Vocabulary Theory, Patterning and Teaching 9781788923750

This book presents the current state of knowledge in the vibrant and diverse field of vocabulary studies, summarising th

191 54 13MB

English Pages 296 [279] Year 2021

Report DMCA / Copyright

DOWNLOAD FILE

Polecaj historie

Vocabulary Theory, Patterning and Teaching
 9781788923750

Citation preview

Vocabulary Theory, Patterning and Teaching

SECOND LANGUAGE ACQUISITION Series Editors: Professor David Singleton, University of Pannonia, Hungary and Fellow Emeritus, Trinity College, Dublin, Ireland and Associate Professor Simone E. Pfenninger, University of Salzburg, Austria This series brings together titles dealing with a variety of aspects of language acquisition and processing in situations where a language or languages other than the native language is involved. Second language is thus interpreted in its broadest possible sense. The volumes included in the series all offer in their different ways, on the one hand, exposition and discussion of empirical findings and, on the other, some degree of theoretical reflection. In this latter connection, no particular theoretical stance is privileged in the series; nor is any relevant perspective – sociolinguistic, psycholinguistic, neurolinguistic, etc. – deemed out of place. The intended readership of the series includes final-year undergraduates working on second language acquisition projects, postgraduate students involved in second language acquisition research, and researchers, teachers and policymakers in general whose interests include a second language acquisition component. All books in this series are externally peer-reviewed. Full details of all the books in this series and of all our other publications can be found on http://www.multilingual-matters.com, or by writing to Multilingual Matters, St Nicholas House, 31–34 High Street, Bristol BS1 2AW, UK.

SECOND LANGUAGE ACQUISITION: 152

Vocabulary Theory, Patterning and Teaching Studies in Honour of Norbert Schmitt

Edited by

Paweł Szudarski and Samuel Barclay

MULTILINGUAL MATTERS Bristol • Jackson

DOI https://doi.org/10.21832/SZUDAR3743 Names: Szudarski, Paweł, editor. | Barclay, Samuel, editor. Title: Vocabulary Theory, Patterning and Teaching/Edited by Paweł Szudarski and Samuel Barclay. Description: Bristol; Blue Ridge Summit: Multilingual Matters, [2022] | Series: Second Language Acquisition: 152 | Includes bibliographical references and index. | Summary: “This book presents the current state of knowledge in the vibrant and diverse field of vocabulary studies, summarising the latest empirical studies and providing a firm indication of the future of the field. The chapters cover the key themes of theorizing and measuring vocabulary knowledge, formulaic language, and learning and teaching vocabulary”—Provided by publisher. Identifiers: LCCN 2021027813 (print) | LCCN 2021027814 (ebook) | ISBN 9781788923736 (paperback) | ISBN 9781788923743 (hardback) | ISBN 9781788923750 (pdf) | ISBN 9781788923767 (epub) Subjects: LCSH: Vocabulary—Study and teaching. | Second language acquisition. | LCGFT: Essays. Classification: LCC P53.9 .V67 2022 (print) | LCC P53.9 (ebook) | DDC 401/.93—dc23 LC record available at https://lccn.loc.gov/2021027813 LC ebook record available at https://lccn.loc.gov/2021027814 Library of Congress Cataloging in Publication Data A catalog record for this book is available from the Library of Congress. British Library Cataloguing in Publication Data A catalogue entry for this book is available from the British Library. ISBN-13: 978-1-78892-374-3 (hbk) ISBN-13: 978-1-78892-373-6 (pbk) Multilingual Matters UK: St Nicholas House, 31–34 High Street, Bristol BS1 2AW, UK. USA: Ingram, Jackson, TN, USA. Website: www.multilingual-matters.com Twitter: Multi_Ling_Mat Facebook: https://www.facebook.com/multilingualmatters Blog: www.channelviewpublications.wordpress.com Copyright © 2022 Paweł Szudarski, Samuel Barclay and the authors of individual chapters. All rights reserved. No part of this work may be reproduced in any form or by any means without permission in writing from the publisher. The policy of Multilingual Matters/Channel View Publications is to use papers that are natural, renewable and recyclable products, made from wood grown in sustainable forests. In the manufacturing process of our books, and to further support our policy, preference is given to printers that have FSC and PEFC Chain of Custody certification. The FSC and/or PEFC logos will appear on those books where full certification has been granted to the printer concerned. Typeset by SAN Publishing Services. Printed and bound in the UK by the CPI Books Group Ltd.

Contents

Contributors

vii

Foreword Michael McCarthy

xi

1 Introduction Samuel Barclay and Paweł Szudarski

1

Part 1: Theorising and Testing Vocabulary

2 Innovation in Measures of Breadth and Depth of Vocabulary Knowledge Stuart Webb

11

3 Vocabulary Knowledge: Lexical Depth and Its Relationship with Out-of-Class Exposure Beatriz González-Fernández

21

4 Does One Size Fit All? Comparing Two Computer-Adaptive Algorithms for a Diagnostic Vocabulary Test Benjamin Kremmel

48

5 The Effects of Item Exposure Control on English Vocabulary Size Estimates in Computerized-Adaptive Testing Wen-Ta Tseng

65



Part 2: Formulaic Language

6

Formulaic Sequences and Second Language Learning Batia Laufer

7 The Development of Academic Collocations in Children’s Writing Phil Durrant and Mark Brenchley

89

99

8 Examining L2 Learners’ Confidence of Collocational Knowledge121 Ana Pellicer-Sánchez, Laura Vilkaitė-Lozdienė and Anna Siyanova-Chanturia v

vi  Vocabulary Theory, Patterning and Teaching

9 Predicting Parameters of Variation in the Use of Academic Multiword Expressions in University Student Writing Taha Omidian, Anna Siyanova-Chanturia and Phil Durrant

141

Part 3: Learning and Teaching L2 Vocabulary

10 Vocabulary Learning and Teaching Paul Nation and Averil Coxhead

169

11 The Role of Visual Imagery in Semantic Clustering Tomoko Ishii

176

12 The Deliberate Learning of L2-L1 Word Lists: Effectiveness and Learners’ Perceptions Suhad Sonbul and Marijana Macis

193

13 A Student-Generated Vocabulary Syllabus in an English for Academic Purposes (EAP) Context Ron Martinez

211

14 Conclusion: Looking Back and Moving Forward in Vocabulary Studies Paweł Szudarski and Samuel Barclay

234

Afterword Zoltán Dörnyei

245



248

Appendix: Final Vocabulary Test

Index

265

Contributors

Mark Brenchley is senior research manager at Cambridge Assessment, where he manages research in the areas of writing, lexico-grammar and auto-marking systems. He has a special focus on the development and application of corpora and corpus-based methodologies, managing both the Cambridge Learner Corpus and the Cambridge English Profile Corpus. With Phil Durrant, he co-built the Growth-in-Grammar Corpus, comprising L1 student writing from the compulsory phases of the English education system. He has published on linguistics in various media and has previously taught in the English education system for many years in various contexts, both mainstream and non-mainstream. Averil Coxhead is an associate professor who teaches undergraduate and postgraduate courses in Applied Linguistics and TESOL in the School of Linguistics and Applied Language Studies, Victoria University of Wellington, Aotearoa/New Zealand. She supervises postgraduate research at MA and PhD level. Averil is a co-author of Measuring the Vocabulary Size of Native Speakers with Paul Nation (John Benjamins, 2021), and English for Vocational Purposes: Language Use in Trades Education (Coxhead, Parkinson, Mackay & McLaughlin, Routledge, 2020). Her current research focuses on various aspects of specialised vocabulary, including multiword units, bilingual word lists, trades education and testing. Zoltán Dörnyei is a professor of Psycholinguistics at the School of English, University of Nottingham. He has published extensively on second language acquisition and language learning motivation, and he is the author of more than 25 books, including Research Methods in Applied Linguistics (2007, Oxford University Press), The Psychology of the Language Learner Revisited (2015, Routledge, with S. Ryan), Innovations and Challenges in Language Learning Motivation (2020, Routledge) and Engaging Language Learners in Contemporary Classrooms (2020, Cambridge University Press, with S. Mercer). Phil Durrant’s career in Applied Linguistics began in Turkey, where he taught English for several years. After learning about the vocabulary research of Norbert Schmitt and his colleagues at the University of Nottingham during a teacher training programme, Phil decided to move to Nottingham, where he did MA and PhD degrees under Norbert’s vii

viii  Vocabulary Theory, Patterning and Teaching

s­ upervision. He completed his thesis research on the topic of collocations in second language learning and went on to work at Bilkent University and, subsequently, at the University of Exeter, where he is now an associate professor of Language Education. Beatriz González-Fernández is a lecturer in Applied Linguistics and TESOL at the University of Sheffield. Her research interests focus around the theory, acquisition and pedagogy of vocabulary in second and foreign languages. Recently, she has investigated how foreign language users acquire multiple aspects of vocabulary knowledge and how this data can better inform our understanding and conceptualisation of the nature of overall vocabulary knowledge in second/foreign languages. Her current projects involve researching the acquisition of polysemy and homonymy in foreign languages under deliberate and incidental conditions. Tomoko Ishii finished her PhD degree under Nobert Schmitt in 2005. She is currently an associate professor at Meiji Gakuin University in Japan and coordinates its English-as-a-Foreign-Language (EFL) programme. Her PhD research was on the assessment of vocabulary knowledge, and her current research interests extend to various issues of vocabulary knowledge and learning, including cross-association and the knowledge of part of speech. She is also actively engaged in training sessions for secondary school teachers on vocabulary instruction. Benjamin Kremmel is Head of the Language Testing Research Group Innsbruck (LTRGI) at the University of Innsbruck, Austria, where he teaches and researches language learning, teaching and assessment. His research interests include vocabulary assessment, L2 reading assessment, diagnostic language testing, language assessment literacy and second language acquisition (SLA). His work has been published in Language Testing, Language Assessment Quarterly, Applied Linguistics, Language Teaching and TESOL Quarterly. Batia Laufer is a professor (emerita) of Applied Linguistics at the University of Haifa, Israel. She has lectured, supervised research and published widely on several areas of vocabulary acquisition: effective teaching, contribution to reading, testing, cross-linguistic influence, ease and difficulty in learning, dictionary use and attrition. Marijana Macis is a lecturer in TESOL and Applied Linguistics at the Manchester Metropolitan University. She completed her PhD in Applied Linguistics at the University of Nottingham, UK. Her research focuses on the teaching and learning of vocabulary. More specifically, she is interested in how learners acquire multiword combinations such as collocations as well as in the effectiveness of different teaching and learning methodologies for improving knowledge of formulaic sequences. She has publications in international peer-reviewed journals such as Language Teaching Research, ELT Journal and Reading in a Foreign Language.

