons *reasongs consequence consequence consequences *consecvencies *consecuence *consecuences *consecuenses *consequencies *consequense *consequenses 314 127 186 1 229 100 129 274 194 78 2 60 48 11 1 173 112 61 939 563 374 1 1 319 76 227 1 2 3 2 4 1 3 14.7 6.6 1.802 1105 697 − − 450 223 269 − − − − − − − 54.5 83.Appendix 1: Expressing cause and effect Comparisons based on total number of running words ICLE Abs.2 15 81.3 92. Rel.5 35.6 23. BNC−AC−HUM Abs.6 4. LogL 19.175 577 598 − 500 286 214 − 183 72 111 80.4 40.2 (− −) 5.2 (++) 22.2 (− −) 5.3 (++) Rel.8 26.9 755 492 263 − 550 244 306 1.1 (++) (Continued) .6 87.

3 9.2 211 (++) 3.4 4. 1.9 (++) 2.3 2.4 (− −) 499 140 106 220 33 51 25 10 14 2 116 61 20 21 13 1 14 3 2 9 0 20 8 4 3 5 0 42.8 (− −) LogL effect effect effects efect result result results *resut outcome outcome outcomes implication implication implications TOTAL NOUNS verbs cause cause causes caused causing bring about brings brings brought brining contribute to contribute contributes contributed contributing *contribuates generate generate generates generated generating give rise to give gives gave given giving 395 214 179 2 381 167 213 1 28 21 7 12 4 8 3.2 . 33.220 Appendix 1 ICLE Abs.5 20.8 10 8.4 170.03 (− −) 1 12.8 67.8 4.5 Rel.124 32.8 0. Rel.3 2.612 259.2 227 63 23 119 22 6.7 101 23 21 32 18 7 3 6.6 1.9 BNC−AC−HUM Abs.7 24.4 570 133 66 317 54 125 44 6 64 11 276 52 18 82 26 − 17.830 1249 581 813 502 311 − 143 135 8 411 93 318 268 8.4 (− −) 1. 55 84.

67 19 5 35 8 671 161 105 334 71 115 14 13 82 6 4. 1.7 221 LogL 31.2 (− −) (Continued) .1 (++) 46 (− −) 115.3 BNC−AC−HUM Abs.Appendix 1 ICLE Abs.1 (− −) 666.3 114 30 33 32 5 0.3 161 38 11 102 10 − − − 327 104 18 138 67 88 31 16 34 7 5.2 Rel.2 4.5 Rel.6 15 7 8 0 0 356 184 83 72 17 12 4 2 3 3 4.8 0 30.8 (++) 1 3.7 171 145 31 28 30 4 52 14.9 0.4 9. 2 2.7 39.5 22.1 (− −) 2.6 20. induce induce induces induced inducing lead to lead leads led leading prompt prompt prompts prompted prompting provoke provoke provokes provoked provoking provocate provocated provoqued result in/from result results resulted resulting yield yield yields yielded yielding make sb/sth do sth# arise from/out of arise arises arose arisen arising derive derive derives derived deriving *derivated 39 12 8 15 3 1 489 8 4 2 2 0 0 3.2 50 14 8 16 8 1 2 1 8.4 476 77 68 297 34 − 2 2 0 0 0 42 0.

3 (− −) 0.8 (++) 0.174 125.9 5.7 7 (− −) Rel.5 53 344 397 1.7 (++) 10. 10 171 181 0.4 (++) .6 21.9 0.9 13.1 6.2 (− −) LogL emerge emerge emerges emerged emerging follow from 33 2.3 (++) 4.7 66.7 13.6 17 0.7 0 2 1.6 68 14 22 23 9 4.4 0.7 0.3 1 0.8 11 6 15 1 4 follow follows followed following 1 0 2 1 8 trigger triggers triggered triggering 5 0 3 0 7 stem stems stemmed stemming 1 5 0 1 1.6 95 599 195 196 22 1 66 52 109 35 22 24 1.6 (++) 175.8 229.8 0.6 3.6 (− −) 360.1 0.847 158.7 56 14 3 27 12 2.3 0. 14 126.1 (++) 1.1 31. 466 107 95 221 43 74 33 35 5 1 1.19 433.6 (− −) 0.1 8.5 0.7 39.7 1. Rel.2 23.6 0.321 18 5.7 (− −) trigger stem from TOTAL VERBS adjectives consequent responsible (for) TOTAL ADJ.7 15.9 14.89 531 246 79 7 5 8 17 7 199 3 7 1.3 BNC−AC−HUM Abs.109 45.6 10. prepositions because of due to as a result of as a consequence of in consequence of in view of owing to in (the) light of thanks to on the grounds of on account of TOTAL PREP.4 12 3.5 0.222 Appendix 1 ICLE Abs.1 (++) 3 0. 2.

3 7.4 18.6 31.5 0.5 54.9 1121 955 883 1.7 28.2 15.4 326.9 (− −) 2.1 180 41.8 3 0.2 1.1 (− −) 243.066 .3 8.407 28.4 1. Rel.4 2.3 25 (− −) 809. Adverbs therefore therefore *therefor accordingly consequently consequently *consecuently thus hence so thereby as a result as a consequence in consequence by implication TOTAL ADVERBS conjunctions because because *becausae *becaus since## as ## 223 BNC−AC−HUM LogL Rel.8 (− −) 55.9 17. 701 689 12 26 183 179 4 446 42 1.8 (++) 36.7 (−) 132.9 (++) for so that PRO is why that is why this is why which is why on the grounds that TOTAL CONJ.1 (++) 989.1 1.9 4.9 16.036 696 52 22 18 12 83 5.74 26.7 (++) 24.8 (++) 43.3 (++) 2.7 130 143 3.894 182 101 20 14 35 5.2 1.2 21 1.436 15 103 35 11 0 2.912 26.1 (++) 2.4 (++) 38.6 123.3 3.4 1553.Appendix 1 ICLE Abs. TOTAL 220 189 18 12 5 3.1 (++) 1 325.6 0.207 66.56 0.3 (− −) 457.998 60.2 (− −) 33.2 1.7 (++) 34.767 283 1.5 3 0.7 0.493 1 1 428 331 58 273 214 2.495 2.2 8.9 0 257.3 (++) 3.4 5 23.412 42.4 359 (++) 381.5 178 794.981 53.5 1 0.8 21.5 57 5. Abs.810 13.7 (++) 0.

3 7.2 2.1 570 125 276 227 101 67 671 115 161 327 88 171 145 476 466 74 56 68 4.5 (− −) 8.174 2.3 0.6 (− −) 16.802 450 1.7 6.9 499 51 116 14 20 15 356 12 50 114 2 489 8 39 33 4 8 7 1.175 500 183 1.9 0.5 9. prepositions because of due to as a result of as a consequence of 531 246 79 7 4.5 1.9 755 550 1.7 0.830 813 143 446 8.6 1.1 4. nouns cause factor source origin root reason consequence effect result outcome implication TOTAL NOUNS Verbs cause bring about contribute to generate give rise to induce lead to prompt provoke result in yield make sb/sth do sth# arise from/out of derive emerge follow from trigger stem TOTAL VERBS adjectives consequent responsible (for) TOTAL ADJ.9 (− −) 56 (− −) 463.8 84.9 (− −) 9 (− −) 1.1 0.1 1.847 3.6 (− −) 16.9 0.9 0.8 24.4 0.9 0.4 (− −) 274 (− −) 231.5 1.9 (− −) 4.1 0.6 1.7 0.1 0.4 1.6 (− −) 0 0.5 0.9 145.9 2.1 39.1 0.8 1.1 14.8 0.2 0.6 0.8 0.7 32.1 0.224 Appendix 1 Comparisons based on total number of ‘cause and effect’ lexical items ICLE Abs.55 1.1 599 195 196 22 2.9 0.8 2.4 2.5 (++) 263.5 1.5 (− −) 23.612 2. % LogL .2 0.4 0.3 0.3 0.8 (− −) 0.4 53 344 397 0.1 10 171 181 0.3 (− −) 36.3 2.4 3 2.2 1.2 (− −) % BNC−AC−HUM Abs.1 0.4 (++) 1.8 1.3 0.2 106.2 (− −) 314 229 274 60 173 939 319 395 381 28 12 3.1 23.4 1.3 (− −) 153.7 0.1 93.0 3.124 2.4 (++) 1.8 (− −) 14.5 (− −) 192.7 (− −) 201 (− −) 36.7 0.7 6.2 0.3 (++) 2.3 0.6 (++) 71.1 0.9 (− −) 10.4 0.1 1.3 (++) 95.1 2.9 3.9 0.2 0.7 23.0 0.2 0.3 15.3 1.4 0.7 0.6 6.3 1.

0 0.3 (− −) 507. .3 (++) ## Estimations based on Gilquin (2008).3 22.0 0.7 164.7 1.45 20.5 0.2 1.5 Abs.1 0.1 0. in consequence of in view of owing to in (the) light of thanks to on the grounds of on account of TOTAL PREP.6 3.5 6.0 29.9 19.1 0.8 (++) 1. TOTAL # 225 BNC−AC−HUM LogL % 0.9 (++) 73.894 182 101 20 14 35 5.0 21.1 22.2 100 2.7 (++) 6 1.1 1.1 (− −) 2.5 (++) 14. 1 66 52 109 35 22 24 1321 % 0 0.3 2.4 0.1 0.998 5.5 0.1 0.412 130 143 1.495 428 331 58 273 220 189 28 3 5 3.1 0.6 (++) 2.1 3.1 7.4 (++) 1.4 50.3 11 0.6 (++) 182. Estimations based on an analysis of the first 200 occurrences of the word in each corpus.9 2.1 0.4 (− −) 158.1 8.Appendix 1 ICLE Abs.4 3.3 0.92 (−) 262.036 696 52 22 18 12 83 5.6 0.0 0.981 5.4 (− −) 72.407 8.4 0.066 19.1 0.7 0.1 1.8 (++) 294.7 0.2 (− −) 144.5 0.436 15 103 35 11 0 2.2 0.4 (− −) 26.810 13.6 0.3 2.0 0.912 26.767 283 1.3 0.7 (− −) 70.3 3.4 0.1 0 23 1.7 1.2 (++) 21.4 0.2 (− −) 270.59 (− −) 10. adverbs therefore accordingly consequently thus hence so thereby as a result as a consequence in consequence by implication TOTAL ADVERBS Conjunctions because since## as## for so that PRO is why that is why this is why which is why on the grounds that TOTAL CONJ.3 39.2 0.8 0.4 100 790.1 (++) 5 8 17 7 199 3 7 1109 701 26 183 446 42 1.4 3.4 0.1 (− −) 0.2 0.1 5 6.3 28.4 2.2 0.1 0.207 955 883 1.

6 9.3 116 100 16 212 106 106 − − 147 76 71 19 10 9 − − 175 133 42 522 470 52 311 249 62 − − 1.1 0.6 1.3 2.7 14.4 0.2 0 2.4 2.1 33.6 3.3 3.19 3.4 7.3 54 (− −) 2 5.3 0 2.5 6.3 0.3 4 1.3 15.2 (− −) 54.7 24.49 3 0.2 0. LogL 4.9 (− −) Rel.19 35. BNC−AC−HUM Abs.1 0.5 0.9 82. nouns resemblance resemblance resemblances similarity similarity similarities *similarieties *similiraty parallel parallel parallels parallelism parallelism parallelisms *paralelism *parallelim analogy analogy analogies contrast contrast contrasts comparison comparison comparisons *comparaison *comparision difference difference differences *differencies *difference 3 3 0 25 18 7 38 36 0 1 1 394 187 191 6 3 3 1 0 1 1 6 6 0 2 2 0 25 7 16 1 1 0.1 0.3 0.2 1.9 (− −) 178.1 0.Appendix 2: Comparing and contrasting Comparisons based on total number of running words ICLE Abs.1 0 0.3 (− −) 49.5 1.5 0.1 1.1 0 0.1 0.318 802 516 − − 3.5 0 0.5 8 (− −) .38 3.1 15.1 0.3 (− −) 39.5 0. Rel.4 0.6 0.3 0.1 0.8 16 16.

9 30.1 25.9 0. − − − − − − − 76 72 4 − 595 498 97 10 559 − 28 27 1 85 58 27 56 4.058 160 157 2 1 1 275 16 12 5 23 1 1.4 (− −) 8.5 (− −) 55 1055 223 137 52 98 63 2.1 1.3 12.1 0.7 127.1 3.2 0 0.8 0.1 75.4 73.7 30.3 (− −) 268 (++) 2.9 15 2.8 6.3 (− −) 10.1 130 129.2 21.1 0.5 0.2 0.496 2496 − − − 72 1.6 (− −) 283.1 0.6 0.6 Rel.1 0.3 2.1 0.1 23.3 0.3 16.17 22. 227 LogL 28.9 (− −) 3.8 0.7 4.2 0.6 1.1 0.4 0.6 4.2 0.75 (− −) (Continued) .1 0.5 (+) 2.Appendix 2 ICLE Abs.6 3 1.3 30.229 0.7 (− −) 18.7 13.515 1510 2 1 2 4 90.1 0.7 31.1 3.027 77.3 (− −) 148.78 3.1 1.4 2 0.1 0.9 75.8 3 2.8 13.3 0.4 1 0.9 3 2 0 1 47 38 9 2 246 1 17 16 1 44 40 4 5 860 1 1 1 1 1 1 1 Rel.4 (− −) 59.8 (++) 110.1 17.3 2.5 1.3 0.1 0. 0.8 BNC−AC−HUM Abs. *diference *difference *difference *differency *differene *difference *diffrences differentiation differentiation differentiations *differenciation distinction distinction distinctions distinctiveness (the) same *similars (the) contrary contrary contraries (the) opposite opposite opposites (the) reverse TOTAL NOUNS Adjectives same similar similar *similiar *simmilar analogous common comparable identical parallel alike contrasting different different *differents *differrent *diffrent differing 1.580 1.8 (− −) 20.4 0.2 1.1 0.1 4.8 (− −) 31.

3 (− −) 9.4 4.8 3.3 (− −) 6.3 0.2 0.4 (− −) LogL distinct distinct *distinc distinctive distinguishable unlike contrary opposite reverse reverse *reversed TOTAL ADJECTIVES verbs resemble resemble resembled resembles resembling correspond correspond corresponded corresponds corresponding look like look like looks like looked like looking like compare compare compared compares comparing parallel parallel parallels paralleled paralleling contrast contrast contrasted contrasts contrasting 9 7 2 13 2 2 7 53 7 3 4 3.3 1.52 2.1 0.1 0.1 0.4 BNC−AC−HUM Abs.4 3.2 1.3 0.6 0.1 3.4 0.7 0. 8.7 4.2 1.5 0.8 0.1 6.2 0.5 0.4 11 (− −) .5 1.6 4.1 5.7 (− −) 45.1 11.1 1.2 0.7 40.8 1.9 0.1 0.1 111.4 4.9 (++) 6.4 6.3 0.1 0 0 0.3 0.2 0.552 257.6 0.5 1.2 4.5 1.2 2.3 0. 278 278 − 163 33 43 27 127 23 Rel.4 0.4 8.6 0.6 9.1 0.9 (− −) 14.4 0.3 0 0 138 51 18 46 23 137 73 16 48 28 102 42 38 19 3 278 140 71 17 50 56 9 4 38 5 137 31 47 42 17 4.4 31 16 3 11 1 41 27 3 4 7 106 72 21 12 1 129 75 36 2 16 2 1 0 0 1 7 3 4 0 0 2.5 1.1 6.6 0.3 0.1 8.4 0. Rel.6 0.8 0.9 (− −) 0.3 0.5 1.2 1.163 8.1 2.7 1.3 0.1 0.6 (+) 21.2 1.228 Appendix 2 ICLE Abs.0 0.8 58.1 1. 0.3 271.8 3.1 0.9 1 1.