Contributors ix

Ron Martinez is an English Language Specialist under the U.S. Department of State. He completed his PhD in English in 2011 at the University of Nottingham under the supervision of Norbert Schmitt and Michael McCarthy. He specialises in English for Academic Purposes (EAP) at the Universidade Federal do Paraná in Brazil where he founded and currently directs the writing centre. He also lectures at the University of California, Berkeley, where he teaches writing for academic publication on the Summer English Language Studies programme. Michael McCarthy is an emeritus professor of Applied Linguistics, University of Nottingham, UK. He is author/co-author/editor of 53 books, including Touchstone, Viewpoint, The Cambridge Grammar of English, English Grammar Today, From Corpus to Classroom, The Routledge Handbook of Corpus Linguistics, Innovations and Challenges in Grammar and titles in the English Vocabulary in Use series. He is author/coauthor of 120 academic papers. He was co-founder of the CANCODE and CANBEC spoken English corpora projects. His current research focuses on grammar, EAP and spoken fluency. He has taught in the UK, Europe and Asia, has given talks and workshops in 46 countries and has been involved in language teaching and applied linguistics for 55 years. Paul Nation is emeritus professor of Applied Linguistics in the School of Linguistics and Applied Language Studies at Victoria University of Wellington, New Zealand. He has taught in Indonesia, Thailand, the United States, Finland and Japan. His specialist interests are language teaching methodology and vocabulary learning. He has written many books about language teaching, including What Should Every EFL Teacher Know? and How Vocabulary is Learned. Taha Omidian is a PhD student at Victoria University of Wellington (New Zealand). He specialises in the use of corpus linguistic and computational methods to explore systematic patterns in language data. His research interests include corpus linguistics, quantitative linguistic research methods, English grammar, vocabulary, phraseology, register variation, language learning, language for specific purposes, multilingualism and academic writing. Ana Pellicer-Sánchez is an associate professor of Applied Linguistics and TESOL at the UCL Institute of Education. Her research centres around the teaching and learning of vocabulary in a second or foreign language, with a particular focus on learning from reading and multimodal input. Her recent research has made use of eye-tracking to explore the cognitive processes involved in vocabulary learning and reading comprehension. Anna Siyanova-Chanturia is a senior lecturer in Applied Linguistics at Victoria University of Wellington (New Zealand) and is a guest professor in Linguistics and Applied Linguistics at Ocean University of China (China). Anna’s research interests include psychological aspects of second

x  Vocabulary Theory, Patterning and Teaching

language acquisition, bilingualism, usage-based approaches to language acquisition, processing and use, vocabulary and multiword expressions, quantitative research methods (corpora, eye movements and EEG/ERPs). Suhad Sonbul is an assistant professor in the English Language Centre at Umm Al-Qura University, Saudi Arabia. She holds a PhD degree in Applied Linguistics from Nottingham University, UK. She has published in several international journals including Language Learning, Bilingualism: Language and Cognition, Applied Psycholinguistics and System. Her research interests include vocabulary teaching/learning, formulaic language with a specific focus on collocations and phrasal verbs, psycholinguistic measures of vocabulary development and quantitative research methods. Her latest project is a co-authored monograph Research Methods in Vocabulary Studies (John Benjamins). Wen-Ta Tseng is an associate professor in the Department of Applied Foreign Languages, National Taiwan University of Science and Technology. He received his MA and PhD from Nottingham University under the supervision of Professor Norbert Schmitt. His current research interests include the application of psychometrics to developing and validating new test instruments and the integration of psychometrics and meta-analysis in the area of second language acquisition. Laura Vilkaitė-Lozdienė is an assistant professor at Vilnius University, Lithuania. Laura’s current research focuses on vocabulary acquisition and use, and especially on formulaic language. She is also interested in questions of language processing and applying research-based knowledge to language pedagogy. Stuart Webb is a professor of Applied Linguistics at the University of Western Ontario. Before teaching applied linguistics, he taught English as a foreign language in Japan and China. His research interests include vocabulary studies, second language acquisition and extensive reading, listening and viewing. His latest books are How Vocabulary is Learned (with Paul Nation) and The Routledge Handbook of Vocabulary Studies.

Foreword

A Bit of History

I shall never forget our first meeting. I was Director at the Centre for English Language Education (CELE) at the University of Nottingham, a role I took over in 1991. The Centre’s administrative assistant tapped on my office door one day and announced a visitor who had slightly alarmed her. She told me a big American man in motorbike leathers was hoping to talk to me. I popped my head out and, sure enough, a tall, muscular figure was striding down the corridor towards my room, leather-clad, crash-­ helmet under his arm. Evel Knievel? Terminator? Not really. With a warm smile, Norbert Schmitt introduced himself. At that time, Norbert had already left behind his farming background, had decided on an academic future and was keen to study the acquisition of English vocabulary with a view to a doctorate. He was, so to speak, doing the rounds, testing the waters to see if and where anyone shared his interests. I am delighted to say he chose Nottingham and completed his PhD in 1997. It may seem strange to the ears of younger readers to be told that, in the early 1990s, vocabulary had not by any means fully morphed out of its ‘Cinderella’ status in language teaching and learning, the decades of ‘unredeemed neglect’ that Ron Carter and I complained of in 1988 (Carter & McCarthy, 1988: 41). If I may be forgiven a brief historical excursion, it is worth dwelling on this point, as it explains so much of the nature of the environment in which Norbert Schmitt’s vocabulary work began and took root, and the immense contributions Norbert has made to the field ever since. Carter and I were writing at the tail end of a long era during which structuralism, with its emphasis on syntax and phonology, had prevailed in foreign and second language teaching. Although the great pioneer of our applied linguistic profession Harold Palmer had, in the earlier part of the 20th century, put forward a set of systematic principles for the teaching of vocabulary, referring to himself in 1936 as being in ‘a state of revolt against the giving to pupils as a first vocabulary any haphazard assortment of words, as if taken at random out of a sack’ (Palmer, quoted in Smith, 1999: 36), Palmer’s heritage was seen mostly in his emphasis on pronunciation, grammar and structural drilling, as manifested in his

xi

xii  Vocabulary Theory, Patterning and Teaching

works published in the immediate aftermath of the First World War (Palmer, 1922, 1924). One vestige of the structuralist heritage was seen in the situational syllabuses that threaded through the pedagogical fabric of the 1960s and later, which Mitchell criticised: such syllabuses offered ‘little more language than the learning by heart of situationally relevant phrases and vocabulary, and failed to lead the learner systematically to linguistic independence and creativity’ (Mitchell, 1994: 35–36). The communicative movement, ushered in in the 1970s, in its strongest versions did not help either: it was assumed in many circles that linguistic needs (grammar, vocabulary) would be met as they emerged during the practice of communicative utterances – language as ‘doing’. The notion of an organised vocabulary syllabus or an understanding of the psycholinguistic dimension of vocabulary acquisition had low priority (McCarthy, 1984; Mitchell, 1994: 36). That is not to say there were no voices crying in the wilderness. Most notably, Jack Richards (1976) struck a new chord in considering psycholinguistic and sociolinguistic questions to be germane to the notion of lexical competence. Richards asserted that the question of ‘knowing’ a word involved not only the syntactic and derivational features of a word but also the problem of how words are remembered, and word knowledge necessarily embraced psychological and social dimensions such as word association and functional and situational use. This inclusive approach, based neither entirely in descriptive linguistics nor entirely in psycholinguistics or sociolinguistics, mapped out territory for a potentially greatly expanded world of research in vocabulary studies. Richards’ conclusion, at a time when vocabulary was in the pedagogical doldrums, was that ‘beyond the elementary levels of instruction, a major feature of a second language program should be a component of massive vocabulary expansion’ (Richard, 1976: 84). Richards’ paper was followed by another in TESOL Quarterly on the subject of vocabulary instruction, by Elliott Judd (Judd, 1978). This article too challenged the neglected status of vocabulary in second language pedagogy and questioned the assumption that vocabulary teaching and learning should be swallowed up in reading and listening classes and the parallel belief that vocabulary expansion should only be a concern once the basics of syntax were under control. Judd argued for vocabulary knowledge to be taught as a skill per se and that vocabulary teaching should be commenced at an early stage in a course of instruction. Properly organised, direct vocabulary teaching would avoid the ‘randomized appearance of vocabulary’ (1978: 73), a criticism that echoed Palmer’s metaphor of plucking vocabulary at random out of a sack. Anne-Marie Cornu (1979), in a paper emanating from the Katholieke Universiteit te Leuven, Belgium, offered another important, though perhaps now undervalued, contribution to the debate. Cornu advocated a

Foreword xiii

semantic field approach to vocabulary and at the same time recognised the need for psycholinguistic understanding with regard to the problem of vocabulary retention, concluding that vocabulary learning could not just be ‘the assimilation of lists of words’ (Cornu, 1979: 272). This approach was continued by Joanna Channell (1981) and later by another Louvain scholar, Ludo Beheydt (Beheydt, 1987). The (Re)birth of Vocabulary Study

One figure whose presence has been felt in our field for a very long time and whose work influenced Norbert Schmitt (and most of us) profoundly is Paul Meara. Meara (1980) is an article considered to represent a launchpad for the research agenda in second language vocabulary acquisition studies which Norbert and so many after him have embraced. Seen from the perspective adopted by Ronald Carter and me in our work on vocabulary, Meara’s work was strikingly different. For Carter and me, the 1980s was indeed a period of ferment in the study of vocabulary and language teaching and learning. But we were more interested in the role of linguistic description in the understanding of the lexicon, spurred on by our mentor John Sinclair, whose corpus linguistic research was giving empirical substance to Firthian notions of meaning in context, collocation and chunking, an inspiration which I hope was reflected in our 1988 book. I had also around that time become interested in discourse analysis under Sinclair’s guidance, and I advocated a more discourse-sensitive approach to lexical meaning and function and how it might be applied to Englishas-a-Foreign-Language (EFL) vocabulary teaching (McCarthy, 1984). Meara took a different line, one that was to have greatly different consequences. Meara was interested in acquisition, an aspect of vocabulary study which, he pointed out, had been given ‘short shrift from applied linguistics’ (Meara, 1980: 221). And his comment on the type of statistical-­ descriptive work on frequency, which at the time was becoming facilitated by developments in computing, placed him in stark contrast to what I have described as the approach taken by Carter, myself and others enamoured of the burgeoning field of corpus linguistics: ‘this work is characteristic in that it concentrates on what is basically a problem to do with the management of learning, rather than with the learning process itself – i.e. the object of this type of research is to decide what words are to be taught, not to find out how words are actually learned’ (Meara, 1980: 224). He had a point. Among the issues which needed addressing for Meara were vocabulary size (just how many words can learners acquire in a given time, and is there an upper limit?), the need to understand ‘the complex patterns of meaning relationships that characterise a proper, fully formed lexicon, as opposed to a mere word list’ (1980: 225), the implementation of proper