6 0.1 0.2 0.5 (− −) 2.3 0 3.4 0.1 0.1 0 0 0 1.1 35.5 1.8 (− −) 3.8 3.7 0.9 (− −) 13.1 1.8 1.01 5 0.2 0.1 0.1 0.8 11. 7.4 9.7 0.5 47.5 0.3 62.2 0.9 2.2 1.6 3.2 1.1 0.2 45.9 3.6 9 (− −) 4.7 (− −) 54. 242 112 73 57 404 164 116 36 88 − − 74 22 6 31 15 − 1.7 2.3 (− −) 7.8 15. 7.16 4.4 (− −) 0.9 5.9 394 11.1 0.2 0.7 12.3 0.7 0.2 6 1.4 4.1 1.6 1.1 2.7 (− −) 2.5 0 9.1 0.6 0.2 2.1 0 0.568 Rel.2 17.29 3.2 1.2 0.4 2.6 (− −) 2 2 29 − 118 56 3 97 185 116 69 0 23 14 9 69 − 4 25 372 0.4 0.1 0. differ differ differs differed distinguish distinguish distinguished distinguishes distinguishing *distinquish *distingush differentiate differentiate differentiates differentiated differentiating *differenciate TOTAL VERBS adverbs similarly similarly *similarely *similarily *similary analogously identically correspondingly parallely likewise in the same way contrastingly differently by/in contrast by contrast in contrast by way of contrast by/in comparison by comparison in comparison comparatively comparatively *comparitively contrariwise distinctively on the other hand 31 26 1 1 3 1 0 0 1 9 38 0 42 9 2 7 1 0 0 0 14 13 1 0 1 418 2.09 1.4 2.9 3.7 98.4 (− −) − 30.0 24.9 0.3 (− −) 258.3 34.3 (+) 1.2 2.3 (++) (Continued) .2 0.0 0.2 0.8 0.5 0.1 (− −) 1.9 86 57 29 0 107 70 16 12 6 2 1 18 12 1 2 1 2 527 Rel.3 (− −) 9.0 0.1 0.9 0.3 2.Appendix 2 ICLE Abs.1 0 0 0.4 1.7 229 LogL 0.2 BNC−AC−HUM Abs.7 13.1 0.

6 (− −) 8. Rel.5 0.157 206 99.4 0.4 104.4 (− −) 124.812 244 8 121 46 82 73 9 53 66 0 0 52 14 4 21 14 3.230 Appendix 2 ICLE Abs.1 0.4 12.6 0.3 0.3 14 BNC−AC−HUM Abs.5 (− −) 0.3 3.250 0 1.7 7.7 0.9 0 0 133.435 26 0 7 0 23 15 8 7 18 2 3 39 28 11 0 0 1.560 123.8 37.5 2.6 0 2 1.5 76.9 (++) LogL (on the one hand) *on the other side *on the opposite on the contrary on the contrary *on the contray *on the contrairy Other expressions with contrary *in contrary *by the contrary *to the contrary quite the contrary *in the contrary rather the contrary *quite contrary *contrary reversely conversely TOTAL ADVERBS Prepositions like# unlike in parallel with as opposed to as against in contrast to/with in contrast to in contrast with versus contrary to *in contrary to *opposite to by/in comparison with in comparison with in comparison to by comparison with in comparison with TOTAL PREP.6 0.1 0 0 0 0 4.5 (++) 45.8 (++) 62 (++) 8.9 29.4 (− −) .2 3.484 84.4 (− −) 27.1 2.9 38 185.9 38. Conjunctions as # while # 100 23 3 164 160 3 1 13 1 1 2 4 2 1 1 1 1 6 875 1.9 4 3.7 (− −) 0.7 Rel.4 0.6 1.1 0 0 2.1 0.6 2 0.1 (+) 30.3 0.8 (− −) 4.3 1.8 2.5 7.4 2.045 1264 151. 136 0 0 95 95 − − 2 0 0 0 2 0 0 0 0 0.7 (++) 1.9 8.1 0.7 0 62 1.1 (++) 158. 8.9 (− −) 231. 4.5 (++) 14.7 (+) 12.4 (− −) 62 (++) 1.3 17.2 0.4 2.2 0 0.7 5.1 0 0 0 0.1 (++) 12.9 127.2 0.6 2 0 0 1.6 1.

854 110.751 203.3 11 (++) 12.6 0.26 1.5 0. 1.5 0.249 83.3 0.2 0.26 1 0.4 21.4 0.4 1.5 2.2 1. 442 Rel.6 4. 11. as in the same way as/that compared with/to compared with compared to CONJ compared to/with as compared to/with when compared to/with if compared to/with TOTAL 1.287 19 49 12 37 14 5 3 6 9.3 (− −) 15.6 231 LogL 6.14 4.5 845.500 Rel.8 (+) 0.3 0.5 (++) 1.3 (− −) 67.2 281. 13.67 3.8 11.7 BNC−AC−HUM Abs.0 880.Appendix 2 ICLE Abs.2 2. .6 0.2 128.3 1.0 3. whereas whereas wheras TOTAL CONJ.766 38 155 113 42 32 11 20 1 29.4 1. Other expressions as .5 137 135 2 1. .24 (− −) .

Nouns resemblance similarity parallel parallelism analogy contrast comparison difference differentiation distinction distinctiveness (the) same (the) contrary (the) opposite (the) reverse TOTAL NOUNS Adjectives same similar analogous common comparable identical parallel alike contrasting different differing distinct distinctive distinguishable unlike contrary opposite reverse TOTAL ADJECTIVES Verbs resemble correspond look like compare parallel 31 41 106 129 2 0.5 0.5 52.1 19.8 1.7 (+) 20.2 0.5 0.6 (−) 1.7 0.4 0.2 4.2 0.1 0.1 1.2 (++) 8.1 0.7 116 212 147 19 175 522 311 1.8 0.7 (− −) 202.3 (− −) 1.8 0.3 2.9 3.318 76 595 10 559 28 85 56 4.7(− −) 15.5 0.5 0.9 (− −) 45.1 (− −) 0.027 55 1055 223 137 52 98 63 2.1 0.1 2.5 0.5 0.0 0.2 0.4 26.5 (− −) 168.2 14.580 1.8 (− −) 3.1 8.6 0.82 (++) 2 25 6 3 3 25 38 394 3 47 2 246 17 44 5 860 0.0 2.8(− −) 24.9 (− −) 138.9 (− −) 0.2 8.3 0.0 0.496 72 278 163 33 43 27 127 23 8.5 0.1(− −) 56.5 0.4 1.4 1 0.5 5.1 0.1 11.229 0.3 0.4 0.7 (− −) 9.2 (− −) 10. % LogL .3 (− −) 51.1 (− −) 4.1 0.3 0.0 0.1 0.1 29.4 4 0.1 0.6 29 (− −) 307.1 0.1 0.0 2.0 0.9 0.5 (− −) 106.8 (− −) % BNC−AC−HUM Abs.5 0.4 0.7 1.2(++) 98.0 0.0 0.1 0.163 10.6 0.2 0.1 32.515 4 9 13 2 2 7 53 7 3.5 0.5 63.5 11.7 0.3 1 0.0 0.0 1.3 (− −) 14.3 0.1 4.4 1.5 0.2 (− −) 37.552 0.6 0.1 1.0 15.2 3.7 (++) 21.232 Appendix 2 Comparisons based on total number of ‘comparison and contrast’ lexical items ICLE Abs.2 28.8 79.0 0.1 (−) 2.3 0.6 (− −) 32.2 (− −) 29.3 0.5 0.2 0.058 160 1 275 16 12 5 23 1 1.0 138 137 102 278 56 0.

7 5.0 0.3 8.4 BNC−AC−HUM Abs.3 0. 137 242 404 74 1.4 0 0.3 2.1 0.0 0.2 1.6 0 92.3 0.1 0.3 (++) 166.2 0.8 (− −) 275.812 244 8 121 46 82 53 66 0 9.4 0.8 (++) 33 (++) 63.1 0 0.0 1.8 1.250 1.3 3.3 (− −) 4.1 1.9 0. contrast differ distinguish differentiate TOTAL VERBS Adverbs similarly analogously identically correspondingly parallely likewise in the same way contrastingly differently by/in contrast by contrast in contrast by way of contrast by/in comparison by comparison in comparison comparatively contrariwise distinctively on the other hand (on the one hand) *on the other side *on the opposite on the contrary Other expressions with contrary reversely conversely TOTAL ADVERBS Prepositions like# unlike in parallel with as opposed to as against in contrast to/with versus contrary to *in contrary to 1.1 394 2 2 29 0 118 56 3 97 185 116 69 0 23 14 9 69 4 25 372 136 0 0 95 2 0 62 1.2 0.0 0.5 (Continued) .3 (− −) 10.2 0.1 0.7 (− −) 0.1 0 0.6 0.6 6.4 0.3 0.6 7 86 107 18 527 % 0.1 0 0.5 0.0 2.9 (− −) 0.0 0 0 0.8 0.3 (− −) 0.0 0.4 (++) 1.2 (++) 2.0 0.2 5.8 13.7 35.3 (− −) 26.2 0.0 4.8 (++) 25.1 1.3 5.0 0.3 0.1 1.1 2.5 0 0 0.1 0.2 0.4 (− −) 0 0.2 0.4 0.4 (− −) 0.4 0.0 0 0.4 (++) 8.8 28.4 233 LogL 42.0 0.6 0.9 (− −) 0.8 12 (−) 258.2 4.0 0 0.8 (++) 42.3 0 0.2 0 31 1 0 0 1 9 38 0 42 9 2 7 1 0 0 0 14 0 1 418 100 23 3 164 13 1 6 875 0.8 (− −) 2.8 59.7 0.7 1.3 0.435 26 0 7 0 23 7 18 2 14.1 0.2 0.4 0.6 (++) 155.1 8.Appendix 2 ICLE Abs.2 5.9 1.568 % 0.2 16.0 0.

560 1.1 0.3 1.7 0.751 17.2 0.3 4.5 (− −) 110. Other expressions as … as in the same way as/that compared with/to CONJ compared to/ with TOTAL 3 39 1.7 231.157 206 137 1.2 0.4 15.2 11.5 0.766 38 155 32 29.5 0.6 (− −) 1.1 100 2.2 5.1 1.3 (+) 13.1 150. Conjunctions as# while# whereas TOTAL CONJ.5 0.4 (+) 83.500 11.5 23.8 BNC−AC−HUM Abs.045 1264 442 6.854 13.0 0.9 8.1 100 87. % 0.6 (− −) 0.9 (++) LogL *opposite to by/in comparison with TOTAL PREP.6 # Estimations based on an analysis of the first 200 occurrences of the word in each corpus.8 (++) 2.7 2. .484 % 0 0. 0 52 3.1 0.4 15.287 19 49 14 9.249 9.234 Appendix 2 ICLE Abs.

lancs. This specific set of abstract nouns has variously been referred to as ‘signalling words’ (Jordan.html (accessed 2 August 2009).uib. 1986). Social Sciences. ‘anaphoric nouns’ (Francis. The four corpora are equivalent in the sense that they were compiled using the same corpus design and sampling EAGLES96/corpustyp/node18. Reading and Oxford Brookes under the directorship of Hilary Nesi and Sheena Gardner (formerly of the Centre for Applied Linguistics at Warwick University). and there may be only 2. For more information about these corpora. it has 75. See http://ucrel. with funding from the ESRC (RES-000-23-0800). Oxford Brookes). 2000) and ‘discourse-organising words’ (McCarthy. 1991).000 different words (called types) in the .cnr.000 tokens. The BAWE corpus contains 2761 pieces of proficient assessed student writing. ‘carrier nouns’ (Ivanic ˇ. But a lot of these words will be repeated. If a text is 75. Holdings are fairly evenly distributed across four broad disciplinary areas (Arts and Humanities. Reading) and Paul Wickens (Westminster Institute of Education. Life Sciences and Physical Sciences). 1991). ‘shell nouns’ ( Chapter 2 1 2 3 4 5 6 7 The BAWE Pilot Corpus was a pilot for the ESRC funded project ‘An investigation of genres of assessed writing in British higher education (RES-000-23-0800).hit.000 words long. Sentence examples are taken from the Longman Dictionary of Contemporary English (2005) See the definition of a reference corpus proposed by the Expert Advisory Group on Language Engineering Standards (EAGLES96) at http://www. Each of these corpora consists of one million words of British or American written English. The British Academic Written English (BAWE) corpus was developed at the Universities of Warwick.Notes Chapter 1 1 2 See Stein (2008) for a review of major twentieth-century projects aimed at developing a controlled vocabulary for foreign language (accessed 2 August 2009). Thirty-five disciplines are represented. Paul Thompson (Department of Applied Linguistics.ilc. see http://khnt.html for a list of tags used in CLAWS C7 tagset (accessed 2 August 2009). It was created in 2001 under the directorship of Hilary Nesi. 1984). with support from the University of Warwick Teaching Development Fund.

but falling as low as 40 per cent for Spanish speakers (Granger et al.lextutor. ‘multiple occurrences of a content word or phrase in a single-text document. which is contrasted with the fact that most other documents contain no instances of this word or phrase at all’. causes problems for certain types of linguistic enquiries. 2000). Turkish and Tswana mother tongue backgrounds (cf. Norwegian. but only for words in a single file. 2009: 11–12). especially when the lexical items under study are closely linked to specific parts of texts (e. Quantitative comparisons between the BNC and ICLE thus have to be treated with caution. See Stefan Evert’s webpage (http: //www.html (accessed 2 August 2009)) for a comprehensive list of measures of association and their mathematical interpretation. Chapter 4 1 2 3 In f[n. n the node and c the collocate. the ‘close proximity of all or some individual instances of a content word or phrase within a document exhibiting multiple occurrences’.236 8 Notes 9 10 Katz (1996: 19) distinguishes between ‘document-level burstiness’. Chapter 3 1 2 3 4 5 A random sample of 20 essays from each of the 16 L1 sub-corpora available in the second version of ICLE were submitted to a professional rater who was asked to rate them on the basis of the Common European Framework of Reference for Languages (CEF) descriptors for writing. Japanese. words and phrasemes used to introduce the main topic or a conclusion).000 words were sampled so as to allow for a wider coverage of text types and avoid over-representation of idiosyncratic uses. i. ICLE also comprises a Bulgarian sub-corpus. f is the frequency. These three nouns are listed under the first sense of ‘classic’ in LODCE4. Granger et al. i. This design criterion. based on an arbitrary division of a text into 8 segments of equal size. 2009). reaching 100 per cent for students with Swedish mother (accessed 2 August 2009).e. A number of studies in the field of English for academic purposes have shown that words may behave differently and display different preferred lexicogrammatical environments in different sections of a text (see.g. the proportion was much higher in some (accessed 2 August 2009). essays written by Bulgarian-speaking learners were mainly written without the help of reference tools and were therefore not included in the analysis. c]. Scott’s (2004) WordSmith Tools 4 can compute Juilland’s D values. and ‘within-document burstiness’ or ‘burstiness proper’.collocations. Texts longer than 45. ICLEv2 now also includes texts written by students with Chinese. While 60 per cent of the sample essays were rated as advanced (C1 or C2).. for example. Gledhill. Available at http://www. http://www. . however.