xiv  Vocabulary Theory, Patterning and Teaching

longitudinal studies, the absorption of words over time into the learner’s lexicon, the role of bilingualism in assisting our understanding of second language learners, especially on the question of storage in the mental lexicon, the sematic relations between words, word association and how the formal aspects of words are acquired (Meara & Ingle, 1986). This was a research agenda vast enough to preoccupy a century of researchers – it has impressively remained current for almost half of one – and it led Meara to reflect just over a decade later that studies of vocabulary acquisition had mushroomed and that it was almost impossible to keep up with the field even if you did nothing else (Meara, 1995). We should remember that much of the content of the present volume, representing as it does the voices of a well-established research community, in the 1980s would have been revolutionary. The early 1980s saw the emergence on to the vocabulary studies scene of another giant on whose shoulders Norbert has been proud to stand: Paul Nation. Nation (1982) is another of those landmark papers which anyone interested in vocabulary teaching and learning should read. Distinguishing between indirect and direct vocabulary learning, Nation asks pertinent questions which at that time were not at the top of the profession’s research agenda: ‘How does indirect learning of vocabulary by reading compare with and complement indirect learning by listening? Are vocabulary exercises which directly teach the production of vocabulary in speaking and writing more sure and efficient in developing a productive vocabulary than indirect learning? Is directly learned vocabulary retained in the memory for a longer time than indirectly learned vocabulary?’ (Nation, 1982: 15). The paper also looks at learning word pairs (L2 words and their L1 translation equivalents) in relation to the number of encounters with them and their retention, and the most efficient ways of dividing up the learning task, as well as addressing the question of word difficulty, and challenging the belief that words should always be taught in context. Nation concluded by warning against inadequate control of variables in experimentation and called for future research to approach vocabulary learning ‘with a richer idea of what it means to know a word and of how words are learned’ (Nation, 1982: 32). The breadth and depth of Nation’s work is best appreciated in his 2001 book, where we see not only the psycholinguistic preoccupations rehearsed in the 1982 paper treated in detail but also an awareness of the significant role of description in bringing to light the phenomena of chunking and collocation, as well as formal aspects of affixation in his work with Bauer (Bauer & Nation, 1993; Nation, 2001). What we see as the 1980s draw to a close is a head of steam propelling concerns with the place of vocabulary in language teaching and learning and dissatisfaction over its perceived neglect. This was the world that Norbert Schmitt entered: vocabulary description, vocabulary teaching and vocabulary acquisition, all rubbing shoulders in the post-structuralist world. How was he to make sense of its almost chaotic complexity?

Foreword xv

Seeing the Wood for the Trees

A short preface like this one can scarcely do justice to Norbert Schmitt’s ability to immerse himself in the dynamic and multipronged research agenda that had been established in the 1980s in lexical description and second language lexical acquisition and pedagogy – an agenda necessarily only briefly outlined in the previous sections. Norbert’s total immersion in this milieu, with its attendant danger of prompting a butterfly existence in research, hopping from one promising bloom to another, in fact led him to forge a rigorous pathway combining the best of description with a profound understanding of the issues involved in acquisition while never losing sight of the practical world of vocabulary teaching. It is not the purpose of this preface to survey Norbert’s work. The plentiful references to it in the chapters of this book provide more than an adequate paean to his achievements. Here, I will simply point to a few of his studies in an attempt to capture the character and breadth of his research. With his wife, Diane, Norbert compiled a list of vocabulary learning strategies which, in my opinion, has not been surpassed, an account of which may be found in Schmitt and Schmitt (1993) and further in Schmitt (1997). Apart from providing the full catalogue of strategies, the survey of learners’ reactions to them provide some surprises, with things in vogue at the time (e.g. the keyword approach, word imaging, group work on words) scoring poorly. The study characterised Norbert’s attention to method and control of detail, even in this ‘early’ work. He is not, and has never been, one to rush to conclusions without careful thought and deliberation. The other feature of the 1993 paper which, for me, is the hallmark of Norbert’s philosophy, is his willingness and ability to collaborate. Norbert and I worked together on an edited book (Schmitt & McCarthy, 1997) and, although I had supervised his PhD at Nottingham, like all good doctoral students, he had soon overtaken his mentor in knowledge and vision. We wanted our book to represent that research agenda which the 1980s had thrown up and which is described in this preface. Norbert had no difficulty navigating and editing the descriptive sections, where the growing influence of corpus linguistics came to the fore, but equally handled with aplomb the chapters on vocabulary acquisition and pedagogy and was able to see connections across the sections of the book and to tease out recurrent themes which my lazy brain had missed. Norbert always seemed to be able to see the wood for the trees. One of his great talents has always been the ability to make sense of formal descriptive issues and see their connections and relevance to language learners, as seen, for example, in his collaboration with Cheryl Boyd Zimmerman in their study of learners’ grasp of derivational word forms (Schmitt & Zimmerman, 2002).

xvi  Vocabulary Theory, Patterning and Teaching

Nottingham provided an ideal environment for Norbert in many ways (a situation that was ideal for Ron Carter and me, too, in that we held on to Norbert, who became a valued colleague). We already had a solid tradition of descriptive vocabulary studies in the work of Ron Carter and Joanna Channell, we were building spoken corpora and we wanted to use our expertise and resources to explore the role and status of lexis in language learning and teaching. Carter and I had become interested in chunks or formulaic sequences, which seemed to be ubiquitous in our corpus data. Norbert immediately saw the profound implications of formulaic sequences, and his 2004 edited book on the subject covers not only the full range of research in the field but shows the breadth of Norbert’s knowledge in the area in the fact that he co-authored four of the chapters, a truly remarkable achievement (Schmitt, 2004). Once again, the work takes on the imperatives of the 1980s research agenda but enhances the exegesis with the latest developments in corpus linguistics, psycholinguistic underpinning and sociocultural theory. If one could summarise Norbert’s approach in a few words, it would be that his intense curiosity about language learning, born of his years as a language teacher in Japan, has kept human subjects at the forefront of his investigations of questions of vocabulary. Not content with the mechanistic output of corpus frequency lists, his work with Bruce Dunham (Schmitt & Dunham, 1999) explores native and non-native speakers’ intuitions of word frequency, with some fascinating and insightful results delivered with characteristic caution and modesty. As mentioned, a central feature of Norbert’s research has always been collaboration and, in later years, the nurturing of the next generation of scholars in the field of vocabulary studies. The present volume bears witness to the profound respect with which his colleagues, associates and former students around the world regard him. He is one of the people who has made sure that the great debate of the 1980s over the need to reinvigorate vocabulary teaching and learning has been translated into action in an enduring research agenda which continues to push the field forward. In his ability to embrace and grasp the two worlds of description and acquisition and to see what they mean in relation to each other and to pedagogy, he is a man for all seasons. I am privileged to know him as supervisor, colleague and friend. Michael McCarthy Cambridge, June 2020 References Bauer, L. and Nation, P. (1993) Word families. International Journal of Lexicography 6 (4), 253–279. Beheydt, L. (1987) The semantization of vocabulary in foreign language learning. System 15 (1), 55–67.

Foreword xvii

Carter, R.A. and McCarthy, M.J. (1988) Vocabulary and Language Teaching. Harlow: Longman. Channell, J. (1981) Applying semantic theory to vocabulary teaching. ELT Journal XXXV (2), 115–122. Cornu, A.-M. (1979) The first step in vocabulary teaching. The Modern Language Journal 63 (5/6), 262–272. Judd, E.L. (1978) Vocabulary teaching and TESOL: A need for reevaluation of existing assumptions. TESOL Quarterly 12 (1), 71–76. McCarthy, M.J. (1984) A new look at vocabulary in EFL. Applied Linguistics 5 (1), 12–22. Meara, P. (1980) Vocabulary acquisition: A neglected aspect of language learning. Language Teaching 13 (3–4), 221–246. Meara, P. (1995) Editorial: Single-subject studies of lexical acquisition. Second Language Research 11 (2), i–iii. Meara, P. and Ingle, S. (1986) The formal representation of words in an L2 speaker’s lexicon. Second Language Research 2 (2), 160–171. Mitchell, R. (1994) The communicative approach to language teaching: An introduction. In A. Swarbrick (ed.) Teaching Modern Languages (pp. 33–42). London: Routledge. Nation, I.S.P. (1982) Beginning to learn foreign vocabulary: A review of the research. RELC Journal 13 (1), 14–36. Nation, I. (2001) Learning Vocabulary in Another Language. Cambridge: Cambridge University Press. Palmer, H.E. (1922) The Oral Method of Teaching Languages. Cambridge: W. Heffer and Sons Ltd. Palmer, H.E. (1924) A Grammar of Spoken English on a Strictly Phonetic Basis. Cambridge: W. Heffer and Sons Ltd. Richards, J.C. (1976) The role of vocabulary teaching. TESOL Quarterly 10 (1), 77–89. Schmitt, N. (1997) Vocabulary learning strategies. In N. Schmitt and M. McCarthy (eds) Vocabulary: Description, Acquisition and Pedagogy (pp. 199–227). Cambridge: Cambridge University Press. Schmitt, N. (ed.) (2004) Formulaic Sequences. Amsterdam: John Benjamins. Schmitt, N. and Dunham, B. (1999) Exploring native and non-native intuitions of word frequency. Second Language Research 15 (4), 389–411. Schmitt, N. and McCarthy, M. (eds) (1997) Vocabulary: Description, Acquisition and Pedagogy. Cambridge: Cambridge University Press. Schmitt, N. and Schmitt, D.R. (1993) Identifying and assessing vocabulary learning strategies. Thai TESOL Bulletin 5 (4), 27–33. Schmitt, N. and Zimmerman, C. (2002) Derivative word forms: What do learners know? TESOL Quarterly 36 (2), 145–171. Smith, R.C. (1999) The Writings of Harold E. Palmer: An Overview. Tokyo: Hon-no-Tomosha.

1 Introduction Samuel Barclay and Paweł Szudarski

Research into second language vocabulary acquisition has a long history, with an almost exponential growth in studies of vocabulary learning and teaching over the past few decades. Some features of the field have changed little during this time; a time when Norbert Schmitt, the researcher whose work this volume celebrates, became interested in foreign language teaching. For instance, in 1980, Paul Meara recognised that vocabulary causes language learners considerable difficulty, with learners of all levels claiming vocabulary acquisition to be the source of many of the problems they face (Meara, 1980). In a similar vein, many years later, Webb and Nation stated that learners place a great deal of importance on vocabulary learning and that it is a central aspect of second language competency (Webb & Nation, 2017). Vocabulary was, is and will continue to be a major component of foreign language learning. Other aspects of vocabulary studies, however, have undergone considerable change in this time. Paul Meara began his seminal paper by suggesting that vocabulary acquisition had been largely ignored by the research community (Meara, 1980). The landscape today is remarkably different, with much of the research on vocabulary learning conducted over the past century taking place within the last 20 years (Nation, 2013). Thankfully, vocabulary now receives the kind of research attention that reflects its significance to learners and their development. Reflecting this rise in popularity, we now have several publications focusing exclusively on vocabulary studies. There are research manuals (Nation & Webb, 2011; Schmitt, 2010), a handbook (Webb, 2020), special editions of journals (e.g. the January 2017 issue of Language Teaching Research, the January 2018 issue of the ITL International Journal of Applied Linguistics, the 2021 issue of Second Language Research) and edited volumes, of which this book is an example. With this rapid increase in research output, it is necessary to take stock of the way in which the field has developed (see McCarthy, this volume). Doing so helps us appreciate that the increasing significance placed on vocabulary learning and teaching is not coincidental, nor is it just a natural function of the importance of the subject matter. Development has occurred because a few key figures have relentlessly championed and evidenced the centrality of vocabulary in second language learning. This volume contributes to the 1

2  Vocabulary Theory, Patterning and Teaching

literature on, and explores the growth of, vocabulary studies by celebrating the career of one of these key figures, Norbert Schmitt. Norbert Schmitt’s Role in the Development of the Field

It is beyond the scope of this introduction to discuss in detail Nobert’s influence on the field of vocabulary studies and the role his research, and that of his doctoral students, has played in creating the space many of us inhabit today; besides, the foreword to this volume has accomplished this so eloquently already. Rather, here we explore his impact on the field with the help of two illustrations (Figures 1.1 and 1.2). These maps, which detail co-citation analysis, have been kindly provided by Paul Meara. Co-citation analysis involves investigating citation patterns to find authors (i.e. nodes) cited together and who are therefore likely to share a similar research focus. It also demonstrates the relative centrality of an author with the betweenness centrality index. This is a measure of how often a node would appear in a path between two random nodes in the network and is indicated by the size of a node.