. The underuse of the conjunctions as and while reported here must be treated with caution as it results from estimations based on an analysis of only the first 100 occurrences of each conjunction in each corpus. which learners then tend to work into their essays.Notes 4 237 5 6 7 These figures are based on disambiguated data. that there are no idioms. the differences in use are only significant for a few groups. Estimations based on an analysis of the first 200 occurrences of the preposition in the BNC-AC-HUM. Siegel (2002) and Biber et al. AKL words are printed in bold in these examples. (1999: 562) for specific functions of like in speech. which was widely used in French colleges throughout the 1980s and early 1990s. similes. See Miller and Weinert (1995). however. commonplaces and allusions to proverbs and quotations in academic prose. 2004) in specialized texts. proverbs. overuse in learner corpora in general. though moderate. Studies focusing on terminological terms used in English for Specific Purposes have also revealed the pervasiveness of compounds (e. This does not mean.. The relative frequencies of for instance and example are higher in most learner corpora than in the BNC-AC-HUM in most learner corpora. Estimations based on an analysis of the first 200 occurrences of the conjunction in the BNC-AC-HUM. The instances of illustrate used in the sense of ‘to put pictures in a book. although they may be close and intimate and friendly and all that. See Müller (2005: 197–228) for an analysis of like as a discourse marker. however. Chapter 5 1 2 3 4 The ‘word list’ option of WST4 was used to search for any misspelt form of the words under study in the ICLE. As shown by Gläser. Nathan 1984). are not the same as a relationship between members of a family. phrasal verbs. So we’ve got some examples here of some patterns that we want to learn using the N tuple method and tuple and tuple. ‘authors of scientific writing are prone to modify idioms. compounds. Other verb co-occurrents that are quite frequent in the BNC-SP but not found in the BNC-AC-HUM are the verbs get and think. Bourigault et al. Aggregated frequencies thus also help to reveal general. etc’ are not included. (BNC-SP) 5 6 7 8 The noun root is overused in the ICLE largely because it appears in an essay title given to some of the EFL learners. article. When the learner corpora for different mother tongues are analysed separately. France) kindly pointed out to me that the sequence according to me also appeared in published textbooks such as Ok! (Lacoste and Marcelin. and quotations for intellectual punning and sophisticated allusions’ (1998: 143).g. relationships between members. John Osborne (Université de Haute Savoie. (BNC-SP) Again think of the example of erm erm a social club you know. ‘In the words of the old song: “Money is the root of all evil”’.

See Gilquin et al. By contrast. According to von Mayer. Osborne (2008) compared adverb placement in the various interlanguages represented in the first version of the International Corpus of Learner English and found that ‘V-Adv-O order is most frequent in the productions of learners whose L1 has verb-raising ( methodes/ Expressions_et_mots_de_liaison.(2007a) for a detailed discussion of the role of corpora. 242–62) as well as a full chapter on ‘Grammar and Academic English’ (pp.1. Czech and Bulgarian) in between’ (Osborne. they sometimes erroneously use a comma after the conjunctions although or (even) though (e. with speakers of nonraising languages (Russian. As shown in Section 4.htm (last accessed: 30 July 2009). It is noteworthy that. When I compare these languages I do not consider English as an easy language. http://www2c. although there is a chapter on textual cohesion (‘Grammar across turns and sentences’. following Granger and Paquot (2008a). although.g. I prefer to avoid using the adjective ‘collocational’ to refer to sequences of co-occurrents. Cohesion is often dealt with in grammars. ICLE-IT). 2006). The ‘Improve your writing skills’ section in the MED2 shows how a rigorous corpus-based method can help users achieve higher levels of accuracy and fluency . Italian and Spanish). 2008: 77). See Paquot (2008b and in preparation) for details on the corpus linguistics methods and statistical measures used to operationalize Jarvis’s (2000) framework on learner corpus data.2. and more specifically. the information is often neither corpus-based nor confirmed by corpus data. what matters is relative poverty* that is to say* the sudden decrease of wealth. Chapter 6 1 2 3 4 5 6 The quality of the teaching material on the use of connectors in English that is freely available on the Internet is generally quite alarming. no attention is given to lexical cohesion. 266–94). [P] indicates a new paragraph in learner writing A related problem is that of punctuation. in the new corpus-based Cambridge Grammar of English (Carter and McCarthy. learner corpora in the design of EAP materials and for possible explanations of the relatively modest role that corpora have played so far. I do admit that I have noticed some things that are easier about English than about the other languages that I had the chance to learn. when the preferred sentence position of individual connectors is taught. where the focus is always on connectors. especially given that students increasingly use the Internet for study purposes. ICLE-PO).g. pp. German and Swedish).ac-lille. and least frequent with speakers of V2 languages (Dutch. The figures should be treated with caution as the LOCNESS corpus is quite small. however. The results reported here are only preliminary. Polish.238 9 Notes 10 11 12 13 14 Gledhill (2000) uses the term ‘collocational cascade’ but. EFL learners sometimes omit commas after sentence-initial subordinate clauses or connectors or before and after appositives such as that is and that is to say (e.

1992). . Chapter 7 1 2 3 The Centre for English Corpus Linguistics launched the LONGDALE project in January 2008. the Louvain EAP Dictionary (LEAD) (see Granger and Paquot. distinguish between input and intake as follows: ‘“Input” is everything around us we may perceive with our senses. to achieve maximum efficiency.. with the intention of building a large longitudinal database of learner English containing data from learners with a wide range of mother tongue backgrounds. 2008b and 2010). However.).Notes 239 in academic writing. the same students will be followed over a period of two to three years. De Bot et al. etc. medicine. and “uptake” or “intake” is what we pay attention to and notice’ (2005: 8). Kamimoto et al. it is essential to explore ways of integrating this type of description into the microstructure of dictionaries rather than inserting it as a separate middle section. 1992: 211. This project is innovative in two main respects: it allows for both onomasiological (via the lexeme) and semasiological (via the concept) access and is customizable according to the learner’s mother tongue and the field in which he or she is specializing (business. In the LONGDALE project. The major role of L1 frequency has been identified in a few transfer studies focusing on phonology and syntax (Selinker. The Centre for English Corpus Linguistics (Université catholique de Louvain) has therefore recently launched a new dictionary project which consists of a web-based EAP dictionary-cum-writing aid tool.

and Schreuder. Computer Learner Corpora. (ed. 1–15. pp.). Amsterdam and Atlanta: Rodopi. . D. Archer. Phraseology: Theory. K. 247–57. (2006). Amsterdam: John Benjamins. S. Ädel. Second Language Acquisition and Foreign Language Teaching.pdf. (ed. in Granger S. Bailey. (2009a). 53. (eds). pp.). B. Edinburgh: Edinburgh University Press. Aarts.). (2008). Aijmer. D. ‘The use of adverbial connectors in advanced Swedish learners’ written English’. (2002). A. (2006). Bahns. Learner English on Computer. 55–76. London and New-York: Addison Wesley Longman. P. (1998). Aston.comp. and Granger. S. Farnham: Ashgate. ‘I think as a marker of discourse style in argumentative Swedish student writing’. in Archer. D. Amsterdam: Rodopi. pp. 35–53. Language Learning and Language Teaching 6. in Aijmer.. Analysis and Applications. (2001). M. Hung. Introduction to the USAS category system. and Diez-Bedmar B. in Gilquin.). (ed. B. J. ‘Morphological influences on the recognition of monosyllabic monomorphemic words’. Ädel. (ed. D. Baayen. 496–512. G. Studia Linguistica. Oxford: Oxford University Press. ELT Journal. H. (eds). S.). pp. Amsterdam and Philadelphia: John Benjamins. Altenberg. The BNC Handbook. Papp. Studies in Honour of Göran Kjellmer. ‘On the phraseology of spoken English: the evidence of recurrent word-Combinations’. S. (1998). in L.) (2009). Göteborg: Acta Universitatis Gothoburgensis. L. Altenberg. Farnham: Ashgate. (2002). G.. Wilson. What’s in a Word-list? Investigating Word Frequency and Keyword Extraction. pp. A. 38. 101–22. 47 (1). From the COLT’s Mouth . Aijmer. pp. J.References1 Aarts. ‘Modality in advanced Swedish learners’ written interlanguage’. 132–41. A. ‘Involvement features in writing: do time and interaction trump register awareness’.. and Tapper. ‘Causal linking in spoken and written English’. K. J. A Wealth of English.. P. and Burnard. Breivik and A. in Granger S. Altenberg. London and New York: Routledge. Available from http://www. (ed. Metadiscourse in L1 and L2 English. . (ed. . (1993). B. London and New York: Addison Wesley Longman. (1998). ‘Lexical collocations: a contrastive view’. Archer. Learner English on Computer. Hasselgren (eds). ‘Does corpus linguistic exist? Some old and new issues’. ‘Tag sequences in learner corpora: a key to interlanguage grammar and discourse’. (2002). What’s in a Wordlist? Investigating Word Frequency and Keyword Extraction. A. in Granger. pp. 20–69. Journal of Memory and Language. and Petch-Tyson. ‘Does frequency really matter?’. Linking up Contrastive and Learner Corpus Research. Academic Writing: A Handbook for International Students (2nd edition). Archer. R. and Others. pp. K. S. (1998). M. F. J. in Cowie. (2006). and guide. 80–93. (1984). 1–17. R.lancs. Feldman. 56–63.

Working with Specialized Text: A Practical Guide to Using Corpora. Vocabulary and Applied Linguistics. H. Harlow: Longman. (2004). S. Bartning. 4. Applied Linguistics. Available from http:// w3. 32 (4). Baker. AILE. Barkema. . 9. Leech. Amsterdam and Philadelphia: John Benjamins. D. S. S. 1–17. G. Longman Grammar of Spoken and Written English. S. Language Learning. P. A. ‘If you look at . Conrad. (eds). D. P.. Out of Corpora: Studies in Honour of Stig Johansson. (2002). and Oksefjell. Biber. and Béjoint. (1992). Biber. L. Fourth Edition. D. S. J. ‘Querying keywords: questions of difference. (2004). in Flowerdew. pp.: lexical bundles in university teaching and textbooks’. M. N. E. Bhatia. Actes del I Symposium Internacional de Lexicografia. 25 (3). and Charlet. Biskup. (eds). Billurog lu.. London: Macmillan.univ-tlse2. S. ‘The comparative fallacy in interlanguage studies: the case of systematicity’. ‘L1 influence on learners’ renderings of English collocations: a Polish/ German empirical study’. and Pearson. Amsterdam: Rodopi. De lexicografia. International Journal of Lexicography. Variation across Speech and Writing. (1988). BNL 2709: The most commonly used words in Eng˘ lish. D. Bauer. D. in Hasselgård. ‘Sub-technical vocabulary and the ESP teacher: an analysis of some rhetorical items in medical journal articles’. Bley-Vroman. in Freedman. Bowker. ‘Idiomaticity and terminology: a multi-dimensional descriptive model’. ‘Corpus-based applications for translator training: exploring the possibilities’. 371–405. (eds). 181–90. R. pp. and Petch-Tyson. Revue d’Intelligence Artificielle. Beheydt. Bourigault.) Academic discourse. Journal of English Linguistics. (2005). ‘Systems of genres and the enactment of social intentions’. (1999). Amsterdam and Philadelphia: John Benjamins. London and New York: Routledge. L. pp. I. 253–79. ‘A generic view of academic discourse’. 241–50. Barcelona: Institut universitari de linguistica applicada. and Medway. S. (2003). P. 85–93. (1983). (1996). and Conrad. P. and Cortes. (ed. (2002).. J. (1997). Biber D. pp. J. and Finegan. ‘The development of an academic vocabulary’. P. H. Corpus-based Approaches to Contrastive Linguistics and Translation Studies. D. S. 125–60. Lerot. 91–105. A. Genre and the New Rhetoric. in V. (1988). (2004). (1993).. S. Biber. L. in Arnaud. H. Cambridge: Cambridge University Press. 87–110. Série actvitats 15. Studia Linguistica. Conrad. 18 (1). ‘L’apprenant dit avancé et son acquisition d’une langue étrangère: tour d’horizon et esquisse d’une caractérisation de la variété avancée’.References 241 Baker. L. C. ‘Word families’. . London: Taylor and Francis. 79–101. University Language: A Corpus-based Study of Spoken and Written Registers. 9–50. . pp. ‘Lexical bundles in conversation and academic prose’. (eds). and DeCesaris.. and Nation. frequency and sense in keyword analysis’. Bazerman. (2007). Harlow: Longman. 6 (4). J. Biber. V. I. Aussenac-Gilles. 21–39. Johansson. (1999). 50 (2). . (eds). 169–83. in Granger.. ‘Construction de ressources terminologiques ou ontologiques à partir de textes: un cadre unificateur pour trois études de cas’. (1994). pp. Reading in a Foreign Language. Nicosia: Rüstem Kitabevi. and Neufeld. (2006). 346–59.doc Bowker. 33. D. J.

Amsterdam: John Benjamins. and McCarthy. 52 (1): 57–63.. 11 (1). and Larsen-Freeman. (eds). R. Text and Technology: In Honour of John Sinclair. Cook. (1998b). D. ‘The learning and use of academic English words’. Reference Guide for the British National Corpus (XML edition). 43–56. Berlin: Erich Schmidt Verc C lag. System. Contrastive Rhetoric: Cross-Cultural Aspects of Second-Language Writing. R. 26 (2). Wellington: NZCER. Cambridge: Cambridge University Press. Amsterdam: John Benjamins. 3rd Conference on Applied Natural Language Processing. (1998 [1987]) Vocabulary: Applied Linguistic Perspectives (2nd edition). E. M. J. and Eskey. pp. Francis. (1971). 52 (1). International Journal of Corpus Linguistics. ‘The importance of corpus-based research for language teachers’. M. and culture’. S. ‘Lexis and discourse: vocabulary in use’. 47 (4). ELT Journal. Ferrera.. Spoken and Written English Grammar and Usage. ‘The use of conjunctive adverbials in the academic papers of advanced Taiwanese EFL learners’. L. Corson. . Carter R. (2006). (1993) ‘From Firth principles: computational tools for the study of collocation’. Available from http:// citeseer.natcorp. Phraseologie. 201–20. Language Learning. D. R. E. J. Conrad. Glasman. Pratiques. ‘Technical vocabulary in specialised texts’. (1998). R.html Burger. (1997). ‘Corpus Linguistics. C. G. 152–67. Eine Einführung am Beispiel des Deutschen. (1998). ‘Introduction et gestion des exemples dans les textes à thèse’. ELT Journal. (1988).. Carter. 27. and McCarthy. pp. J. in Carter. An Academic Vocabulary List. Carter. Charles. Carter.) How to Use Corpora in Language Teaching. pp. in Baker Campion. M. (ed.. G. Language Variation. 1–18.. Cohen. in Sinclair. (1999). The Grammar Book: An ESL/EFL Teacher’s Course (2nd edition). and Fine. ‘Reading English for specialised purposes: discourse analysis and the use of student informants’. New York: Longman. (1996). communication. Available from http://www. J. E. (1988). ‘The uses of reality: a reply to Ronald Carter’. J. Cambridge Grammar of English: A Comprehensive Guide. 15 (2). W. D. 58. (1999). Boston: Heinle and Heinle. Conrad. Proceedings of ANLP-92. Clear. (1992). (2007). Devine.psu. B. M.nflrc. and Nation. Available from http://www. Cambridge: Cambridge University Press. T. New York: Cambridge University Press. English for Specific Purposes. 67–85. (2004). pp. Rosenbaum-Cohen.hawaii. 671–718. U.242 References Brill. London: Routledge. ˇˇ Burnard. Reading in a Foreign Language. D. P. ‘Orders of reality: CANCODE. Celce-Murcia. McH. Chen. and Tognini-Bonelli.ox. W. (2006). 271–92. (eds). ‘Argument or evidence? Disciplinary variation in the use of the noun “that” pattern in stance construction’. (1988). (eds). in Carrrell P. ‘A simple rule-based part of speech tagger’. S. 23–41. Interactive Approaches to Second Language Reading. and McCarthy. Vocabulary and Language Teaching. 203–18. Coltier. H. (2007). D. H. (2003). E. M. and Elley. and Language Teaching’. P. Chung. 113–30.