Figure 1.1  Co-citation analysis of Applied Linguistics from 2000 to 2004. Node size indicates betweenness centrality (see Meara, 2018)

Introduction 3

Figure 1.2  Co-citation analysis of The Routledge Handbook of Vocabulary Studies (Webb, 2020)

The first co-citation map (Figure 1.1) is based on papers appearing in the journal Applied Linguistics between 2000 and 2004. This shows that there were several clusters of co-citations but that vocabulary research was still a small and somewhat marginal aspect of the Applied Linguistics enterprise. This is demonstrated by the relative size of the vocabulary cluster and also by the limited number of connections with the other parts of the map. The vocabulary cluster is to the north east of Mike Long, with Norbert between Paul Meara, Mari Wesche and Batia Laufer. Interestingly, a similar analysis today would paint quite a different picture, reflecting the increasingly central position that vocabulary studies occupy within applied linguistics and the many names currently researching it. In fact, the picture some years later is very different. The second cocitation map (Figure 1.2) illustrates an analysis of the 34 chapters of the recently published The Routledge Handbook of Vocabulary Studies (Webb, 2020). This map indicates the 102 most cited sources in this handbook and the frequency with which each pair of nodes was co-cited. This figure demonstrates that vocabulary studies is now a thriving research community with a considerable number of contributors. It is interesting to note, however, that many of the researchers from the vocabulary cluster in Figure 1.1, including Paul Nation, Paul Meara and Batia Laufer, are similarly central figures of the field today, with one notable addition, Stuart Webb. Another key researcher is Norbert Schmitt, and Figure 1.2 shows just how important Norbert is to the field, revealing that he was cited in every chapter of the book and showing that he was co-cited with every author. Again, the nodes are sized according to their betweenness centrality. The figure demonstrates that Nobert has the highest betweenness centrality score, highlighting not just his importance to the field in

4  Vocabulary Theory, Patterning and Teaching

general, but the diversity of areas he has explored in his research. Looking at these figures, it is clear that the field has changed dramatically over the past 20 years but that Norbert’s presence has remained constant and developed over the period. It seems only fitting, therefore, that we celebrate his work with this volume. If we may be permitted a slight diversion at this point, another aspect of Nobert’s career that warrants comment is his championing of early career researchers. One way he has done this is through the Vocabulary Research Group at the University of Nottingham. This was a group of academics and PhD students who met once a week to discuss vocabulary issues. We reviewed papers together, read each other’s work, discussed thorny problems and supported one another in our various research endeavours. This is not to say, however, that the weekly sessions were all sunshine and rainbows. Giving a presentation to the group was famously nerve-racking, with even the briefest of talks provoking a barrage of questions. It was not unusual, for instance, for a 15-minute runthrough of a conference presentation to elicit 2 hours of questions. Such thorough interrogation, daunting though it was, not only led to more interesting and robust research projects but also taught the student members about criticality and collaboration and gave us confidence in our work. In many ways, it was an academic internship just as much as it was a studentship. We have used the past tense to discuss the Vocabulary Research Group because it ceased to meet regularly in 2019, after the last of Norbert’s students, Beatriz González-Fernández (Chapter 2) completed her period of study. However, empowered by their own experiences in the group, many former members have started similar enterprises at their own institutions, building on previous lexical findings, taking research into new and innovative directions and, to borrow Nation and Coxhead’s metaphor of a tree (this volume), using the seeds planted by Norbert to feed further developments within the field. This has led to vocabulary research hubs around the world and an ever-widening academic family: exciting times indeed. In many ways, this volume is a product of the Vocabulary Research Group. The authors of the experimental chapters are largely former members, or students of former members. There are also several collaborations between former students, some who did not even study at the same time but who are connected by their shared experience of the group, Sonbul and Macis (Chapter 10) being a case in point. The volume shows the extent to which ripples caused by one academic can be maintained many years later and likewise cause ripples of their own. All the authors of the experimental chapters of this volume, in one way or another, can trace their development back to the Vocabulary Research Group. It is fair to say that we all learnt a great deal from attending those meetings in CRAL Meeting Room 1 in the Trent Building, University of Nottingham (evidence of which can be seen on Norbert’s website: www.norbertschmitt.co.uk).

Introduction 5

Goals and Structure of the Volume

The central aim of this volume is to contribute to the growing literature on vocabulary studies and to inspire future research endeavours in this area. It takes the career of Norbert Schmitt as a framework, dividing the chapters into sections that represent the focus of much of his research contribution: theory and assessment, formulaic language and teaching and learning. The book is largely intended for researchers and graduate students of applied linguistics; however, as each chapter contains a thorough literature review, those unfamiliar with vocabulary studies will be able to gain a broad understanding of the field. Additionally, pedagogical suggestions are given throughout, making this useful reading for English language teachers and language practitioners. Finally, the volume also makes several suggestions for future research projects, meaning that it actively seeks to aid the development of the field. The volume is divided into three parts, each with a consistent theme. These themes have been selected due to their importance to the field and because they are areas to which Norbert has made considerable contribution throughout his career. Each part contains three research chapters that have been authored, or co-authored, by former PhD students of Norbert – his academic children, as he likes to call them. The sections begin with a discussion chapter written by central figures from the field of vocabulary studies – those who have, along with Norbert, played such a crucial role in raising the field to its current position of prominence. These discussion chapters consider the experimental studies in their respective sections, reflect on Norbert’s contribution to the theme of the section and make suggestions for future research. Importantly then, the volume does not just look backwards but includes cutting-edge studies and looks to inspire research and researchers in the future – something to which Norbert has dedicated his whole career. The first part deals with theorising and testing vocabulary. There is currently a call for a practical theory of vocabulary acquisition (Schmitt, 2019) and for vocabulary tests with increasingly robust validity arguments (Schmitt et al., 2020). Part 1 contributes to this discussion. It begins (Chapter 2) with a commentary by Stuart Webb outlining key aspects of the vocabulary knowledge construct and highlighting the need for robust assessments of these aspects. This discussion paves the way for the three experimental studies that follow. In Chapter 3, Beatriz GonzálezFernández investigates the multidimensionality of vocabulary knowledge and considers the association between different aspects of word knowledge and various extramural activities such as viewing English TV. In Chapters 4 and 5, the focus shifts to measuring vocabulary knowledge via computer-adaptive tests. Benjamin Kremmel (Chapter 4) compares two adaptive formulae for a computer-based diagnostic test of vocabulary knowledge. Wen-Ta Tseng (Chapter 5) then explores the effect of another test variable, item exposure, on test-taker performance.

6  Vocabulary Theory, Patterning and Teaching

Part 2 of the volume looks at formulaic language and phraseology, another key area of investigation within vocabulary studies and one to which Norbert has made considerable contributions (e.g. Schmitt, 2004; Siyanova-Chanturia & Pellicer-Sánchez, 2019; Wray, 2002). It opens with a discussion by Batia Laufer (Chapter 6) who considers the research on, and Norbert’s contribution to, various aspects of formulaic language: knowledge and use, processing, acquisition and selection. The three experimental studies that follow directly relate to these dimensions. In Chapter 7, Phil Durrant and Mark Brenchley explain the development of collocations in the writing of English L1 children. In Chapter 8, three former students of Norbert – Ana Pellicer-Sánchez, Laura VilkaitėLozdienė and Anna Siyanova-Chanturia – collaborate on a study investigating the association between collocational knowledge and the level of confidence learners have in this knowledge. In the final study of the section (Chapter 9), Taha Omidian, Anna Siyanova-Chanturia and Phil Durrant address variation in the use of academic formulas by university students in the UK. Part 3 explores vocabulary teaching and learning. This is an area of particular significance to Norbert who has striven throughout his career not only to conduct research with definable pedagogical application but also to communicate these findings to educational stakeholders through conference presentations, practitioner publications and spending time discussing issues directly with teachers. In Chapter 10, Paul Nation and Averil Coxhead discuss the three experimental studies in this section, consider Norbert’s contribution to current pedagogical practice and make several interesting recommendations for future studies. In Chapter 11, Tomoko Ishii explores the effect of visual and semantic relatedness on learning. Suhad Sonbul and Marijana Macis (Chapter 12) then investigate the effectiveness of word lists for vocabulary learning. Finally, Ron Martinez (Chapter 13) details a project using student-generated word lists in an English for Academic Purposes course. The volume concludes by bringing together the common themes from the various sections (Chapter 14) and closes with an afterward by Zoltán Dörnyei in which he reflects on the hardships endured by working alongside his colleague, Norbert Schmitt, for all these years. Overall, the volume acts as a microcosm of the field as it currently stands. It contains studies from almost every corner of the world, reflecting the truly global nature of the research interest in vocabulary studies. Many of the experimental studies employ advanced research methods such as eye-tracking and computer-adaptive testing, highlighting the breadth of research designs and the increasing methodological rigour of the field. Several studies also employ advanced statistical analyses such as mixed effects and item response theory modelling, again mirroring the move within vocabulary studies to analyses that allow us to better understand the complexity of lexical acquisition. Finally, the diverse foci of the

Introduction 7

studies in this volume illustrate the way the field has expanded. Not only do we now have a greater quantity of studies from which to draw (Nation, 2013), but the subject matter touches on an increasingly broad range of areas (the recently published The Routledge Handbook of Vocabulary Studies [Webb, 2020] had as many as 34 entries, for example). Crucially, these features of the volume also reflect Norbert Schmitt’s academic life. He has supervised PhD students from across the world, and his research has engaged a global audience. He has embraced innovative methodologies and has consistently asked complex questions in his studies. And his research has touched on a diverse range of subjects, contributing to a broad knowledge base. Therefore, we hope the volume acts as a fitting celebration of his prolific (and ongoing) contribution. References Meara, P. (1980) Vocabulary acquisition: A neglected aspect of language learning. Language Teaching 13 (3–4), 221–246. doi: 10.1017/S0261444800008879 Meara, P. (2018) Mapping vocabulary research. Journal of Foreign Language Education and Technology 3 (1), 280–308. Nation, I.S.P. and Webb, S. (2011) Researching and Analysing Vocabulary. Boston: Heinle, Cengage Learning. Nation, P. (2013) Learning Vocabulary in Another Language. Cambridge: Cambridge University Press. Schmitt, N. (2010) Researching Vocabulary: A Vocabulary Research Manual. Basingstoke: Palgrave Macmillan. Schmitt, N. (2019) Understanding vocabulary acquisition, instruction, and assessment: A research agenda. Language Teaching 52, 261–274. Schmitt, N. (ed.) (2004) Formulaic Sequences: Acquisition, Processing and Use (Vol. 26). Amsterdam: John Benjamins. Schmitt, N., Nation, P. and Kremmel, B. (2020) Moving the field of vocabulary assessment forward: The need for more rigorous test development and validation. Language Teaching 53 (1), 1–12. doi:10.1017/S0261444819000326 Siyanova-Chanturia, A. and Pellicer-Sánchez, A. (2019) Understanding Formulaic Language: A Second Language Acquisition Perspective. New York and London: Routledge. Webb, S. (2020) The Routledge Handbook of Vocabulary Studies. London: Routledge. Webb, S. and Nation, P. (2017) How Vocabulary Is Learned. Oxford: Oxford University Press. Wray, A. (2002) Formulaic Language and the Lexicon. Cambridge: Cambridge University Press.

2 Innovation in Measures of Breadth and Depth of Vocabulary Knowledge Stuart Webb

Norbert Schmitt’s publications were essential reading during my doctoral studies. Norbert’s studies were well-written, provided detailed literature reviews and were carefully designed, so with each study that I came across, I learned a great deal. However, at the same time, this was also somewhat frustrating, because he was publishing about many of the things that I was planning to research. Being new to research, I thought that this would diminish my chances of getting manuscripts published. It took me a while to learn that there is space for multiple articles about one topic, and indeed, this diversity in studies of a single topic is a good thing. With each article of Norbert’s that I read, I became more and more interested in Norbert Schmitt’s academic journey. I soon learned that we had several things in common: we both had lived in Japan, played ice hockey and were interested in research that informed the teaching and learning of vocabulary. I also learned that we had some differences. I cannot build and fly airplanes, nor do I play the bagpipes. Norbert’s great success with his research proved a useful model to work toward. However, while Norbert’s repeated success with his research and publications provided useful goals, his exceptional production made those goals unreachable. One of the features that is notable about Norbert Schmitt’s research is a consistent focus on methodology and assessment. Developing new research instruments and research designs rather than simply using what was available and had been previously used before is one of the biggest reasons for his great success as a researcher. His studies of depth of word knowledge involved developing new approaches to evaluating different aspects of L2 vocabulary knowledge (e.g. Schmitt, 1998a; Schmitt & Zimmerman, 2002), examining the relationships between different aspects of depth (González-Fernández & Schmitt, 2020) and looking at issues related to testing vocabulary depth (e.g. Schmitt, 1998b, 1998c, 2010). His studies of breadth of L2 vocabulary knowledge involved 11

12  Part 1: Theorising and Testing Vocabulary

improving on an existing test (Schmitt et al., 2001), examining the degree to which different test features might influence scoring (Gyllstad et al., 2015; Kremmel & Schmitt, 2016), as well as examining the relationship between depth and breadth of vocabulary knowledge (Schmitt, 2014). This chapter highlights several of Norbert Schmitt’s contributions to research on depth and breadth of vocabulary knowledge. It also previews the following chapters by his former students Beatriz González-Fernández, Benjamin Kremmel and Wen-Ta Tseng. Finally, it provides some discussion of several topics related to measures of breadth and depth of vocabulary knowledge that deserve further research. Depth of Vocabulary Knowledge

Depth of vocabulary knowledge has been defined in several ways but is typically characterized as different types of word knowledge that need to be learned in order to fully know a word (e.g. Henriksen, 1999; Miller, 1999; Nation, 2001; Read, 2004; Richards, 1976). These different components of knowledge reflect a multidimensional view of word knowledge. For many years, vocabulary learning was operationalized as gains in knowledge of form-meaning connection, whether participants could recall or recognize the meanings of words when presented with their L2 forms, or recall or recognize the L2 forms of words when presented with their meanings. However, after Nation (1990, 2001) provided comprehensive descriptions of the different aspects of vocabulary knowledge, researchers began to explicitly focus on measuring dimensions of vocabulary knowledge apart from form-meaning connection. Tests were created to explicitly measure depth of vocabulary knowledge as a whole (e.g. Read, 1993, 1998; Wesche & Paribakht, 1996), and studies of vocabulary learning began to measure multiple aspects of knowledge (e.g. Pigada & Schmitt, 2006; Schmitt, 1998a; Schmitt & Meara, 1997; Webb, 2005, 2007a, 2007b). One of the innovative methodological features of studies by Norbert Schmitt was measuring multiple aspects of vocabulary knowledge to more accurately assess lexical development. Schmitt and Meara (1997) investigated Japanese English-as-a-Foreign-Language (EFL) learner knowledge of word parts (inflectional and derivational suffixes), associations (including both semantic and syntactic associates) and form-meaning connection (Vocabulary Levels Test (VLT) scores) at the start and end of a school term. Schmitt (1998a) examined the lexical development of 11 words by interviewing three EFL learners to determine their knowledge of written form, associations (semantic and syntactic associates), grammatical functions (identifying word class, creating derivatives) and concept and referents (providing the different meanings of polysemous words). Pigada and Schmitt (2006) and Pellicer-Sánchez and Schmitt (2010) also employed interviews to measure different aspects of word knowledge. Most recently,

Innovation in Measures of Breadth and Depth of Vocabulary Knowledge  13

González-Fernández and Schmitt (2020) examined the relationships between four aspects of vocabulary knowledge (collocation, concept and referents [the different meanings a word conveys], word parts and formmeaning connection) by measuring the extent to which participants could recognize and recall each dimension. Three more studies by Norbert Schmitt (Schmitt, 1998b, 1998c; Schmitt & Zimmerman, 2002) highlighted some of the methodological challenges of measuring different aspects of productive vocabulary knowledge. Schmitt (1998b) described designing and scoring word association tests and shed further light on how many responses should be elicited and how these responses might be scored. Schmitt (1998c) did the same for productive knowledge of collocation by highlighting important features of research that need to be considered such as the span between the node word and the collocate, the measures of strength of co-occurrence and decisions that need to be made when scoring responses. Schmitt and Zimmerman (2002) looked at productive knowledge of word parts (production of derivatives) using two tests. This study also focused on different ways to measure knowledge (two tests were used, one with context and one without), how target items (prompt words) were selected and how responses might be scored. These early studies are important because they provide a foundation from which research investigating depth of vocabulary knowledge could be developed. There are also studies by other researchers that provide different models of how aspects of vocabulary depth might be measured. Webb (e.g. 2005, 2007a, 2007b) investigated the degree to which five different aspects of vocabulary knowledge (written form, form-meaning connection, association, collocation and grammatical function) were learned through incidental and intentional vocabulary learning tasks using a battery of 10 paper-and-pencil tests. Chen and Truscott (2010) measured four aspects of word knowledge: written form, association, grammatical function and form-meaning connection to determine the effects of repetition and lexicalization on vocabulary learning through reading. Koizumi and In’nami (2013) measured four aspects of depth: word parts, association, collocation and form-meaning connection to look at the relationship between vocabulary depth and speaking proficiency. Milton and Hopkins (2006) examined how knowledge of words varies when the spoken forms of item cues are provided in relation to their written forms. Together, these studies reveal how vocabulary knowledge varies across different aspects of lexical mastery. They also provide additional criteria to consider when measuring different aspects of vocabulary knowledge. One positive outcome of the studies of vocabulary depth has been the increased awareness of the need to investigate knowledge of aspects of vocabulary knowledge beyond form-meaning connection. In addition, studies of depth of word knowledge have also shown that a single test may not effectively measure gains in knowledge. Today, there is usually an

14  Part 1: Theorising and Testing Vocabulary

expectation in research that multiple tests are used to measure vocabulary learning. However, although studies of vocabulary learning published in recent years usually use multiple measures, they still tend to measure knowledge of form-meaning connection alone. Thus, there remains a need for studies to measure more than one aspect of vocabulary knowledge to more accurately assess vocabulary learning. The value in testing aspects other than form-meaning connection is apparent when considering the challenges that learners face when using language. The quality of L2 speech and writing samples is often associated with how well words are used, which is a function of their written and spoken forms, grammatical functions, and collocations, as well as their form-meaning connections. Beatriz González-Fernández’s chapter in this volume builds nicely on earlier studies of vocabulary depth and recent studies that have focused on the relationship between exposure to different types of input and language learning (e.g. Lindgren & Muñoz, 2013). She looks at the extent to which 20 target words are known by L2 learners and the degree to which these participants are exposed to different types of L2 input outside of the classroom. She measures knowledge of concept and referents, collocation, word parts and form-meaning connection at two levels of test sensitivity: recall and recognition. González-Fernández surveys participants’ exposure to English language books, TV, videos, films, music, social networks and video games. This is a very interesting topic for research, particularly because researchers have often suggested that gains in many aspects of vocabulary knowledge are likely to occur primarily through encounters with input. Breadth of Vocabulary Knowledge

Research investigating breadth of L2 vocabulary knowledge has also been innovative. Perhaps the most important development in measures of breadth of L2 vocabulary knowledge was the creation of the Vocabulary Levels Test (VLT), which is the most widely used test of L2 vocabulary knowledge. The VLT was originally created by Nation (1983) for the purpose of determining the degree to which test takers had knowledge of different sets of words. In particular, there was great value in the first test level: knowledge of the most frequent 2000 word families. The most frequent 2000 word families account for a huge proportion of spoken and written English (Adolphs & Schmitt, 2003; Webb & Nation, 2017). Thus, evaluating knowledge of these words is extremely important; test scores at the 2000 word level revealed where emphasis should be placed in teaching vocabulary. The VLT also included 3000, 5000, 10,000 and University Word List (Xue & Nation, 1984) levels. Including a level designed to measure knowledge of academic vocabulary is an important tool for English for Academic Purposes programs. One limitation of the original VLT was that there were only 18 items at each level. Subsequent research showed