ELT Journal. A. and Saville. 72–86. K. W. and Moran. in Teubert. The Native Speaker: Myth and Reality. ‘Phraseological dictionaries: some east-west comparisons’. DeRose. Eldridge. ‘Recurrent sequences of words in native speaker and advanced learner spoken and written English: a corpus-driven approach’. M. 14. (2003). Cowan.References 243 Cortes. S. Cutting. (1998). in Flowerdew. (ed. Analysing Learner Language. V. Research Perspectives on English for Academic Purposes. 5 (3). and Peacock. ‘Lexical and syntactic research for the design of EFL reading materials’. A. and Verspoor. ‘The illogic of logical connectors’. Assessment. M. The Corpus Approach to Lexicography. Coxhead. in Olesky. P. Cambridge: Cambridge University Press. A. M. Phraseology: Theory. Developments in English for Specific Purposes. S. ‘Written errors of international students and English native speaker students’. 8 (4). and St Johns. pp. Computational Linguistics.. Contrastive Pragmatics. D. G. (2007). S. T. W. London and New York: Routledge. (2000). Frankfurt am Main: Peter Lang. Tse’s “Is there an ‘academic vocabulary’?”’. D. ‘Lexical behaviour in academic and technical corpora: implications for ESP development’. (eds). TESOL Quarterly. Council of Europe (2001). TESOL Quarterly. (1989). Curado Fuentes. Davies. (1990). Analysis and Applications. R. and Hirsh. Oxford: Oxford University Press. J. J. in Reppen. and Raupach. M. Louvain-la-Neuve: Université catholique de Louvain. Fitzmaurice. in Blue. S. (2001). p. pp. Oxford: Oxford University Press. (forthcoming). M. 34 (2). Second Language Productions. 213–38. Ellis. (1998). J. (eds). R. 44 (4). pp. De Cock. ” A reader responds to K.. ‘Collocational blends of advanced language learners: a preliminary analysis’. H. P. in Cowie. Lowie. K. M. Clevedon: Multilingual Matters. and Nation. Coxhead. in Dechert.. Cambridge: Cambridge University Press. A. Crewe. . S. ‘A new Academic Word List’. Teaching. Language Learning and Technology. H. ‘Computer learner corpora and monolingual learners’ dictionaries: the perfect match’. H. (ed.. 97–113.’ but . Coxhead. (1974). Second Language Acquisition: An Advanced Resource Book. (2002). 31–9. ‘Grammatical category disambiguation by statistical optimization’. and Lennon. J. A. P. A. De Cock.. Amsterdam: John Benjamins. (eds). Common European Framework of Reference for Languages: Learning. (1988). Revue française de Linguistique Appliquée. pp. (2004). Special issue of Lexicographica. (eds). 106–29. Dechert. ‘The specialised vocabulary of English for Academic Purposes’. Hyland and P. (2005). J. TESOL Quarterly. Assessing English for Academic Purposes. and Barkhuizen. (2000). G. Milton. Using Corpora to Explore Linguistic Variation. Amsterdam: John Benjamins. W. 252–67. (2001). 211–30. ‘“No. 131–68. J. (eds). ‘Second language production: six hypotheses’.). D. (1984). M. there isn’t an ‘academic vocabulary. De Bot. Möhle. Tübingen: Gunter Narr. P. P. Byrd. Cambridge: Cambridge University Press. R. J. W. 109–13. . 389–400. Bunting. A. 316–25. ‘A pilot science-specific word list’. 42 (1). (2003). Dechert. 12 (2). J. 131–45.. 20. and Granger. A. (2005). ‘Lexical bundles in Freshman composition’. and Mahlberg.). Boston: University of Michigan Press. . Cowie. (2008). Dudley-Evans. The Academic Word List: Collocations and Recurrent Phrases. pp. 209–28. and Biber. Unpublished PhD thesis. 65–78. Coxhead.

Linking Up Contrastive and Learner Corpus Research.collocations. 30–41. London and New York: Longman. and Wagner. Leech. L. ‘From EFL to ESL: evidence from the International Corpus of Learner English’. Available from http://www. ‘The integrated contrastive model. Evans. English for Specific Purposes.D. 345–62. ‘Vocabulary in ESP: a lexical analysis of the English of electronics and a study of semi-technical vocabulary’. Ph. communication and (some) fundamental concepts in SLA research’. 363–79. International Journal of Corpus Linguistics. (2006).de/phd. (2008). Amsterdam and Philadelphia: John Benjamins. (ed. ‘On discourse.. (eds). 285–300. Modern Language Journal. and Granger. 3–17.. pp. Gilquin. (1997). 8 (3). R. Francis. Amsterdam and Atlanta: Rodopi.). G. S. (1979). Harlow: Longman. Studies in Corpus Linguistics 29.html Farrell. (2008). Y. and Smith. ‘Signalling nouns in a learner corpus’. (eds). Birmingham: English Language Research. (1998). ‘A comparison of internal conjunctive cohesion in the English essay writing of Cantonese and native speakers of English’. pp. 1–17. G. Flowerdew. P. 23 (1). System. 11 (3). ‘Combining contrastive and interlanguage analysis to apprehend transfer’. in Coulthard. Gilquin. (eds). and Sampson. G. Ghadessy. 3–33. L. Garside. (eds). G.O. N. Institut für maschinelle Sprachverarbeitung. The Computational Analysis of English. ‘Introduction: approaches to the analysis of academic discourse in English’. ‘The fallacy of word counts’. and Roseberry.17. J. 1–83. CLCS Occasional Paper. (1997). Corpus Annotation: Linguistics Information from Computer Text Corpora. Flowerdew. pp. Amsterdam: John Benjamins. ‘The statistics of word cooccurrences: word pairs and collocations’. (1993). B. word lists and materials preparation: a new approach’. (1968). (1987). J. J. Flowerdew. 24–7. pp. Languages in Contrast. S. ‘Frequency counts. (2000/2001). (1986) Anaphoric Nouns. R. 3 (1). A. M. RELC Journal. and Papp. 25. ‘Integrating ‘expert’ and ‘interlanguage’ computer corpora findings on causality: discoveries for teachers and students’.. ‘The exploitation of small learner corpora in EAP materials design’. (2001). Gilquin. in Ghadessy. G. 213–31. (2006). Small corpus studies and ELT. Paper presented at the First Triennial Conference of the . 21 (2). University of Stuttgart. ‘Labelling discourse: an aspect of nominal-group lexical cohesion’. in Garside. L. (1990). (1999). University of Birmingham.K. L. 243–64. ‘Problems in writing for scholarly publication in English: the case of Hong-Kong’. and Green. C. (2008) Corpus-based Analyses of the Problem-Solution Pattern: A Phraseological Approach. International Review of Applied Linguistics. pp. 95–123. R. J. Evert. ‘The CLAWS word-tagging system’. Flowerdew. Advances in Written Text Analysis. G. and Yip. Garside. ‘A hybrid grammatical tagger: CLAWS4’. Journal of Second Language Writing. (1994). Firth. 17 (4). New York: Addison Wesley Longman. 329–345. Flowerdew.244 References Engels. 81. S. (2004).M. M. in Diez-Bedmar. M. in Flowerdew. thesis. in Garside. ‘Concordancing as a tool in course design’. Gilquin. 102–21. ‘Why EAP is necessary: a survey of Hong Kong tertiary students’. Journal of English for Academic Purposes. G. English Teaching Forum. M.) Academic Discourse. Flowerdew. J. (ed. and McEnery. Spicing up your data’. J. (2002). Discourse Analysis Monograph 11. 10. L. Field. Francis. G. Leech. 231–44. R. A. London and New-York: Routledge. G. Flowerdew. 6. S. R. 15–28. (1992). pp. 82–101.

‘From CA to CIA and back: an integrated approach to computerized bilingual and learner corpora’.). Proceedings of an International Symposium. and Paquot. Granger.. 3–33. English for Academic and Technical Purposes: Studies in Honour of Louis Trimble. Languages in Contrast: Text-based Cross-linguistic Studies. 3–18. London and New York: Addison Wesley Longman. Granger. pp. J. E. Second Language Acquisition and Foreign Language Teaching. S. Language Learning and Language Teaching 6. (eds). 538–46. (1998a). Amsterdam and Atlanta: Rodopi. Amsterdam and Philadelphia: John Benjamins. 123–45. E. (1997). Gläser.. Germany. Oxford: Oxford University Press. (ed. A. M. (1981). Stockholm: Almqvist and Wiksell International.). pp. 37–51.. Tarone. (2003). S. (1996a). 37 (3). Analysis and Applications. TESOL Quarterly. Learner English on Computer. and Paquot. (ed. S. P. G. Oxford: Oxford University Press. Granger. . I. J. Gilquin. Amsterdam and Atlanta: Rodopi. (eds). Gledhill. in Cowie. ‘On identifying the syntactic and discourse features of participle clauses in academic English: native and non-native writers compared’.. J. 319–35. (2002). (ed. Granger.. G. (2007a) ‘Learner corpora: the missing link in EAP pedagogy. and Upton. London and New York: Addison Wesley Longman.). S. Phraseology: Theory. in Rundell. in Svartvik. S. (eds). Computer Learner Corpora. Studies in English Language and Teaching. (ed. 8–11 October 2008. S. (2000). (editor in chief) Macmillan English Dictionary for Advanced Learners (2nd edition). Collocations in Science Writing. ‘The International Corpus of Learner English: a new resource for foreign language learning and teaching and second language acquisition research’. in Selinker L. Granger. M. Albert-Ludwigs-Universität Freiburg. Granger. in Thompson. ‘Romance words in English: from history to pedagogy’. and Wekker. Granger. pp. Granger. pp. pp. pp. and Johansson. (eds). University of Hanover. (eds).). (2007b).). in Granger. (2006). S. Lund: Lund University Press. Granger. (ed. ‘A taxonomic approach to the lexis of science’. H. Analysis and Applications. Gilquin. ‘The stylistic potential of phraseological units in the light of genre analysis’. M. 105–21. Phraseology: Theory. B. Tuebingen: Gunter Narr Verlag. ‘Too chatty: learner academic writing and register variation’. 1 (1). Altenberg.) (1998). Words. ‘Computer learner corpus research: current status and future prospects’. S. in Aijmer K. G. S. S. Granger. in Aarts. and Paquot. P. P. M. 125–43. ‘The computer learner corpus: a versatile new source of data for SLA research’. Language in Performance 22. S. 41–61. Granger. S. V. (2004). S. A. ‘Improve your writing skills: writing sections’. 5–7 October 2006. 23–39. English Text Construction. (1998b). Hung. 145–60. and Petch-Tyson. (1998). pp. (ed. Learner English on Computer. ‘A bird’s-eye view of learner corpus research’. (1996b). ‘Lexico-grammatical patterns of EAP verbs: how do learners cope?’ Paper presented at Exploring the Lexis-Grammar Interface. Applied Corpus Linguistics: A Multidimensional Perspective. A.References 245 International Society for the Linguistics of English. Goodman. in Cowie. Corpus-based EAP pedagogy. Gilquin. (2008). T. pp. in Connor. and Hnazeli.. pp. IW4–IW28. R. C. U. pp. and Payne. Rowley MA: Newbury House. S. Lund Studies in English 88. 6 (4). Oxford: Macmillan Education. Special issue of the Journal of English for Academic Purposes. de Mönnink. M. 185–98. S. ‘Prefabricated patterns in advanced EFL writing: collocations and formulae’. Granger. in Granger.

K. Learner English on Computer. (2010). University of Crete Publications. (2008b). Gries. M. Hamp-Lyons. Granger. ‘Procedural vocabulary in law case reports’. S. Journal of English for Academic Purposes. K. new applications. S. (2003). and Paquot. ‘From dictionary to phrasebook?’. 1 (2). Academic Writing: At the Interface of Corpus and Discourse. ‘Lexical verbs in academic discourse: a corpus-driven study of learner use’. (2008a).). (2009a). S. 6. Granger. 257–77. and Paquot. (ed. M. (2007).246 References Granger. Granger. ‘Connector usage in the English essay writing of native and non-native EFL speakers of English’. Continuum.. S.ucl. J. Granger S. Handbook and CD-ROM. Granger. L. Meunier. ‘In search of General Academic English: a corpusdriven study’. in Katsampoxaki-Hodgetts. (1998). in Granger. . 23. in Doughty. (eds). 19–29. in Granger. 193–214.pdf Granger. 459–79. Dagneaux. R. pp. Gregg. The International Corpus of Learner English. and Hasan. (2005). ‘SLA theory: construction and assessment’. (2008). 15– Groom. and Paquot. 108–20. (1997). Harris. E. 15–19 July 2008. Handbook of Second Language Research. (1988). pp. 831–65. pp. English for Specific Purposes. Available from http://cecl.S.. M. pp. Proceedings of the eLex2009 Conference.. ‘Dispersions and adjusted frequencies in corpora’. London: Longman. S. and Paquot. S. M. and Meunier. M. 119–31. Granger. S. S. ‘The contribution of learner corpora to second language acquisition and foreign language teaching: a critical evaluation’.P practitioners Conference academic_english. The International Corpus of Learner English. 1345–55.). (2006). S. S. Amsterdam and Philadelphia: John Benjamins. pp. Cohesion in English. and Paquot. CD-ROM and Handbook. E. and Hunston. ‘Pattern and meaning across genres and disciplines: an exploratory study’. M. F. Dagneaux.. R. and Swallow. ‘Customising a general EAP dictionary to learner needs’. ‘Through the looking glass and into ˘lu.). and Eldridge. H. ‘Exploring variability within and between corpora: some methodological considerations’. S. (2009). Louvain-la-Neuve: Presses universitaires de Louvain. (eds). Langage et l’Homme. Hanciog N. 109–51. M. London: Blackwell. M. S. 27 (4). Spain. English for Specific Purposes. in Charles. and Rayson. Granger. Proceedings of the XIII EURALEX International Congress. 16 (4). S. Version 2. Presses universitaires de Louvain: Louvain-laNeuve. Granger. ‘Disentangling the phraseological web’. S. and DeCesaris. S. the land of lexico-grammar’. F. and Long. 13 (4). (2009b). and Tyson. (2008). 403–37. S. (eds). (eds) (2002). and Heasley. M. 94–108. in Bernal. (eds) eLexicography in the 21st century: new challenges. Barcelona. N. Pecorari . Phraseology: An Interdisciplinary Perspective. Options and Practices of L. 27–49. F. S. J.). K. (ed. and Meunier. Halliday. 4. Study Writing: A Course in Writing Skills for Academic Purposes. Louvain-la-Neuve: Presses universitaires de Louvain. Corpora. World Englishes.fltr. E. Amsterdam and Philadelphia: John Benjamins. C.. pp. H. M. ‘Automatic profiling of learner texts’. Granger. P. Cambridge: Cambridge University Press. Corpora and Language Teaching. London and New York: Addison Wesley Longman. and Paquot. (ed. (2009). 15. in Aijmer. Cahiers du Cental. 289–308. S. in Granger. S and Paquot. (eds. E-media. B. D. ‘False friends: a kaleidoscope of translation difficulties’. International Journal of Corpus Linguistics. Gries. (1976). Neufeld. pp. (1996).