Innovation in Measures of Breadth and Depth of Vocabulary Knowledge  15

that this was insufficient to provide reliable evaluation of knowledge of 1000 word levels (Beglar & Hunt, 1999). Schmitt et al. (2001) improved on the earlier version of the VLT by increasing the number of items to 30 per level to provide a more reliable instrument. They also updated the word list used to select items to evaluate academic vocabulary knowledge from the University Word List (Xue & Nation, 1984) to the Academic Word List (Coxhead, 2000) and provided initial validation of two forms of the test. The most recent version of the VLT further improves on the design of the test. Webb et al. (2017) sampled items from Nation’s (2012) BNC/COCA word lists to provide a measure of the frequency of the items in the test that better reflects current English language. Earlier versions derived the lexical frequency of items from texts written in the 1930s and 1940s. Moreover, Webb, Sasao and Ballance’s version of the test measures knowledge of the 1000, 2000, 3000, 4000 and 5000 word levels. The addition of a test of the 1000 word level is particularly important because research shows that this level accounts for a much greater coverage of English than the 2000 word level (Webb & Nation, 2017). For example, Webb and Rodgers’ (2009) study of the vocabulary in television programs found that 85% of the words were from the 1000 word level and 4% of the words were from the 2000 word level. Including tests for both the 1000 and 2000 levels provides a much more useful diagnostic tool for teachers to pinpoint where vocabulary learning should be focused. One common misperception of the VLT is that it is a valid measure of vocabulary size. However, the VLT was not designed for this purpose and is unlikely to provide an accurate assessment of vocabulary breadth. The VLT does not measure knowledge of all word levels between 1000 and 10,000, nor does it measure knowledge of less frequent vocabulary beyond the 10,000 word level. More useful measures of vocabulary size are the New Computer-Adaptive Test of Size and Strength or CATSS (AviadLevitzky et al., 2019), AuralLex (Milton & Hopkins, 2006), Lex30 (Meara & Fitzpatrick, 2000), V_YesNo v1.0 and V_Size v2.01 (Meara & Miralpeix, 2017) and the Vocabulary Size Test (Coxhead et al., 2014; Nation & Beglar, 2007). It is important to also note that tests of vocabulary size are unlikely to provide a reliable evaluation of vocabulary levels. They were not designed for this purpose and do not have a sufficient number of items at a frequency level to do so (Gyllstad et al., 2015). Thus, they may have limited value for pedagogy. However, they should do a good job of differentiating between learners with different vocabulary sizes which can be useful for research and placement purposes. The development of the VLT allowed studies investigating the lexical profiles of texts and corpora to determine their appropriacy for language learners with different vocabulary sizes (e.g. Adolphs & Schmitt, 2003; Nation, 2006; Webb & Rodger, 2009). This line of research builds on studies investigating the lexical coverage necessary to understand spoken

16  Part 1: Theorising and Testing Vocabulary

(van Zeeland & Schmitt, 2013) and written text (e.g. Laufer, 1989; Schmitt et al., 2011). Lexical profiling studies tend to indicate that knowledge of the most frequent 2000 and 3000 word families provides 90% and 95% lexical coverage of most forms of spoken discourse such as conversation, television programs, and movies (Webb & Nation, 2017). Knowledge of the most frequent 4000 word families provides 95% coverage of academic spoken discourse (Dang & Webb, 2014), and knowledge of the most frequent 8000–9000 word families may provide 98% coverage of newspapers and novels (Nation, 2006). However, it is important to note that the vocabulary knowledge figures discussed in lexical profiling studies do not relate to vocabulary size test figures because these tests do not indicate the degree to which knowledge of words at different frequency levels is known. Scores on vocabulary size tests may indicate that test takers have a vocabulary size of n words, but they do not indicate the degree to which these words are high-, mid- and low-frequency words. Vocabulary size figures in lexical coverage studies refer more closely to mastery of frequency levels rather than vocabulary size. Thus, VLT scores are more relevant. However, currently, there are no tests that provide an accurate indication of the degree to which word frequency levels are known up until the 8000–9000 word frequency levels. This would be a useful area for further test development. Benjamin Kremmel’s chapter in this volume shows how differences in item placement can affect the behavior of a test. Kremmel compares two versions of a computer-adaptive test. One presented items by the frequency of the words in English, while the other presented items together in a range of frequencies. Kremmel examines how this difference in item presentation format contributes to measures of vocabulary knowledge at each of the different frequency level and test–retest reliability. There is great value in looking at the features within tests to optimize lexical assessment, and Kremmel’s study is a great example of this type of innovation in assessment. The chapter by Wen-Ta Tseng in this volume examines item exposure control in a computer-adaptive test. Tseng compares three item exposure rates: no control, .25 and .50 (.25 = one out of four test takers will receive the same item, .50 = one out of every two takers will receive the same item). The degree to which exposure to a test and its items affects performance on subsequent administrations of the test is always a concern in the validity of test performance, and so there is great value in looking at this feature in a carefully controlled computer-adaptive test. Future Research on Lexical Assessment

The following three chapters in this volume reveal useful directions for future research on lexical assessment. More broadly, there are two other features of lexical assessment that deserve attention in future studies:

Innovation in Measures of Breadth and Depth of Vocabulary Knowledge  17

innovation in the development of new tests and further validation of existing tests. There have been some important and innovative studies of vocabulary testing in recent years. Schmitt et al. (2001) and Webb et al. (2017) provide models of how an existing test might be reworked and improved upon. Kremmel and Tseng’s chapters in this volume along with AviadLevitzky et al. (2019) and Mizumoto et al. (2019) provide examples of how we might innovate through computerized-adaptive vocabulary testing. Several other excellent examples of innovation in test development are found in Meara and Miralpeix (2017) who provide multiple examples of research directions worth considering in vocabulary test development. Despite all of the research on lexical assessment, there are still many gaps in the research, particularly in the development of tests that may be used to evaluate vocabulary depth. For example, it would be useful to develop tests designed to measure knowledge of derivative forms of base words. Schmitt and Zimmerman (2002) provide two examples of possible test formats. It would also be useful to develop tests designed to measure knowledge of collocations. Gyllstad (2009) and Nguyen and Webb (2017) provide examples of useful collocation test formats. There would also be value in development of tests measuring knowledge of the different meanings of words. Knowledge of secondary and peripheral word meanings are rarely discussed in studies of vocabulary learning. This is an area where further research is clearly warranted. Hoshino (2015) outlined the advantages and disadvantages of several test formats that measure knowledge of multiple meanings. The greatest value in the development of tests of derivative forms, collocation and concept and referents would be to create them according to word frequency levels similar to how the VLT is structured. This would provide a useful contrast to VLT scores and also better reveal the degree to which knowledge of each aspect is developed. Scores for higher frequency items would typically reflect a greater strength of knowledge, while knowledge of lower frequency items would usually reflect less developed knowledge. There is also a need for further validation of existing tests. Test validation should be viewed as a continuing process to determine whether a test provides valid and reliable results for test takers with differing backgrounds and learning contexts (Fulcher & Davidson, 2007). There is limited validity evidence provided for most vocabulary tests and a need for follow-up research evaluating whether the test is working appropriately (Schmitt et al., 2020). Conclusion

I have learned a great deal from reading the work of Norbert Schmitt. One thing that I learned that is particularly relevant now is how to write a contribution for a Festschrift. I reviewed Pellicer-Sánchez and Schmitt (2010) for the special issue in Reading in a Foreign Language in honor of

18  Part 1: Theorising and Testing Vocabulary

Paul Nation. It was an excellent manuscript that was a pleasure to review. It did a great job of highlighting some of Paul Nation’s exceptional work and did a much better job of this than my contribution (sorry about that, Paul). Norbert Schmitt’s articles, chapters and books are informative, interesting and methodologically rigorous. They not only helped me learn about vocabulary studies, but they also provided great models for how to write manuscripts. I am grateful for his and his students’ excellent contributions. Norbert Schmitt’s students have also provided a great model of what a team of vocabulary researchers at one institution might hope to accomplish. Norbert’s students have achieved great success with their research, and he should be very proud. When attending top-tier conferences such as AAAL or EuroSLA, you could always be sure that his students would be well prepared and give great presentations. One of the highlights of these conferences has always been the times spent with Norbert, Diane and their students, and I look forward to this continuing on into the future with the addition of his grandstudents (i.e. his students’ students). References Adolphs, S. and Schmitt, N. (2003) Lexical coverage of spoken discourse. Applied Linguistics 24 (4), 425–438. Aviad-Levitzky, T., Laufer, B. and Goldstein, Z. (2019) The new computer adaptive test of size and strength (CATSS): Development and validation. Language Assessment Quarterly 16 (3), 345–368. Beglar, D. and Hunt, A. (1999) Revising and validating the 2000 word level and university word level vocabulary tests. Language Testing 16, 131–162. Chen, C. and Truscott, J. (2010) The effects of repetition and L1 lexicalization on incidental vocabulary acquisition. Applied Linguistics 31 (3), 693–713. Coxhead, A. (2000) A new academic word list. TESOL Quarterly 34 (2), 213–238. Coxhead, A., Nation, P. and Sim, D. (2014) Creating and trialling six versions of the Vocabulary Size Test. TESOLANZ Journal 22, 2213–2227. Dang, T.N.Y. and Webb, S. (2014) The lexical profile of academic spoken English. English for Specific Purposes 33 (1), 66–76. Fulcher, G. and Davidson, F. (2007) Language Testing and Assessment: An Advanced Resource Book. New York: Routledge. González-Fernández, B. and Schmitt, N. (2020) Word knowledge: Exploring the relationships and order of acquisition of vocabulary knowledge components. Applied Linguistics 41 (4), 481–505. https://doi.org/10.1093/applin/amy057 Gyllstad, H. (2009) Designing and evaluating tests of receptive collocation knowledge: COLLEX and COLLOMATCH. In A. Barfield and H. Gyllstad (eds) Researching Collocations in Another Language: Multiple Interpretations (pp. 153–170). New York: Palgrave Macmillan. Gyllstad, H., Vilkaitė, L. and Schmitt, N. (2015) Assessing vocabulary size through multiple-choice formats: Issues with guessing and sampling rates. ITL-International Journal of Applied Linguistics 166 (2), 278–306. Henriksen, B. (1999) Three dimensions of vocabulary development. Studies in Second Language Acquisition 21, 303–317. Hoshino, Y. (2015) Measuring knowledge of words with multiple meanings. Vocabulary Learning and Instruction 4 (1), 58–65.

Innovation in Measures of Breadth and Depth of Vocabulary Knowledge  19

Koizumi, R. and In’nami, Y. (2013) Vocabulary knowledge and speaking proficiency among second language learners from novice to intermediate levels. Journal of Language Teaching and Research 4 (5), 900–913. Kremmel, B. and Schmitt, N. (2016) Interpreting vocabulary test scores: What do various item formats tell us about learners’ ability to employ words? Language Assessment Quarterly 13 (4), 377–392. Laufer, B. (1989) What percentage of text lexis is essential for comprehension? In C. Laurén and M. Nordman (eds) Special Language: From Humans Thinking To Thinking Machines (pp. 316–323). Clevedon: Multilingual Matters. Lindgren, E. and Muñoz, C. (2013) The influence of exposure, parents, and linguistic distance on young European learners’ foreign language comprehension. International Journal of Multilingualism 10, 105–129. Meara, P. and Fitzpatrick, T. (2000) Lex30: An improved method of assessing productive vocabulary in an L2. System 28, 19–30. Meara, P. and Miralpeix, I. (2017) Tools for Researching Vocabulary. Bristol: Multilingual Matters. Miller, G.A. (1999) On knowing a word. Annual Review of Psychology 50, 1–19. Milton, J. and Hopkins, N. (2006) Comparing phonological and orthographic vocabulary size: Do vocabulary tests underestimate the knowledge of some learners? Canadian Modern Language Review 63 (1), 127–147. Mizumoto, A., Sasao, Y. and Webb, S. (2019) Developing and evaluating a computerized adaptive testing version of the Word Part Levels Test. Language Testing 36 (1), 101–123. Nation, I.S.P. (1983) Testing and teaching vocabulary. Guidelines 5 (1), 12–25. Nation, I.S.P. (1990) Teaching and Learning Vocabulary. New York: Heinle and Heinle. Nation, I.S.P. (2001) Learning Vocabulary in Another Language. Cambridge: Cambridge University Press. Nation, I.S.P. (2006) How large a vocabulary is needed for reading and listening? The Canadian Modern Language Review 63, 59–82. Nation, I.S.P. (2012) The BNC/COCA word family lists. See https://www.victoria.ac.nz/ lals/about/staff/paul-nation#vocab-programs (accessed June 2021) Nation, P. and Beglar, D. (2007) A vocabulary size test. The Language Teacher 31 (7), 9–13. Nguyen, T.M.H. and Webb, S. (2017) Examining second language receptive knowledge of collocation and factors that affect learning. Language Teaching Research 21 (3), 298–320. Pellicer-Sánchez, A. and Schmitt, N. (2010) Incidental vocabulary acquisition from an authentic novel: Do “Things Fall Apart”? Reading in a Foreign Language 22 (1), 31–55. Pigada, M. and Schmitt, N. (2006) Vocabulary acquisition from extensive reading: A case study. Reading in a Foreign Language 18 (1), 1–28. Read, J. (1993) The development of a new measure of L2 vocabulary knowledge. Language Testing 10 (3), 355–371. Read, J. (1998) Validating a test to measure depth of vocabulary knowledge. In A.J. Kunnan (ed.) Validation in Language Assessment (pp. 41–60). Mahwah, NJ: Erlbaum. Read, J. (2004) Plumbing the depths: How should the construct of vocabulary knowledge be defined? In P. Bogaards and B. Laufer (eds) Vocabulary in a Second Language: Selection, Acquisition and Testing (pp. 209–227). Amsterdam: John Benjamins. Richards, J.C. (1976) The role of vocabulary teaching. TESOL Quarterly 10, 77–89. Schmitt, N. (1998a) Tracking the incremental acquisition of second language vocabulary: A longitudinal study. Language Learning 48 (2), 281–317. Schmitt, N. (1998b) Quantifying word association responses: What is native-like? System 26 (3), 389–401.

20  Part 1: Theorising and Testing Vocabulary

Schmitt, N. (1998c) Measuring collocational knowledge: Key issues and an experimental assessment procedure. ITL-International Journal of Applied Linguistics 119 (1), 27–47. Schmitt, N. (2010) Researching Vocabulary: A Vocabulary Research Manual. Basingstoke: Palgrave MacMillan. Schmitt, N. (2014) Size and depth of vocabulary knowledge: What the research shows. Language Learning 64 (4), 913–951. Schmitt, N., Jiang, X. and Grabe, W. (2011) The percentage of words known in a text and reading comprehension. The Modern Language Journal 95 (1), 26–43. Schmitt, N. and Meara, P. (1997) Researching vocabulary through a word knowledge framework: Word associations and verbal suffixes. Studies in Second Language Acquisition 19 (1), 17–36. Schmitt, N., Nation, P. and Kremmel, B. (2020) Moving the field of vocabulary assessment forward: The need for more rigorous test development and validation. Language Teaching 53, 109–120. Schmitt, N., Schmitt, D. and Clapham, C. (2001) Developing and exploring the behaviour of two new versions of the Vocabulary Levels Test. Language Testing 18 (1), 55–88. Schmitt, N. and Zimmerman, C.B. (2002) Derivative word forms: What do learners know? TESOL Quarterly 36 (2), 145–171. van Zeeland, H. and Schmitt, N. (2013) Lexical coverage in L1 and L2 listening comprehension: The same or different from reading comprehension? Applied Linguistics 34 (4), 457–479. Webb, S. (2005) Receptive and productive vocabulary learning: The effects of reading and writing on word knowledge. Studies in Second Language Acquisition 27 (1), 33–52. Webb, S. (2007a) Learning word pairs and glossed sentences: The effects of a single sentence on vocabulary knowledge. Language Teaching Research 11 (1), 63–81. Webb, S. (2007b) The effects of repetition on vocabulary knowledge. Applied Linguistics 28 (1), 46–65. Webb, S. and Nation, I.S.P. (2017) How Vocabulary is Learned. Oxford: Oxford University Press. Webb, S. and Rodgers, M.P.H. (2009) The vocabulary demands of television programs. Language Learning 59 (2), 335–366. Webb, S., Sasao, Y. and Ballance, O. (2017) The updated Vocabulary Levels Test: Developing and validating two new forms of the VLT. ITL - International Journal of Applied Linguistics 168 (1), 34–70. Wesche, M. and Paribakht, T.S. (1996) Assessing second language vocabulary knowledge: Depth versus breadth. Canadian Modern Language Review 53 (1), 13–40. Xue, G. and Nation, I.S.P. (1984) A university word list. Language Learning and Communication 3, 215–229.

3 Vocabulary Knowledge: Lexical Depth and Its Relationship with Out-of-Class Exposure Beatriz González-Fernández

Vocabulary is a multifaceted construct that involves mastering various types of word knowledge, such as spelling, form-meaning mapping, collocation and derivatives (Nation, 2013). Achieving this depth of vocabulary mastery, although challenging, is deemed key for the successful and appropriate use of a second language (Schmitt, 2014). This significance of vocabulary depth for learners’ lexical development attracted Norbert Schmitt’s interest from the beginning of his academic career, when he published his 1998 study examining the productive knowledge of vocabulary components by second language (L2) learners of English. Schmitt’s 1998 foundational study set the basis for the exploration of depth of vocabulary knowledge in second languages. Nevertheless, the complexity of defining and operationalising this construct has meant that successive research has not typically focused on examining the extent of knowledge and the relationships between multiple vocabulary components in a systematic manner. Consequently, it remains unclear how well the various word-knowledge components1 are known by English-as-a-ForeignLanguage (EFL) learners and how they fit together in constructing vocabulary depth. In addition, it has been argued that learners’ engagement with and exposure to the second language (L2) might contribute to their vocabulary knowledge and development (Peters, 2018). Given the large vocabularies that learners require in order to use an L2 (Schmitt et al., 2017) as well as limited classroom time, researchers advocate that vocabulary learning should also occur and be promoted outside of the classroom (Puimège & Peters, 2019). However, there is a limited understanding about how different types of out-of-class exposure to English might influence the knowledge of the various vocabulary components. The present study addresses these gaps by comparing EFL learners’ extent of 21

22  Part 1: Theorising and Testing Vocabulary

knowledge of various vocabulary components, as well as the effect of different out-of-class exposure activities on depth of vocabulary knowledge. Conceptualising Vocabulary Knowledge

The great complexity of the vocabulary knowledge construct implies that it has been conceptualised in multiple and diverse ways. Most of the proposed conceptualisations are based on one of the earlier and broader descriptions of vocabulary, which understands it as the combination of size and depth of knowledge (Anderson & Freebody, 1981). Size refers to the quantity of words a person knows, typically at the form-meaning mapping level, while depth denotes the ‘quality’ or richness of that knowledge, that is, how well those words are known at other levels of mastery. Defining the degree to which a word is known is complex and ambiguous, and thus researchers have proposed multiple classifications of what exactly comprises depth of vocabulary knowledge (e.g. Cronbach, 1942; Richards, 1976). These conceptualisations on the nature of vocabulary knowledge led to the development of the widespread dimensions or components approach, which separates word knowledge into its multiple components. The most widely accepted framework of the components involved in word knowledge is Nation (2013). He proposes a list of nine word-knowledge aspects, each divided into receptive and productive mastery (see Figure 3.1). Nation’s framework provides the most comprehensive and precise description of the various components included in depth of vocabulary knowledge, and thus research investigating this issue has typically favoured his description to conceptualise vocabulary knowledge (e.g. González-Fernández & Schmitt, 2020; Pellicer-Sánchez, 2016; Schmitt, 1998; Webb, 2005, 2007).

Figure 3.1  Nation’s (2013: 49) framework of word-knowledge components

Vocabulary Knowledge: Lexical Depth and Out-of-Class Exposure  23

Due to the prevalence of the components approach in lexical studies, vocabulary knowledge is theoretically described, interpreted and accepted as a multidimensional construct. Yet, even Nation’s (2013) exhaustive and detailed description of overall vocabulary knowledge cannot explain the actual relationships between the various word-knowledge components and their contribution to the overall vocabulary knowledge construct. In order to shed some light on this issue, a recent study by GonzálezFernández and Schmitt (2020) examined the nature of L2 English vocabulary knowledge and the relationships between various components as described in Nation’s taxonomy. Employing Implicational Scaling and Structural Equation Modelling techniques, 2 they found that the wordknowledge components are so highly interrelated with each other that they do not behave as independent dimensions. Instead, the findings suggested that vocabulary knowledge acts as a unidimensional construct in L2 lexical use, where the various types of word knowledge constitute different degrees of difficulty of the same unique concept. This unidimensionality of vocabulary knowledge has also been empirically supported by Spencer et al. (2015) when assessing L1 English children’s knowledge of vocabulary size, associations, morphological knowledge and word use in context (i.e. grammatical functions, collocation and register). According to Spencer et al. (2015), what one knows about a word affects all other types of knowledge of that word, and thus, although they vary in their difficulty, the various vocabulary-knowledge components represent the same underlying unidimensional construct. These findings seem to oppose the most widely accepted theoretical conceptualisation of vocabulary as a multidimensional construct. Despite these recent findings on the structure of vocabulary knowledge, the great complexity of the vocabulary knowledge construct demands further investigation to untangle the nature of vocabulary knowledge and the relationships among its various components. There are only a few studies that have systematically examined various vocabularyknowledge components in order to investigate how they fit together, whether some components are known better than others, and, if so, in which pattern. The studies that have investigated this area are discussed in the next section. Depth of Vocabulary Knowledge

Previous studies have examined the knowledge of multiple vocabulary components concurrently, providing initial insights into learners’ vocabulary depth and the relationship among the word-knowledge components. One of the first attempts to address this issue was undertaken by Norbert Schmitt in 1998. He assessed EFL learners’ productive knowledge of four vocabulary components (i.e. written form, word parts [derivatives], associations and concepts and referents [multiple meanings]) over an academic year.