P. (2006). pp.). N. C. P. ‘Are low-frequency complex prepositions grammaticalized? On the limits of corpus data – and the importance of intuition’. 171–210. D. Analysis and Applications. Hoey. ‘Lexical teddy bears and advanced learners: a study into the ways Norwegian students cope with English vocabulary’. Kohn. in Cowie. Teaching Academic ESL Writing: Practical Techniques in Vocabulary and Grammar. (1996).References 247 Harris Leonhard. 26–45. New Zealand: Victoria University of Wellington. V.evert/PUB/HoffmannEvert2006. (eds).. . E. S. ‘A common signal in discourse: how the word reason is used in texts’. Hunston. 5. Range [Computer software]. (1996). 177–95. Corpus Linguistics with BNCweb – a Practical Guide. ‘Adverbial markers and tone in L1 and L2 students’ writing’. Peter Lang. (1999). London: Lawrence Erlbaum Associates. ‘The phraseology of learners’ academic writing’. ‘Phraseological standards in EAP’. S.victoria. Nottingham: Nottingham University Press. J. T. Available from http://purl. Smith. Huntley. (eds). and Fox. 23 (2). and Evert. G. (ed. (1993). London: Routledge. G. (eds) Academic Standards and Expectations: The Role of EAP. CALICO Journal.. Heatley. Techniques of Description. pp. Oxford: Oxford University Press. New Tools. pp. M. Howarth. Y. Hoey. and Berglund Prytz. (2002).aspx Hegelheimer. Tübingen: Max Niemeyer Verlag. ‘BNCweb (CQP-edition): The marriage of two corpus tools’. A. Hoey. Journal of Pragmatics. Hinkel. (2003). and technology: a sample technology-supported approach to teaching grammar and improving writing for ESL learners’. in Sinclair. M. and Mukherjee. (1994). 1049–68. M. (eds). A. and Mair. D. M. Evert. 237–60. pp. Available from http://www. Ibérica. P. S. Oxford: Oxford University Press.. 3–17. writing. ‘Specificity in LSP’. (2002). and Thompson. volume 3 of English Corpus Linguistics. 4 (2). K. in Bool H. P. 35 (7). (2004). Frankfurt am Main. Hoffmann. Boston: Houghton Mifflin Company. M. in Coulthard. 257– stefan. (2005). 689–96. E. NJ: Lawrence Erlbaum Associates. Hasselgren. Hirsh. and Nation. London and New-York: Routledge. Phraseology in English Academic Writing: Some Implications for Language Learning and Dictionary Making. M..pdf Hoffmann. Bonston: Heinle and Heinle. and Fisher. A. ‘Signalling in discourse: a functional analysis of a common discourse pattern in written and spoken English’. ‘Grammar. Hinkel. H. Advances in Written Text London and New York: Routledge. J. Lexical Priming: A New Theory of Words and Language. (1998). Hinkel. Hoey. Discoveries in Academic Writing. 149–58.). Second Language Writers’ Text: Linguistic and Rhetorical Features. P. Corpus Technology and Language Pedagogy: New Resources. and Luford. Lee D. Frankfurt am Main: Peter Lang. P. Essential Academic Vocabulary: Mastering the Complete Academic Word List. S. New Methods. ‘What vocabulary size is needed to read unsimplified texts for pleasure?’ Reading in a Foreign Language. (2004). P. 67–82. Wellington. Corpus Approaches to Grammaticalization in English. Mahwah. 161–86. International Journal of Applied Linguistics. S. in Braun. (2006). (2006). and Nation. S. (1994). Amsterdam and Philadelphia: John Benjamins. Huckin. Howarth. (2008). pp. Phraseology: Theory. Evaluation in Text: Authorial Stance and the Construction of resources/range.. Howarth. B. S. (2003). 8. (1992). H. Hoffmann. in Lindquist. (eds) (2000). E. (ed.

183–205. K. (1996). (ed. K. 30. (1964). (1991). K. R. K. and Petch-Tyson. (2005). M. (2007). London and New York: Continuum. 437–55. Krishnamurthy. 50 (2). Jordan. Hyland. Katz. Rhetoric of Everyday English Texts. 8 (3): 251–77.) Second Language Writing. Frequency Dictionary of Spanish Words. pp. S. Crosslinguistic Influence in Language and Cognition. E. ‘Persuasion and context: the pragmatics of academic metadiscourse’. pp. S. Criper. and Milton... N. English for Academic Purposes.. Kamimoto. and Howatt. International Review of Applied Linguistics in Language Teaching. (1992). and Tse. Edinburgh: Edinburgh University Press. London: Allen and Unwin. Cambridge: Cambridge University Press. Study Skills in English. 6. in Kroll. (1984). (1998). A. and Rodriguez. I. Corpus-based Approaches to Contrastive Linguistics and Translation Studies. Hyland. in Granger. Interlanguage. Cambridge: Cambridge University Press. Hyland. J. (2008). La Haye: Mouton. in Schreibman. Jarvis. ‘Parallel concordancing and its applications’. ‘Specificity revisited: how far should we go now?’ English for Specific Purposes. 140–53. S. pp. B. Academic Discourse. ‘Methodological rigor in the study of transfer: identifying L1 influence in the interlanguage lexicon’. (2005). (2000). and Unsworth. (1990). (eds). K. Göteborg: Acta Universitatis Gothoburgensis. 6 (2). (2003). 98–122. B. 356–73.. Lerot. 293–313. ‘Directives: argument and engagement in academic writing’. K. 29. (1994). Journal of Pragmatics. S. Approaches to Pedagogic Grammar. Hyland. open. Hyland. ‘A second language classic reconsidered: the case of Schachter’s avoidance’. K. (1999). J. Jordan. 235–53. Hyland. 215–39. and Pavlenko.248 References Hyland. Juilland. (1984). ‘The empirical evidence for the influence of the L1 in interlanguage’. (1997). 289–306. ‘Preparation and analysis of linguistic corpora’. Natural Language Engineering.and closed-system characteristics’. pp. S. Metadiscourse. ‘Distribution of common words and phrases in text and language modelling’. Journal of English for Academic Purposes. Applied Linguistics. T. in Odlin. A Guide and Resource Book for Teachers. ‘Issues in creating a corpus for EAP pedagogy and research’. (2002b). C. A. S. (2002a). 23. R. A. ‘Is there an “academic vocabulary”?’ TESOL Quarterly. 41 (2). ‘Nouns in search of a context: a study of nouns with both ˇ. T. Second Language Research. A Companion to Digital Humanities. Some Aspects of the Vocabulary of Learned and Scientific English. (eds). 21. Oxford: Blackwell. Johansson. R. ‘From printout to handout: grammar and vocabulary teaching in the context of data-driven learning’. Amsterdam and Philadelphia: John Benjamins. Language Learning. Shimura. J. King. 15–59. R. in Davies A. S. Jarvis. Academic Writing Course. class compositions’. (2007). Cambridge: Cambridge University Press. 93–114. 385–95. T. Harlow: Pearson Education Limited.). ‘Qualifications and certainty in L1 and L2 students’ writing’. Kroll. P. C. Hyland. (1978). P. (2000) Disciplinary Discourses: Social Interactions in Academic Writing. P. 245–309. Jordan. E. Ide. Johns. (ed. (1997). Kellerman. New York and London: Routledge. 2 (1). (eds). Siemens. London and New York: Continuum. and Kellerman. 157–68. Ivanic R. E. R. and Kosem. A. R. Harlow: Pearson Education Limited. ‘What does time buy? ESL student performance on home vs. K. (2009). . Journal of Second Language Writing.

. Vocabulary and Applied Linguistics. Lehmann. Leech. N. 1–21. text types. ‘The use of tagging’. (ed. P. 34 (1). pp. (eds). H. Learner English on Computer. 20 (1). S. (2006). R. B. (1999b). (2004). G. Luzón Marco. S. Schneider. Lonon Blanton. English for Specific Purposes. L. U. Special Language: From Humans Thinking to Thinking Machines. . ‘Overstatement in advanced learners’ writing: stylistic aspects of adjective intensification’.). (1996). registers. A Corpus Study of Argumentative Writing. in Granger. Boston: Heinle and Heinle. 53–66. Rayson. B. (2000). London: Macmillan. pp. pp. Lennon. Leech. and Smith. E. (1989). H. 126–132. and Swales. (1997). Lakshmanan. P. 58 (2). G. 23–36. 393–420. pp. ‘BNCweb’. and Wilson.. Adjective Intensification – Learners versus Native Speakers. (1994). 1–18. Language Learning and Technology. 137–44. ‘Analysing interlanguage: how do we know what learners know?’ Second Language Research. Lee. ‘A corpus-based EAP course for NNS doctoral students: Moving from available specialized corpora to self-compiled corpora’. E. 55–75. in Bublitz. ‘What percentage of text-lexis is essential for comprehension?’. G.. H. Amsterdam: Rodopi. Amsterdam and Atlanta: Rodopi. Paris: Presses Universitaires de France.References 249 Lake. ‘Collocational frameworks in medical research papers: a genre-based study’. 23–36. ‘Introducing corpus annotation’. Laufer. Language and Computers: Studies in Practical Linguistics 27. Corpus Annotation: Linguistics Information from Computer Text Corpora. 25. in Flowerdew. ‘How much lexis is necessary for reading comprehension?’. (2001). Word Frequencies in Written and Spoken English. non-native argumentative writing’. K. P. (2000). 316–23. (1999). in Granger. Mieux écrire en anglais. Laufer. and Ventola. Syntactic Wordclass Tagging. Lorenz. Lorenz. A.-M. (1999). Lenk. W. L.). and styles: clarifying the concepts and navigating a path through the BNC jungle’. (eds). and Selinker. and Pemberton. pp. Amsterdam and Philadelphia: John Benjamins. Lorenz. and Tong. A. G. G. Lee. J. R. Leech. pp. ‘Procedural vocabulary: lexical signalling of conceptual relations in discourse’. (1992). (1998). (1999a). 183–96. Composition Practice 3. 17. in Garside. (ed.). Laruelle. U. and Hoffmann. (eds). Leech G. (ed. Dordrecht: Kluwer Academic Publishers. in van Halteren. P. (1998) ‘Preface’. London and New York: Addison Wesley Longman. in Arnaud. D. 19 (1). London and New York: Addison Wesley Longman. Leech. (2001). S. ‘Using ‘on the contrary’: the conceptual problems for EAP students’. ‘Genres. 63–86. C.-L. J. (eds). Applied Linguistics. M. K. English for Specific Purposes. J. Hong Kong: The Hong Kong University of Science & Technology. (2004). P. domains.). 5 (3). in Kirk. 259–66. London: Longman. Corpora Galore: Analysis and Techniques in Describing English. pp. G. L. Entering Text. D. IRAL. pp. ELT Journal. Coherence in Spoken and Written Discourse. Learner English on Computer. and Nordman. Luzón Marco. 37–72. New York: Addison Wesley Longman. (2001). and Béjoint. S. M. M. 56–75. Li. G. J. (2001). How to Create it and How to Describe it. (eds). J. and McEnery.. ‘Learning to cohere: causal links in native vs. A. in Lauren. (ed. ‘An investigation of students’ knowledge of academic and subtechnical vocabulary’. J. Clevedon: Multilingual Matters. ‘Getting ‘easy’ verbs wrong at the advanced level’.

P. R. English for Specific Purposes. . ‘A computer corpus linguistics approach to interlanguage grammar: noun phrase complexity in advanced learner writing’. Fixed Expressions and Idioms in English. in Candlin. (1991). Miller. I. ‘Exploiting L1 and interlanguage corpora in the design of an electronic language learning and production environment’. 10 (1).. 16 (2). (2006). P. R. Martínez. Harlow: Longman. ‘Collocations and lexical functions’. Unpublished PhD thesis. F. ‘Academic vocabulary in agricultural research articles: a corpus-based study’. 91–7. Nation. English Corpus Studies. TESOL Quarterly. (1995). Phraseology: Theory. Analysis and Applications. 28. McCarthy. Available from http://www. Martin. K. Discourse Markers in Native and Non-native English Discourse. N. in Kettemann. J.). pp. G. Mel’c ˇuk. B. Tübingen: Gunter Narr Verlag Tübingen. Oxford: Clarendon Press. K. M. (eds). 23–53. 221–43. J. Coming to Know: Studies in the Lexical Semantics and Pragmatics of Academic English. text coverage and word lists’. 4 (1). Nation. P. and Hwang. and Rohrback. P. Milton. McCarthy. and Hyland. ‘Where would general service vocabulary stop and special purposes vocabulary begin?’ System. C. (2006). ‘Rethinking applied corpus linguistics from a language-pedagogical perspective: new departures in learner corpus research’. in Cowie. M. Discourse Analysis for Language Teachers. Meyer. (eds). Cambridge: Cambridge University Press. S. Mukherjee. and Marko. McEnery. A. Oxford: Oxford University Press. in Schmitt. (2005). London and New-York: Routledge. 25–32. 23–42.) (2006). Planning. P. ‘Teaching academic vocabulary to foreign graduate students’. Processes and Müller. Gluing and Painting Corpora: Inside the Applied Corpus Linguist’s workshop. London and New York: Addison Wesley Longman. (eds). J. Université catholique de Louvain: Louvain-la-Neuve. M. and Waring. S. (2006). (2001). and O’Dell. Milton. Amsterdam and Philadelphia: John Benjamins. Mudraya. J. (1976). Mukherjee. O. Journal of Pragmatics. A. M.uni-giessen. Learning Vocabulary in another Language. (ed. Major. N. and Tono. 6–19. R. A. 186-198. Frankfurt am Main: Peter Lang. (ed. (1973). Narita. (2005). Anglistik. (2000). RELC Journal. pp. Vocabulary: Description. Cambridge: Cambridge University Press. J. I. and Sugiura. R. Learner English on Computer. London and New York: Longman. pp. Cambridge: Cambridge University Press. (1997). R. C. and Panza. Writing: Texts. 13.. ‘Lexical thickets and electronic gateways: making text accessible by novice writers’. (ed. (1998). (2009). and Nation. (1998). Meunier. P. ‘Preparing word lists: a suggested method’. (2006). ‘The native speaker is alive and kicking – linguistic and languagepedagogical perspectives’. Beck.). (1995). pp. M. (1997). ‘The function of like in dialogue’. The Longman Exams Dictionary. 25 (2). 205–32. in Granger. ‘Engineering English: a lexical frequency instructional model’. 23 (1). Y. Cambridge: Cambridge University Press. Moon. F. and Weinert. ‘The use of adverbial connectors in argumentative essays by Japanese EFL college students’. (2008). 183–98. W. (1999). 23. Nation. Xiao. ‘Vocabulary size.250 References Lynn. Corpus-based Language Studies: An Advanced Resource Book. S. English for Specific Purposes. J-M. Acquisition and Pedagogy. 235–56. (1998). Academic Vocabulary in Use. 365–93. 7–23.G. 35–41.

and Mason. R. (eds). pp. ‘Student papers across the curriculum: designing and developing a corpus of British student writing’. A. T. (1998). J-P.. E... M. Phraseology in Language Learning and Teaching. Neff. (2004b).) (2004)... in Tschichold. Díez. ‘A corpus-based study of business English and business English teaching materials’. Phraseological Units: Basic Concepts and their Application. (2008). in Doughty. English Modality in Perspective. R. in Moder. F. ‘The expression of writer stance in native and non-native argumentative texts’. J. People. F. (eds). Cambridge: Cambridge University Press. (2004).. Boston and New York: Houghton Mifflin Company. J. Oxford: Blackwell. and Rica.. Ballesteros. and Granger.. F. Neff J. 269–86. F. P. ‘Linguistic correlates of second language literacy development: evidence from middle-grade learner essays’. D. pp. Amsterdam: Rodopi. F. Frankfurt am Main: Peter Lang. pp. pp. (2007). (2000). Basel: Schwabe. in Fitzpatrick. Academic Word Power (1 – 4). J. J. ‘Formulaic language in English academic writing: a corpus-based study of the formal and functional variation of a lexical phrase in different academic disciplines’. Corpus Linguistics and Society. Computers and Composition. 85–99. (2004a). ‘Cross-linguistic influence’. in Facchinetti. D. Language Transfer: Cross-linguistic Influence in Language Learning.. Obenda.. The Handbook of Second Language Acquisition. Neff van Aertselaer. and Martinovic-Zic. D. A. E. M. Ballesteros. M. ‘Contrasting English-Spanish interpersonal discourse phrases: a corpus study’. in Gerbig. J.14 (1). (2008).) English Core Linguistics. G. M.References 251 Neff. Díez.. (2003). (eds).). (eds). F. H. Allerton. C. M. in Meunier. O. Edinburgh: Edinburgh University Press. Essays in Honour of D. pp. Ballesteros. Literary and Linguistic Computing. (2002). ‘A contrastive functional analysis of errors in Spanish EFL university writers’ argumentative texts: corpus-based study’. M. (2007). Discourse across Languages and Cultures. and Chaudron. ‘Transfer at the locutional level: an investigation of Germanspeaking and French-speaking learners of English’. Nesselhauf. Tschichold. Sharpling. (ed. Nesi. (1989). (2005). Amsterdam and Philadelphia: John Benjamins. pp. N (2004). 267–83. Amsterdam: John Benjamins. 436–86. ‘Contrastive discourse analysis: argumentative text in English and Spanish’. . Odlin. N. Odlin. Statistics for Corpus Linguistics. D. and Prieto. 203–25. J-P. Collocations in a Learner Corpus. C. 22 (1). C. (ed. 21. E. Bern: Lang. 85–100. Nesselhauf. C. ‘Use of the Chi-Squared Test to examine vocabulary differences in English language corpora representing seven different countries’. Nesselhauf. E. 141–61. H. 1–22. C. N. Journal of Second Language Writing. P. Dafouz. Martínez. Oakey.. J. ‘What are collocations?’. Unpublished PhD Thesis. E. Oakes. T. Language. L. (ed. and Wieser. Corpus Linguistics beyond the Word: Corpus Research from Phrase to Discourse (Language and Computers 23). F. Rica. Martínez.. Amsterdam and New York: Rodopi. Dafouz. and Long. pp.. Oakes. (2003). 439–50... in Reynolds. Prieto. and Palmer. (eds). R. Dafouz. J. pp. Dafouz. and Farrow. and Ganobcsik-Williams. Numbers. in Allerton. S. and Rica. Neff J. 19–45. M. Nelson. (2005). F. Manchester: University of Manchester. Amsterdam: John Benjamins. 73–89. Martínez. J. (eds). ‘Formulating writer stance: a contrastive study of EFL learner corpora’.

P. S. R. Language and Communication. H. 519–49. Lancaster University. C. Patterns and Meanings. and Gozdz-Roszkowski. Leech. S. pp. 243–65. is a native speaker?’ Anglistik. London and New York: Addison Wesley Longman. ‘Matrix: a statistical method and software tool for linguistic analysis through corpus comparison’. in Walinski. New York: Pearson Education. S.html Paquot. (1985). 13 (4). and Schmidt. Phraseology in Language Learning and Teaching. Université catholique de Louvain. PALC 2005. Phraseology in Language Learning and Teaching. London: Longman. (1998). ‘Who.comp. Amsterdam and Philadelphia: John Benjamins. M. (2008). M. (ed. 43–64. J. London: Longman. Writing Academic English. (2006). Lodz Studies in Language 13. and Syder.lancs. (eds). Using Corpora for English Language Research and Teaching. A. Schreier. Bergen. J. if anyone. Learner English on Computer. Rayson. J. F. Greenbaum. Available from http://cecl. A Comprehensive Grammar of the English Language. Amsterdam and Philadelphia: John Benjamins. . in Meunier. P. Partington. H. 127–40. M. M. Frankfurt am Main: Peter Lang.. 67–84. I. in Richards. Paquot. (eds). Corpus-based and Computational Approaches to Discourse Anaphora. 12 (2). Amsterdam and Philadelphia: John Benjamins. A. Kredens. in and Svartvik. Amsterdam: Rodopi. Available from http://www. 107—118. ‘Two puzzles for linguistic theory: nativelike selection and nativelike fluency’. Sophia Antipolis. Université de Nice. A..html Rayson. in Granger. Petch-Tyson. 8–22. Oshima. 101–19. Paper presented at the ASKeladden Opening Conference. Quirk. and Bestgen. pp.fltr. (2008). Petch-Tyson. ‘Exemplification in learner writing: a cross-linguistic perspective’. M. (1999). Norway. J. Amsterdam: John Benjamins. (in preparation). and Jucker. (eds). Paquot. pp. and Meunier. (2009). S. ‘Unveiling L1-induced effects with the help of learner corpora: Transfer of lexical priming’. ‘Lifting the “methodological fog” that covers transfer studies: a combination of Granger’s (1996) integrated contrastive model and Jarvis’s (2000) unified framework for transfer research’. M. A. (2001).. Unpublished PhD thesis. K. be/publications. S. pp. (2007a). pp. (2007b).. J. American University Word List. Paquot. ‘EAP vocabulary in native and learner writing: from extraction to analysis. Perdue. A. in Botley. ‘Reader/writer visibility in EFL persuasive writing’.ucl. P. 92. ‘Towards a productively-oriented academic word list’. Paquot. Corpora and ICT in Language Studies. Unpublished PhD Thesis.252 References Osborne. (2004). W. Piller. ‘Comment rendre compte de la “logique” de l’acquisition d’une langue étrangère par l’adulte?’ Etudes de Linguistique Appliquée.). (eds). ‘Distinctive words in academic writing: a comparison of three statistical tests for keyword extraction’. and Hogue. (eds). (eds). M. Corpora: Pragmatics and Discourse. S. 29–59. M. ‘From key words to key semantic domains’. 109–21. G. Pecman. Pawley. D. Praninskas. ‘Phraséologie contrastive anglais-français: analyse et traitement en vue de l’aide à la rédaction scientifique’. (1972). Paquot. ‘Demonstrative expressions in argumentative discourse – a computer-based comparison of non-native and native English’. S. International Journal of Corpus Linguistics. Y. (2008b). (1993). F. M. (1998). S. 24–25 June 2008. (2003). ‘Phraseology effects as a trigger for errors in L2 English: the case of more advanced learners’. in Granger. A phraseologically-oriented approach’. Unpublished PhD thesis. (1983). and McEnery. and Granger. (2008a). F. London and New York: Longman. A. pp.

Fairon. B. D. (eds). 199–219. Ruetten. How to Use Corpora in Language Teaching. N. H. Le Poids des Mots: Proceedings of the 7th International Conference on Statistical Analysis of Textual Data (JADT 2004). (2005).. Oxford: Macmillan Education. and Schmitt. Cross-linguistic Similarity in Foreign Language Learning. K. (ed. Schmitt. London: Longman. Saville-Troike. J. and Dister. McH. ‘Developing and exploring the behaviour of two new versions of the Vocabulary Levels Test’. Developing Composition Skills. Ringbom. (2005). and Sinclair.. Macmillan English Dictionary for Advanced Learners. Patterns. in Lüdeling.. (2005). P. Rundell. D. U.) (2007). J. and Stewart. pp. and Marinova. W. M. (1996). and Kytö. ‘Dictionary use in production’. A. M. (eds). (2008). 14. A.. C. Berlin: Mouton de Gruyter. 18 (1). Boston: Heinle. Bernardini. London and New York: Longman. 128–43. London and New York: Addison Wesley Longman.Corpora and Language Learners. ‘Comparing real and ideal language learner input: The use of an EFL textbook corpus in corpus linguistics and language teaching’. Amsterdam: John Benjamins. in Granger. S. N. Rhetoric and Grammar.. (2007). pp. (1999). pp.). M. Language Testing. The Role of the First Language in Foreign Language Learning. in Tyler. and Clapham. pp. 12 (1). and Francis. (2004). R. M. 2004. . Scarcella. G. Berridge. (2nd Edition). 112–30. in Aston. Schleppegrell. (eds). (ed. ‘Cognates. Amsterdam: John Benjamins. (2005). (1998). H-J. ‘Vocabulary frequencies in advanced learner English: a crosslinguistic approach’. S. Clevedon: Multilingual Matters. C. Focus on Vocabulary: Mastering the Academic Word List. Ringbom. Language in Use: Cognitive and Discourse Perspectives on Language and Language Learning. English Corpus Linguistics: Studies in Honour of Jan Svartvik. in Aijmer. Schmitt. ‘Conjunction in spoken English and ESL writing’.References 253 Rayson. G. ‘Collocational frameworks in English’. Applied Linguistics. B. U. Ringbom. 151–68. (1984). U. Washington: Georgetown University Press. Louvain-la-Neuve: Presses universitaires de Louvain. pp. (1987). Römer. Contexts and Didactics. Pedagogy. ‘Corpora and language teaching’. Amsterdam: John Benjamins. Römer. A Corpus-driven Approach to English Progressive Forms. Rundell. (2004b). Corpus Linguistics. English Abstract Nouns as Conceptual Shells: From Corpus to Cognition. 35–53. C. (1991). D. 123–36. Clevedon and Philadelphia: Multilingual Matters. M. C. Berlin: Mouton de Gruyter. H. A. Journal of Second Language Writing. U. M. (eds). Belgium. D. (2001). Schmitt. and Altenberg. International Journal of Lexicography. An International Handbook (volume 1). A. Y. in Purnelle. 19–45. 926–36. 185–99. (eds). D. March 10–12. Reynolds.). pp. 55–88. Römer. B. (2004a). pp. ‘What really matters in second language learning for academic achievement?’ TESOL Quarterly 18 (2). 41–52. (ed. (2003). H.. (2000). 271–85. D. Renouf. Schmid. Progressives. J. ‘A corpus-driven approach to modal auxiliaries and their didactics’. M. in Sinclair. Functions. Learner English on Computer. cognition and writing: an investigation of the use of cognates by university second-language learners’. and Zimmerman. ‘Extending the Cochran rule for the comparison of word frequencies between corpora’. Takada. ‘Linguistic correlates of second language literacy development: evidence from middle-grade learner essays’. [HSK series]. Kim. Louvain-la-Neuve. Römer. 17 (3).

L. B. J. (2008). pp. (1986). (ed. Available from http://ahds. Siepmann. English for Specific Purposes. Shaw. Soler.html Siegel. Current Issues in Linguistic Theory 17. Sinclair.). (1991). J. M. Scott. (1999). J. M. Corpus and Discourse. Berlin and New York: Mouton de Gruyter. M. Amsterdam and New York: Rodopi. (2006). M. ‘PC analysis of keywords and key keywords’. P. (2006). Amsterdam and Philadelphia: John Benjamins. ‘Analysing adjectives in scientific discourse: an exploratory study with educational applications for Spanish speakers at advanced university level’. IRAL. Papers from the 23rd International Conference on English Language Research on Computerized Corpora. ‘Interlanguage’.. frequency distributions through the WordSmith Tools suite of computer programs’. (ed. A. M. ‘The lexical item’.ac. ‘Intuition and annotation – the discussion continues’. 1–16. (1997). Oxford: Oxbow Books. (ed. Scott. 1–24.). . Tübingen: Stauffenburg M. Amsterdam: John Benjamins. (ed. Developing your English Vocabulary.). (ed.). Trust the Text: Language. 98–115. and Altenberg. (ed. London: Collins. M. 2577–90. New York: Routledge. pp. L. D. J. A Systematic New Approach. in Weigand. M. 19 (1).. cancer experience and internet use: a comparative keyword analysis of interviews and online cancer support groups’. and Roseberry. M. 35–71. C. pp. 145–65. in Wynne. (2001). (2004a). Advances in Corpus Linguistics. M. (1992).dk/ engelsk/naes2004/papers. pp. ‘Like: the discourse particle and semantic’. M. 149–63. Ziebland. J. ‘The development of Swedish university students’ written English. M. Directions in Corpus Linguistics. (2002). Educational Linguistics. ‘Technical. Looking up. (2005). M.254 References Scott. May 27–29. (2004b). in Svartvik.hum. (2002). M. appropriacy. System. ‘The nature of the evidence’. P. Selinker. Available from http://www. Sinclair. J. E. Oxford: Oxford University Press. Denmark. ELT Journal. and Charteris-Black. Rediscovering Interlanguage. J. 62. in Proceedings of the Ninth Nordic Conference for English Studies. ‘Corpus and text–basic principles’. (eds). K. Sinclair. in Ghadessy. Strevens. Textual Patterns: Key Words and Corpus Analysis in Language Education. Oxford and New York: Blackwell. Social Science and Medicine. 209–31. J. lexical competence and nuclear vocabulary’. Sinclair. (2005). ‘The empty lexicon’. 150–9. Sinclair. 25 (2). and scientific English’.htm. WordSmith Tools 4. (2004). C. Aarhus. 233–45. J. Stubbs.). M. X (3). Selinker. London and New York: Routledge. Scott. Journal of Semantics. Seale. J. Oxford: Oxford University Press. Sinclair. Discourse Markers across Languages: A Contrastive Study of Second-level Discourse Markers in Native and Non-native Text with Implications for General and Pedagogic Lexicography. Contrastive Lexical Semantics. pp. Collocation. (1987). 223–34. Small Corpus Studies and ELT. in Sinclair. J. Sinclair. technological. ‘The automatic analysis of corpora’. ‘Language development. (1992). (2004).). scope and coherence’. Henry. Corpus. pp. in Stubbs. 27 (3). G. in Aijmer. London and New York: Longman. M. pp. (1973). 21. and Tribble. 47–67. ‘Gender. M. collocations. Stein. S. Developing Linguistic Corpora: A Guide to Good Practice. in Sinclair. Concordance. (1972). L. 39–59. 378–97. Amsterdam: John Benjamins. ‘Comparing corpora and identifying key words. (eds).



Summers, D. (1996), ‘Computer lexicography: the importance of representativeness in relation to frequency’, in Thomas, J. and Short, M. (eds), Using Corpora for Language Research: Studies in Honour of Geoffrey Leech. London: Longman, pp. 260–6. Sutarsyah, C., Nation, P. and Kennedy, G. (1984), ‘How useful is EAP vocabulary for ESP? A corpus based case study’. RELC Journal, 25, 34—50. Swales, J.M. (1990) Genre Analysis: English in Academic and Research Settings. Cambridge: Cambridge University Press. Swales, J. M. (2002), ‘Integrated and fragmented worlds: EAP materials and corpus linguistics’, in Flowerdew, J. (ed.), Academic Discourse. Harlow: Pearson Education, pp. 150–64. Swales, J.M., Ahmad, U., Chang, Y., Chavez, D., Dressen, D. and Seymour, R. (1998), ‘Consider this: the role of imperatives in scholarly writing’. Applied Linguistics, 19, (1), 97–121. Swales, J.M. and Feak, C. B. (2004), Academic Writing for Graduate Students: Essential Tasks and Skills (2nd edition). Ann Arbor, MI: University of Michigan Press. Tan, M. (2005), ‘Authentic language or language errors? Lessons from a learner corpus’. ELT Journal, 59 (2), 126–34. Tankó, G. (2004), ‘The use of adverbial connectors in Hungarian university students’ argumentative essays’, in Sinclair, J. (ed.), How to Use Corpora in Language Teaching. Amsterdam and Philadelphia: John Benjamins, pp. 157–81. Thurstun, J. and Candlin, C. (1997), Exploring Academic English. A Workbook for Student Essay Writing. Sydney: NCELTR Publications. Thurstun, J. and Candlin, C. (1998), ‘Concordancing and the teaching of the vocabulary of Academic English’. English for Specific Purposes, 17 (3), 267–80. Tognini-Bonelli, E. (2001), Corpus Linguistics at Work. Amsterdam and Philadephia: John Benjamins. Tognini-Bonelli, E. (2002), ‘Functionally complete units of meaning across English and Italian: towards a corpus-driven approach’, in Altenberg, B. and Granger, S. (eds) Lexis in Contrast: Corpus-based Approaches. Amsterdam and Philadephia: John Benjamins, pp. 73–95. Tribble, C. (2001), ‘Small corpora and teaching writing: towards a corpus-informed pedagogy of writing’, in Ghadessy, M., Henry, A. and Roseberry, R. (eds), Small Corpus Studies and ELT: Theory and Practice. Amsterdam and Philadelphia: John Benjamins, pp. 381–408. Trimble, L. (1985), English for Science and Technology. Cambridge: Cambridge University Press. Tseng, Y-C. and Liou, H-C. (2006), ‘The effects of online conjunction materials on college EFL students’ writing’. System, 34, 270–83. Tutin, A. (forthcoming), ‘Evaluative adjectives in academic writing in the humanities and social sciences’. Paper presented at Interpersonality in Written Academic Discourse: Perspectives across Languages and Cultures, Jaca, 11–13 December 2008. Available from interlae_2008_tutin.pdf Van Roey, J. (1990), French-English Contrastive Lexicology: An Introduction. Louvain-la-Neuve: Peeters. Vassileva, I. (1998), ‘Who am I/who are we in academic writing? A contrastive analysis of authorial presence in English, German, French, Russian and Bulgarian’. International Journal of Applied Linguistics, 8 (2), 163–90.



Voutilainen, A. (1999), ‘A short history of tagging’, in van Halteren, H. (ed.), Syntactic wordclass tagging. Dordrecht: Kluwer Academic Publishers, pp. 9–21. Wang, J., Liang, S., and Ge, G. (2008), ‘Establishment of a Medical Academic Word List’. English for Specific Purposes, 27 (4), 442–58. Wang, K. and Nation, P. (2004), ‘Word meaning in academic English: homography in the Academic Word List’. Applied Linguistics, 25 (3), 291–314. Ward, J. (1999), ‘How large a vocabulary do EAP engineering students need?’ Reading in a Foreign Language, 12 (2), 309–24. Ward, J. (2009), ‘A basic engineering English word list for less proficient foundation engineering undergraduates’. English for Specific Purposes, 28, 170–82. Weissberg, R. and Buker, S. (1978), ‘Strategies for teaching the rhetoric of written English for Science and Technology’. TESOL Quarterly,12 (3), 321–9. West, M. (1937), ‘The present position in vocabulary selection for foreign language teaching’. The Modern Language Journal, 21 (6), 433–7. West, M. (1953), A General Service List of English Words. London: Longman. Widdowson, H. G. (1983), Learning Purpose and Language Use. Oxford: Oxford University Press. Widdowson, H. G. (1991), ‘The description and prescription of language’, in Alatis, J. E. (ed.), Linguistics and Language Pedagogy: The State of the Art. Washington, D.C.: Georgetown University Press, pp. 11–24. Widdowson, H. G. (2003), Defining Issues in English Language Teaching. Oxford: Oxford University Press. Wilkins, D. A. (1976), Notional Syllabuses. Oxford: Oxford University Press. Wilson, A. and Thomas, J. (1997), ‘Semantic annotation’, in Garside, R., Leech, G. and McEnery, A. (eds), Corpus Annotation: Linguistics Information from Computer Text Corpora. New York: Addison Wesley Longman, pp. 53–65. Winter, E. (1977), ‘A clause relational approach to English texts: a study of some predictive lexical items in written discourse’. Instructional Science, 6, 1–92. Wray, A. (2002), Formulaic Language and the Lexicon. Cambridge: Cambridge University Press. Xue, G. and Nation, P. (1984), ‘A University Word List’. Language Learning and Communication, 3 (2), 215–29. Yang, H. (1986), ‘A new technique for identifying scientific / technical terms and describing science texts’. Literary and Linguistic Computing, 1 (2), 93–103. Zamel, V. (1983), ‘Teaching those missing links in writing’. ELT Journal, 37 (1), 22–9. Zemach, D. and Rumisek, L. (2005), Academic Writing: From Paragraph to Essay. Oxford: Macmillan. Zhang, M. (2000), ‘Cohesive features in the expository writing of undergraduates in two Chinese universities’. RELC Journal, 31, 61–95. Zhang, H., Huang, C. and Yu, S. (2004), ‘Distributional consistency: A general method for defining a core lexicon’, in Proceedings of the Fourth International Conference on Language Resources and Evaluation, Lisbon, Portugal, 26–28 May 2004. Available from Zwier, L. J. (2002), Building Academic Vocabulary. Michigan: The University of Michigan Press.


All cited internet sources were correct as of 2 August 2009.

Author index

Note: Page numbers in italics denote illustrations. Aarts, J. 35, 143 Ädel, A. 69, 72, 150, 157, 161 Aijmer, K. 152, 157, 176 Altenberg, B. 83, 106, 121, 126, 150, 152, 180 Archer, D. 42, 45 Aston, G. 73 Baayen, R. H. 145 Bahns, J. 204 Bailey, S. 206 Baker, M. 17, 19, 20, 21, 22, 24 Baker, P. 48 Barkema, H. 84, 100 Barkhuizen, G. 67 Bartning, I. 79 Bauer, L. 12 Bazerman, C. 72 Beheydt, L. 14, 27 Bestgen, Y. 62 Bhatia, V. 26 Biber, D. 2, 29, 55, 83, 122, 137, 143, 179, 211, 237n. 3 Billuroglu, A. 16 ˘ Biskup, D. 185 Bley-Vroman, R. 70 Bourigault, D. 237n. 7 (Ch. 4) Bowker, L. 35, 206 Brill, E. 37 Buker, S. 81 Burger, H. 83, 121 Burnard, L. 73 Campion, M. E. 11 Candlin, C. 206 Carter, R. 11, 23, 85, 207, 238n. 3 Celce-Murcia, M. 169 Charles, M. 214 Chen, C. W. 126, 152, 174 Chung, T. 14, 18 Clear, J. 102 Cohen, A. D. 18 Coltier, D. 88 Connor, U. 152, 190 Conrad, S. 59, 83, 85, 121, 179, 180 Cook, G. 207 Corson, D. 13 Cortes, V. 1 Cowan, J. R. 17, 18 Cowie, A. P. 213 Coxhead, A. 3, 5, 9, 10, 11, 12, 13, 16, 17, 20, 21, 25, 27, 28, 31, 34, 44, 63, 82, 122, 212 Crewe, W. 169, 174, 175, 176, 193, 201 Curado Fuentes, A. 46, 213 Cutting, J. 72 Davies, A. 71 De Bot, K. 239n. 3 Dechert, H. 155, 168 De Cock, S. 30, 72, 86, 121, 157 DeRose, S. 37 Dudley-Evans, T. 211 Eldridge, J. 26, 214 Elley, W. B. 11 Ellis, R. 67 Engels, L. K. 11 Evans, S. 1 Evert, S. 75, 76, 78 Farrell, P. 17, 18, 20 Farrow, M. 48, 50, 62 Feak, C. B. 24


Author index
Huntley, H. 9, 11, 16, 82 Hwang, K. 11, 13, 14 Hyland, K. 3, 24, 25, 26, 31, 32, 72, 90, 92, 93, 99, 147, 157, 189, 201, 211, 214 Ide, N. 37 Ivanic, R. 235n. 2 (Ch. 1) ˇ Jarvis, S. 4, 182, 183, 184, 185, 197, 216, 238n. 13 Johansson, S. 31 Johns, T. 214 Jordan, M. P. 23, 235n. 2 (Ch. 1) Jordan, R. R. 1, 81, 82, 85, 201, 202, 203, 211 Juilland, A. 50 Kamimoto, T. 239n. 2 Katz, S. 48, 236n. 8 Kellerman, E. 197 King, P. 206 Kosem, I. 62 Krishnamurthy, R. 62 Kroll, B. 69 Lake, J. 169, 170, 201 Lakshmanan, U. 70, 71 Larsen-Freeman, D. 169 Laruelle, P. 91 Laufer, B. 10 Lee, D. 73, 74, 132 Leech G. 11, 34, 35, 70, 72 Lennon, P. 165, 168 Li, E. S.-L. 18 Liou, H.-C. 206 Lonon Blanton, L. 85 Lorenz, G. 72, 101, 143, 146, 150, 152, 157, 169, 173, 177, 193 Luzón Marco, M. J. 22, 83, 137 Lynn, R. W. 11 Major, M. 11 Martin, A. 19, 20, 21, 27 Martínez, I. 15, 27, 34, 82, 212 McCarthy, M. 9, 23, 211, 235n. 2 (Ch.1), 238n. 3,

Field, Y. 171, 177, 193 Firth, A. 70 Fisher, D. 204 Flowerdew, J. 2, 60, 61, 172, 178, 201, 214 Flowerdew, L. 23, 199, 201, 204 Francis, G. 22, 23, 59, 235n. 2 (Ch. 1) Garside, R. 37, 38, 39 Ghadessy, M. 11 Gilquin, G. 1, 7, 70, 71, 151, 153, 195, 197, 207, 207, 208, 209, 210, 225, 238n. 5 Gläser, R. 237n. 7 (Ch. 4) Gledhill, C. 83, 102, 119, 123, 161, 236n. 4, 238n. 9 Goodman, A. 17 Granger, S. 4, 26, 32, 65, 67, 68, 70, 71, 72, 84, 100, 102, 118, 122, 123, 126, 143, 145, 150, 151, 152, 155, 157, 168, 169, 170, 177, 179, 182, 184, 185, 194, 197, 202, 204, 206, 213, 214, 215, 216, 236n. 1, 236n. 2, 238n. 9 Green, C. 1 Gregg, K. R. 218 Gries, S. 48, 50 Groom, N. 215 Halliday, M. 203 Hamp-Lyons, L. 206 Hancioglu, N. 15, 16, 27, 63, 212 ˘ Harris, S. 22 Harris Leonhard, B. 85 Hasan, R. 203 Hasselgren, A. 147 Heasley, B. 206 Heatley, A. 44 Hegelheimer, V. 203 Hinkel, E. 1, 3, 33, 59, 148 Hirsh, D. 10, 34 Hoey, M. 23, 26, 192, 216, 217 Hoffmann, S. 75, 76, 86 Hogue, A. 85 Howarth, P. 119, 165, 217 Huckin, T. 26, 214 Hunston, S. 118

Author index
McEnery, A. 30, 35, 76 Mel’cuk, I. 83 ˇ Meunier, F. 143, 150 Meyer, P. G. 24, 27 Miller, J. 237n3 Milton, J. 72, 147, 157, 179, 201, 202, 203, 206, 213 Moon, R. 121 Mudraya, O. 13, 14, 16, 17, 19, 31, 34 Mukherjee, J. 70, 71, 160 Müller, S. 237, Narita, M. 126, 152, 174, 177, 179, 202 Nation, I. S. P. 12 Nation, P. 1, 3, 10, 11, 13, 14, 16, 17, 18, 23, 26, 44, 82, 185 Neff, J. 73, 75, 152, 157, 194, 195 Neff van Aertselaer, J. 73 Nelson, M. 46 Nesi, H. 31, 32, 33 Nesselhauf, N. 73, 78, 101, 164, 166, 185 Neufeld, S. D. 16 Oakes, M. P. 48, 50, 62, 76 Oakey, D. 213 Obenda, D. 82 O’Dell, F. 9 Odlin, T. 185, 204 Osborne, J. 238n. 12 Oshima, A. 85 Paquot, M. 15, 26, 36, 62, 84, 100, 118, 122, 123, 135, 151, 153, 157, 168, 190, 195, 197, 204, 213, 214, 238nn. 9,13, 239n. 6 Partington, A. 15 Pavlenko, A. 185 Pawley, A. 71 Payne, E. 17 Pearson, J. 35 Pecman, M. 119 Pemberton, R. 18 Perdue, C. 192 Petch-Tyson, S. 142, 145, 157 Piller, I. 71 Praninskas, J. 11 Quirk, R. 179


Rayson, P. 29, 30, 37, 38, 43, 47, 50, 61, 76, 145, 150 Renouf, A. 102 Reynolds, D. W. 1 Ringbom, H. 182, 185, 192 Rodriguez, E. C. 50 Rohrback, J.-M. 160 Römer, U. 85 Ruetten, M. 85 Rumisek, L. 85 Rundell, M. 201, 207 St Johns, M. J. 211 Saville-Troike, M. 26 Scarcella, R. C. 13 Schleppegrell, M. J. 180 Schmid, H.-J. 235n. 2 (Ch.1) Schmitt, D. 11, 16 Schmitt, N. 11, 16 Scott, M. 2, 45, 46, 47, 48, 69, 236n. 9 Seale, C. 46 Selinker, L. 70, 71, 182, 183, 239n. 2 Shaw, P. 69 Siegel, M. 237n. 3 Siepmann, D. 82, 88, 101, 107, 126 Sinclair, J. M. 2, 26, 35, 82, 101, 102, 118 Smith, N. 35, 37, 39 Soler, V. 118 Stein, G. 16, 235n. 1 (Ch.1) Strevens, P. 13 Stubbs, M. 10 Sugiura, M. 126, 152, 174, 177, 179, 203 Summers, D. 207 Sutarsyah, C. 26 Swales, J. M. 24, 31, 61, 86, 92, 132, 189, 207 Swallow, H. 185 Syder, F. H. 71 Tan, M. 71 Tankó, G. 147 Tapper, M. 126, 150, 152 Thomas, J. 42

L. 35. 85 Zhang. 170. H. 20. 194 Zimmerman. S. 18. 177. 21 Tse. E. 60 Widdowson. K. 13. 206 Tognini-Bonelli. 80 West. G. 176. 11. 22 Wray. L. 62. V. A. L. 169. 27. 193 Zamel. 34 Waring. 126. H. 152. 201 Zemach. 17 Yip. 12. 185 Vassileva. 207. O. 206 Tutin. 86 Xue. J. 1. A. J. R. W. 10. E. 24. 85 Thompson. 26. 81 Wilson. 118 Tribble. 3. J. 16. 92 Tseng.-C. 61. 184 Van Roey. 50 Zhang. 11 . 152 Voutilainen. G. 30. R. D. 36. A. I. G. M. 25. 13 Zwier. 118 Tyson. 22. 16 Yang. 15. R. P. Y. H. 38 Wagner. M. 14. 212 Wilkins. 46. 11.260 Author index Weinert. P. 34 Wang. 26 Ward. 118 Thompson. 177. M. 2 Thurstun. 255n. 177. C. 42 Winter. D. 237n. 171. 70 Wang. A. J. J. B. J. C. 44. 69 Trimble. A. 3 Weissberg.

overused and underused in ICLE 144 academic literacy 231 academic vocabulary 7 vs. 31. 3–4. 7. 114–18 expressing a concession in 109 expressing possibility and certainty 118–20 reformulating in 109 see also British National Corpus (BNC) . core vocabulary and technical terms 10–13 definition of 212 fuzzy vocabulary categories 13–17 meaning of 9. 47 bilingual dictionaries 204 Billuroglu-Neufeld-List (BNL) 16 ˘ blend 168 BNC-AC-HUM 75. 102 comparing and contrasting in 112–14 expressing cause and effect in 110–11. 82 nouns and 56 overused and underused clusters with 156 and rhetorical functions 81–7 words distribution. 20. 123 automatic semantic analysis. 16. 217 academic discourse community 31 Academic Keyword List (AKL) 5. 15. 59 adverbials/adverbs 93. 15. 63. 122. 24. 214. 118 in the Academic Keyword List 57 as co-occurrents of academic nouns 100. of AKL 82–3 Baby BNC Academic corpus (B-BNC) 31. 122 academic vocabulary and 60 automatic semantic analysis of 82–3 distribution. 133. 27–8. 43 part-of-speech annotation 30. 41. 5. 59 semantic misuse and 139–40 sentence position 179 annotation 34–6. 43 semantic annotation 35. 34. 213 in the Academic Keyword List 58 mono-lexemic 91 multiword linking 121 potential academic 58. 37–42. 43–4. 82. 38. 167 potential academic 57. 59. 34–5. 90. 28 and sub-technical vocabulary 17–21 Academic Word List (AWL) 3. 63. 24–5. 40. 27. 25. 119. in GSL and AWL 60 words. 100. 53 association measures 76. 37. 36. in ICLE 143 exemplificatory discourse markers in 88 grammatical distribution categories in 55 need for concordancing in 61 need for pedagogic mediation 61. 101 attitudinal formulae 84. 102. 78. 32. Academic Corpus 11–12 academic discourse 2. 122. 11. 17. 212 activity verbs 59 adjectives 101.Subject index Note: Page numbers in italics denote illustrations. 55–61. 40. 60. 95. 12.

31. 164.262 Subject index Common European Framework of Reference for Languages (CEF) 236n. 106. 216. 146. 169. 202. 87. 4. 134. 190. 166. 120 overuse of 146 sentence initial position of 194 connectors 140. 77 booster 157 British Academic Written Corpus 217 British Academic Written English (BAWE) Pilot Corpus 32–3. 76. 84. 178. 30. 73. 77. 101. 100. 238n. 193 preferred co-occurrences in EFL writing 160–8 core vocabulary 3. 167. 30. 235n. 1–3 medial position for 180 overuse of 201–2 semantic misuse 201 sentence 22 sentence position 141. 203 Constituent Likelihood Automatic Wordtagging System (CLAWS) 37–42. 170. 133. 213. 174–82. 3 advance and retrospective labelling 22 grammatical 203 lexical 18. 208. 203. 215 contrastive rhetoric 2. 3. 76. 152 control corpus 67. 9 collocational framework 102 collocational overlap 165 . 193. 76. 8 Contrastive Interlanguage Analysis (CIA) 4. 215. 191 corpus-based approach 2. 236n. 119. 85. 238n. 210. 148 colligation 168 colligational errors 166. 119–20. 59 content words 10. 213 non-technical words 18 textual 123. 162. 203 BNCweb 75. 130. 59. 100. 87. 172. 93. 133–4. 217. 74. 1 (ch3) communicative phrasemes 84. 149 conceptual frequency 86 concession markers 87 in BNC-AC-HUM 109 conjunctions complex 40. 147 Centre for English Corpus Linguistics (CECL) 207 Clairefontaine Les fiches essentielles du Baccalauréat en anglais 160 CLAWS 37–42. 95. 22. 29. 125. 67. 123. 181. 203. 70. 161. 177. 172. 193–4. 152. 238nn. 236n. 84. 65. 211. 178–9. 34. 165. 61. 192. 78. 201–2. 118. 148. 121–2 comparative fallacy 70. 1. 22. 193. 150. 8 Cambridge Advanced Learner’s Dictionary 207 cataphoric markers 90–1 cause and effect markers 87. 95. 17. 102. 102. 73 co-occurrence 37. 114. 70. 115–17. 186. 76 co-text 2.2 (ch2) British National Corpus (BNC) 4. 78. 22. 102. 188. 202. 15 Corpus de Dissertations Françaises (CODIF) 184. 79. 71. 99. 31. 79. 6 corpus-driven approach 29. 219–25 in BNC-AC-HUM 110–11 EFL learners’ use of 146–8. 96. 180. 137. 119. 204. 160. 188–9 cohesion 22. 207 Baby BNC Academic Corpus (B-BNC) 31 BNC-AC 78 BNC-AC-HUM see BNC-AC-HUM Index 73–5 mark-ups 73 BROWN corpus 47 burstiness 48. 75–6. 59 code gloss 90. 132. 168 collocation 23–4. 174–6. 10–11. 77. 71 comparison and contrast markers 87. 35 Corpus Query Processor (CQP) 75. 238n. 226–34 in BNC-AC-HUM 112–14 EFL learners’ use of 148. 16.

78. 27. 73. 63. 86 . 61. 131. 75. 12. 26. 59. 91 endophoric marker 91. 98. 14. 93. 55. 47. 189–91 ‘extended units of meaning’ 118 fiction 46. 27. 119 English as a Second Language (ESL) 33. 211. 19 developmental factor 4. 60. 14. 98 IMS Open Corpus Workbench 75 International Corpus of Learner English (ICLE) 4. 55. 103. 123 textual formulae 121 free combination 100. 214 endophoric markers 91. 126. 93. 103. 188. 36 data-driven learning 214 derivation 12. 16. 27. 216 directives see imperatives discipline 3. 75. 189–90. 86. 71. 59 document-level burstiness 236n. 45. 143 fuzzy vocabulary categories 13–17 263 General Service List of English Words (GSL) 10–11. 84 attitudinal formulae 84. 18. 159 global keyword 48 local keyword 48 negative keyword 47. 215. 60. 46–8. 95. 62.Subject index data-driven approach 4. 62. 15. 143 ditto-tag 38. 93 discourse-organizing vocabulary 9. 98. 71. 107 as directives with rhetorical purpose 92 first person plural 136. 94. 55. 44. 60 general service word 16. 92. 204 second person 91–2. 86 positive keyword 47. 3 Juilland’s D statistical coefficient 50–3 keyness 4. 121 FLOB corpus 47 formulae 12. 132. 122. 72–3. 181. 95. 16. 13. 5. 135. 119. 61. 159. 26. 84. 23 dispersion see distribution distribution 29. 191. 47. 213. 99. 74. 65. 212 keyword 30. 192. 212 homographs 25. 137. 44. 33. 236n. 20. 102. 37. 125. 2 see also knowledge domain discourse marker 88. 30. 132. 86. 138 cataphoric marker 90. 28. 78. 101. 84. 183. 197. 25. 93. 29. 119 engagement marker 92. 235n. 102. 130. 27. 204 English for Academic Purposes (EAP) 15. 37 idiom 3. 67–9. 83 function words 10. 125. 40. 145 global keywords 48 grammatical cohesion see cohesion graphemic words 40 hedge 2. 157 high-frequency word 5. 123. 46–8. 148. 15. 180. 45. 20. 94. 50–3. 237n. 47 field approach 82 fixed phrase 23–4. 212 genre 2. 214. 10. 93. 99. 44. 45. 85 English for Specific Purposes (ESP) 9. 213. 123 FROWN corpus 47 functional-product approach 82 functional syllabus 81–2. 7 illocutionary nouns 23 imperatives in academic writing 93. 23. 62. 217 epistemic modifiers 147 evenness of distribution see distribution and Juilland’s D statistical coefficient exemplifiers 85–8 in BNC-AC-HUM 88–108 learners’ use of 125–42. 55. 8 EAP material design 221 EAP teaching 26. 45. 18.

9. 49. 184 . 201. 71 native student writing 72 negative keywords 86 n-gram 69 non-technical term 17. 126. 130. 99. 201. 118. 138 in the Academic Keyword List 56 adjectives as co-occurrents of academic 100. 239n2 L1-induced factor see transfer L1 influence 182–92 Jarvis’s unified framework 182–4 labeling 22–4. 195 Macmillan English Dictionary for Advanced Learners 7. 193. 32 monolingual learners’ dictionary (MLD) 206–7 morphosyntactic annotation 34–5 multiword expression 37. 59. 6 meaning 10. 186. 160. 123. 4. 88. 93. 177 lexical cohesion see cohesion lexical extension 213 lexical priming see priming lexical repertoire 3. 140. 125. 129. 161 metalinguistic labels 23. 197. 237n. 76. 204 LOB corpus 47 local keywords 48 logico-semantic relationship verbs 59 log-likelihood 47. 62. 158. 179. 206. 218. 118. 156. 185 delexical meaning 118 figurative meaning 119 over-extension 146. 167 verbs as co-occurrents of academic 95–9. 133. 159. 60. 65. 121. 173. 108. 197 native control corpus 70 native speaker norm. 25. 155. 53. 145–6. corpusapproximation to 70. 190. 125 log-likelihood calculator. 4. 18–19. 35. 152. 90. 197. 194. 144. 20 nuclear words and pragmatic neutrality 14 organizational function see rhetorical function overuse 86. 18 non-technical words 18–19. 118. 151. 118. 20. 137. 5. 10. 97 retrospective labeling 22. 95–6. 84. 143. 14. 184. 195. 238n. 197. 134. 24 nouns 22–3. 73. 83. 160. 31. 215 linking word 3. 170. 194. 78. 19. 44. 48. 164. 215 lexico-grammatical error 155. 190. 137. 96. 59 Michigan Corpus of Upper-level Student Papers 217 Micro-Concord Corpus Collection B (MC) 31. 14. 192–3 lexical teddy bear 147 lexical transfer see transfer lexico-grammar 26. 180. 148. 162–3 novice writing 1. 105.264 keyword analysis see keyness knowledge domain 31. 181. 13. 217 nuclear vocabulary 9. 32 Subject index non-technical meaning 18. 177. 52 mental process nouns 23 mental verbs 59 metadiscourse 24. 121. UCREL 125 LONGDALE project 239n. 161. 194. 100–1. 157. 138. 195–6. 185. 85. 1 Longman Dictionary of Contemporary English 206 Louvain Corpus of Native Speaker Essays (LOCNESS) 32. 214. 102. 203 advance labeling 96. 85. 98 labels 22–3 semantic misuse 172–3 language-activity nouns 23 learners’ dictionary 206–7 lexical bundle 69. 215. 152. 150. 52. 207. 19 technical meaning 18. 142–50. 71. 5 Oxford English Dictionary (OED) 20 paraphrasing and clarifying see reformulation markers L1 frequency 185.

82. 136. 137. 205 semantic annotation 35 semantic misuse 5. 73 reference corpus 46. 138. 24 syntactic annotation 35 tagging see annotation teaching material 82. 138. 217 mental 192 transfer of 192. 90. 1 technical terms 3. 215 reporting verbs 59 retrospective labelling 22 rhemes 97. 26. 212 Perl program 48 personal metadiscourse 161 personal pronoun 98. 24. 123. 90. 68. 119 structural phraseme 121 textual phraseme 84. 99. 108. 188 phraseological competence 217 phraseological infelicity 155 phraseology of rhetorical functions 65. 201 semantic tagging 43 semantic transfer see transfer semi-technical vocabulary 17 sentence connectors 22 sentence stem 97. 118. 161. 208. 193. 195. 121. 168–74. 135 specialised non-technical lexis 17. 21. 192. 101 demonstrative 96 as exemplified item 97 impersonal 168 personal 98. 22. 123. 44–55 preferred co-occurrence 2. 102. 120. 48–50. 213 phraseological ‘cascade’ 161. 110–11. 15. 138. 85. 164. 131. 122. 153. 98. 119. 209 in BNC-AC-HUM 109 register awareness 5. 9. 151. 142. 93. 213. 71. 90. 154. 14. 141. 18 speech 2. 15–16. 213 speech-like lexical item 151–2. 144 priming 192. 123. 108. 161. 125. 11. 177. 217 frequency-based approach to 122 positive keywords 86 potential academic words 29. 172. 14. 132. 106. 84. 78. 203 procedural vocabulary 22. 106. 27. 37–8 see also annotation pedagogic mediation 61. 197. 184. 190. 161. 7. 70. 142. 81. 118. 121. 121. 214. 150. 17. 136. 215 rhetorical overstatement 176 Robert & Collins CD-Rom 204. 62. 121. 213. 94. 193 preferred ways of saying things 83. 17–21 . 81. 148. 122. 238n. 119–20. 30. 13. 160. 61. 151. 84. 195 spoken frequency counts 145 Student Writing Corpus 32. 33 sub-technical vocabulary 17–21. 45. 155. 139. 193. 13.Subject index parsing see syntactic annotation part-of-speech (POS) tagging 34–5. 166. 108. 207. 9. 190. 33. 95. 41. 18 technical vocabulary 13. 157 mono-lexemic phraseme 88. 150–2. 206. 212 pronoun 23. 112–14. 169. 135 referential phrasemes 84. 164 phraseme 83. 197. 193 preposition 97. 155. 22. 132. 145. 143 complex 40. 170. 94. 197. 118. 4. 178. 136. 139. 180. 62. 123. 60. 148. 20. 73. 161 phraseological accent 83 phraseological analysis 83–4. 142. 135 rhetorical function 5. 90. 10. 152. 164 third person 157 265 range 1. 4. 185. 120. 114. 63. 214 communicative phraseme 84. 109. 139–40. 211 production 1. 203. 62. 106. 153. 202. 47. 148. 166. 161 referential phraseme 84. 188. 101. 84. 9. 93. 95. 119 reformulation markers 87. 115–17. 160. 212 Range corpus analysis program 44 reception 1. 69. 160. 125. 76. 12. 211. 216. 121. 145. 203. 192–3.

218 lexical transfer 216 transfer effects 182–5 transfer of form 185 transfer of form/meaning mapping 185. 11. 158. 156. 181. 194. 42–4 . 146. 162–3 forming rhemes with noun 98 lexical 36 linking 59 mental 59 potential academic 57 reporting 59 in sentence-initial infinitive clauses 138 Vocabulary 3 items 22 Web Vocab Profile 59. 144. 190. 197. 102. 118–20. 131. 34.266 Subject index Varieties of English for Specific Purposes dAtabase (VESPA) 217 verbs 24. 69 Concord tool 69 Detailed Consistency Analysis 51 Keywords option 155 WordList option 49 ‘you-know-it-when-you-see-it’ syndrome 182 text coverage 10. 182. 40. 137. 60 within-document burstiness 236n. 16. 17. 53 word families 12. 26–7. 136. 4 text nouns 23 textual formulae 121 textual phrasemes 84. 168. 203. 126. 145. 155. 161 textual sentence stems 97. 16. 216 transfer of style and register 185. 143. 130. 148. 119. 191. 59. 45 in AWL 12. 16–17. 236n. 118. 147. 16. 91. 17 in GSL 11 word form 12. 135. 94. 51. 157. 216. 121. 48. 137. 188. 192 transfer of training 144. 201–3 transfer of use 185 transfer-related factor see transfer typicality 106. 36. 27. 171. 197. 134. 123. 160. 46 in the Academic Keyword List 57 Word Smith Tools 2. 17. 194. 190–1 transfer of meaning 185 transfer of the phraseological environment 185. 197 transfer of primings 192. 107 underuse 86. 216 transfer of L1 frequency 185. 159 underused words see negative keywords University Word List 16 USAS (UCREL Semantic Analysis System) 37. 157–8 activity 59 co-occurrents of academic nouns 95–9. 49. 203. 121. 93. 149. 157 word list 2. 47. 15. 39. 135 tokenisation 38 transfer 4. 8 Wmatrix 36–7.

Sign up to vote on this title
UsefulNot useful