24  Part 1: Theorising and Testing Vocabulary

His results showed that these components were interrelated and developed in a parallel manner. However, he did not find any systematic pattern of knowledge which showed that some components were consistently better known by the learners than others. In a cross-sectional study, Shimamoto (2000) assessed knowledge of three vocabulary components: meaning recall, recognition of associations and recognition of collocations by a group of EFL learners. He found that associations were the best-known component among his participants, followed by collocation recognition, and finally meaning recall, and thus concluded that overall word-­ knowledge recognition seems to be higher than recall knowledge. This assumption was later corroborated by Pellicer-Sánchez and Schmitt (2010), who showed that written form recognition and meaning recognition were better known than grammatical function (i.e. word class) recall and meaning recall. In another cross-sectional study, Chui (2006) examined EFL learners’ depth of vocabulary knowledge by assessing four components: grammatical function (i.e. word class) recognition, meaning recall, collocation recognition and derivative form recall. She found that word class recognition was the best-known aspect, followed by meaning recall and collocation recognition, with her learners exhibiting the lowest scores on derivative form recall. Her findings suggest that aspects such as word class recognition and meaning recall might be easier, and thus potentially learnt earlier, than recognition knowledge of collocation and derivative recall, which seems to partially contradict the results above showing EFL learners’ general recognition knowledge as being higher than recall knowledge. The studies that have examined the greatest number of vocabularyknowledge components to date are Webb’s early studies from 2005 and 2007. In this research, he employed a comprehensive test battery to assess productive and receptive knowledge of five vocabulary components: written form, form-meaning mapping, syntagmatic associations (i.e. collocations), grammatical functions (i.e. word class) and paradigmatic associations. His findings show that written form tends to be the bestknown aspect both receptively and productively but that the extent of knowledge found for each component varies across the different studies. Hence, despite having employed the same test battery, his research has not found a consistent relationship or pattern of knowledge of the different vocabulary components. Only a recent study by González-Fernández and Schmitt (2020) has provided some preliminary evidence of a potential pattern of knowledge across multiple vocabulary components. Using an extensive test battery, the authors measured EFL learners’ knowledge of four word-knowledge components in recall and recognition mastery: form-meaning link, collocation, word parts (i.e. derivatives) and concepts and referents (i.e. multiple meanings). Their findings indicate that the various components

Vocabulary Knowledge: Lexical Depth and Out-of-Class Exposure  25

were very strongly interrelated with each other but that some of them were consistently better known than others. In particular, form-meaning link was found to be the best-known aspect, suggesting that it might develop earlier in the acquisition process, while knowledge of multiplemeanings and derivatives was the most difficult for EFL learners to master. The pattern also suggests that the recognition mastery of all components was better than the recall mastery of any aspect. Nevertheless, their study is limited to only one L1 population, and thus, it is unclear whether this pattern remains consistent with EFL learners of different L1 backgrounds. Taken together, this previous research, which builds on Schmitt’s (1998) study, suggests that some components of vocabulary knowledge can be expected to be better known than others. However, these studies present certain features that limit their conclusions about the interrelationships among vocabulary components. First, the studies have typically assessed different vocabulary components and employed different measures, and, as a result, their findings cannot be easily compared in order to specify the actual relationships between the multiple wordknowledge components. Even when the same vocabulary components have been assessed via the same measures (e.g. Webb, 2005, 2007), the results have failed to return a consistent pattern of knowledge across the various components. Finally, when the studies have been able to suggest a pattern of relationships among vocabulary components (e.g. Chui, 2006; González-Fernández & Schmitt, 2020), these have only been established for one learner population. Consequently, there is not enough evidence to ascertain how the various components of vocabulary knowledge fit together, whether there exists a consistent pattern of knowledge among them, and, if so, whether this pattern is generalisable to EFL learners in general. Such a pattern could begin to inform the development of a much-needed theory of vocabulary knowledge and acquisition in second languages. The Role of Out-of-Class L2 Exposure in Vocabulary Knowledge

It has been argued that exposure to L2 activities outside the classroom plays an important role in learners’ vocabulary development (PellicerSánchez, 2017; Peters, 2018). Among the specific exposure activities that researchers have advocated as having a beneficial effect on vocabulary knowledge and learning are: (extensive) reading, (extensive) TV and video viewing, playing computer games, social networking and listening to music. The positive effect of reading on vocabulary learning has been demonstrated by a number of empirical studies. For example, Pellicer-Sánchez and Schmitt (2010), in their investigation of incidental vocabulary learning through reading an authentic novel, found that EFL learners gained

26  Part 1: Theorising and Testing Vocabulary

knowledge of words at various levels of mastery. Similarly, Webb and Chang (2015) found significant vocabulary gains after an extensive reading treatment, both immediately (44%) and three months later (36.7%). This facilitative effect of reading on vocabulary learning has also been found in narrow reading treatments. Arndt and Woore (2018) examined 42 EFL learners’ acquisition of spelling, meaning and grammatical functions while incidentally reading three short online texts on a related topic (1691 words in total) and found immediate moderate gains in all aspects (65.9% overall), particularly in meaning recognition and spelling. Comparatively, fewer studies have investigated the relationship between more casual out-of-class reading in non-controlled study conditions and vocabulary knowledge. Some exceptions are González Fernández and Schmitt (2015), Macis and Schmitt (2017) and Peters (2018). These three studies found significant correlations between casual engagement in outof-class reading and learners’ vocabulary knowledge, suggesting that this less formal and less prescribed reading exposure also contributes to L2 vocabulary learning. Video and TV viewing has also been proposed as an input activity that can foster vocabulary development, and its examination has experienced an increased interest in recent years, with a special issue dedicated to this topic in The Language Learning Journal in 2019. Indeed, the studies in this area have found a facilitative effect of viewing TV and watching films on vocabulary learning. After viewing 13 episodes of a TV programme, Rodgers (2013) found moderate vocabulary gains at the form-meaning receptive level. Arndt and Woore (2018) explored vocabulary learning under a narrower-viewing situation, where 38 EFL learners viewed 3 short videos (each 2–5 minutes long) on the same topic. They found immediate vocabulary learning gains after the viewing session (69.2% overall), mainly in meaning recognition and word class recognition. More recently, Puimège and Peters (2019) found that both single words and formulaic language were learnt at the form recall, form recognition and meaning recall levels after viewing just 30 minutes of a TV programme. These two studies together suggest that the beneficial effects of video/TV-viewing on vocabulary learning remain even during limited exposure to the L2. As an out-of-class input activity, research has found that viewing films and TV is a popular activity among EFL learners (Peters, 2018; Sylvén & Sundqvist, 2012) and that it is positively correlated with vocabulary knowledge (González Fernández & Schmitt, 2015; Kuppens, 2010; Peters, 2018), which indicates its potential as a means for incidental vocabulary learning outside the classroom. The relationship between vocabulary knowledge and playing computer games has also been a topic of research interest in the past few years (e.g. Sundqvist, 2019; Sundqvist & Wikström, 2015; Sylvén & Sundqvist, 2012). Sundqvist and Wikström (2015) found that frequent computer gamers (>5 hours/week) had higher scores on several vocabulary size and

Vocabulary Knowledge: Lexical Depth and Out-of-Class Exposure  27

use measures than moderate learners ( 6) and significantly more moderately scoring combinations (6 < t > 2) than intermediate students. They also found that very low-scoring combinations (t < 2) were more common in intermediate than in advanced learner writing. Bestgen and Granger (2014) studied the mean t-score for all bigrams in L2 writing but did not find a significant relationship with proficiency. Delta-P has shown stronger results. Kim et al. (2018) found a significant positive correlation between Delta-P of adjacent bigrams and both overall writing proficiency and lexical proficiency scores for L2 writers. Similarly, Durrant et al. (2019) found a significant correlation between the same measure and scores awarded to US freshman composition assignments. In sum, measures of ‘phraseological sophistication’ do appear to be useful in tracing L2 writing development. More advanced L2 writers use fewer combinations which are unattested or infrequent in a relevant reference corpus. This seems to contrast with university-level writing in L1-dominant US university contexts, where greater use of unattested ­combinations or lower-frequency combinations is associated with higher grades, suggesting a possible dissociation between what is valued in L1 and L2 writing. In both contexts, however, mutual information and

104  Part 2: Formulaic Language

Delta-P are associated with higher grades. While there is some evidence that t-score shows a negative correlation with L2 proficiency, findings are not conclusive. The current study aims to move this work forward by studying the development in collocations in children’s writing in an L1 English environment, a context which has not previously been studied. Given the differences that have already emerged between L2 and mature L1 writing, we might expect this context to differ from those of previous research in interesting ways. Specifically, it addresses the following research questions: (1) To what extent do quantitative measures of collocation use capture changes in the written language of school children in England as they progress through the statutory education system? (2) How do developmental patterns in collocation use interact with the genres in which children are writing? Methodology Study corpus

This study is based on a corpus of writing produced by children in schools in England as part of their regular schoolwork. Texts were sampled at the ends of each of the four ‘Key Stages’ of the compulsory English school system (i.e. Years 2, 6, 9 and 11) from classes in English, Science and the Humanities (i.e. History, Geography and Religious Studies). They were classified for genre on the basis of a two-way distinction between ‘literary’ and ‘non-literary’ tasks. We define a ‘literary’ text as one which can be evaluated as successful or unsuccessful without considering any kind of propositional or directive relationship to the world. The primary purpose of a literary text is to be appreciated on its own terms as a piece of stylised writing. Prototypical examples were creative fiction and literary imitations (where students were asked to produce a text mimicking a particular kind of fiction, such as the Gothic novel). ‘Non-literary’ texts, in contrast, do need to bear a propositional or directive relationship to the world in order to be considered successful. The primary purpose of non-literary texts is to (a) accurately depict a particular state-of-affairs, (b) evaluate a particular state-of-affairs or (c) argue for a particular state-of-affairs to be the case. Prototypical non-literary texts included historical accounts, literary criticism and experimental reports. Procedures for collecting and encoding texts are described in more detail on the project website.1 • The full corpus comprises 2898 texts. For the present study, certain texts were excluded. Specifically: • Year 2 texts. These tended to have markedly fewer syntactic combinations of the sort studied here than those produced at higher years, making comparison problematic.

The Development of Academic Collocations in Children’s Writing  105

• Texts that did not constitute continuous prose (e.g. labelled diagrams, sentence exercises, poetry). • Texts that were co-authored. • Texts that were shorter than 100 words. • Texts that had a high proportion of illegible words; specifically, any texts with more than 10% illegible words. The makeup of the resulting corpus is shown in Tables 7.1a-b. Table 7.1a  Corpus makeup – distribution of texts across year groups, genres and disciplines English Year 6

Literary

238

Non-literary

242

Year 9

Literary

217

Non-literary

270

Year 11

Literary Non-literary

Discipline Totals

Humanities

Science

Genre × Year totals 238

93

159

494 217

103

74

63

447 63

366

47

39

452

1396

243

272

1911

Reference corpus

As a reference corpus, we selected the British Academic Written English corpus (BAWE). 2 The BAWE corpus comprises 2761 texts written by bachelors- and masters-level students at four British universities across a range of subject areas. All assignments in the corpus had received at least an ‘upper-second class’ or ‘merit’ grade and so can be deemed examples of successful student writing. The BAWE corpus was selected as representing ‘target language’ use for children’s non-literary writing. Although not all children progress to university study, successful university writing can be seen as a major goal towards which work throughout a child’s school career is ultimately preparing them. Importantly, while the BAWE corpus can be seen as representing a relevant developmental target for non-literary writing, it cannot be seen as such for their literary writing. It therefore also provides a useful means of understanding how children’s genres are differentiated by their use of collocation. We worked with the UTF-8 version of the BAWE corpus, which is tagged in accordance with TEI5 guidelines (TEI Consortium, 2014). As sentences are not reliably marked with full-stops, the tags in this version of the corpus were used to divide each text into sentences. TEI tags were also used to identify and remove lists, abstracts and other pre-body front-matter, block quotations, footnotes, endnotes and formulas/equations.

1911

39

Overall

47

Science Non-Literary

366

63

74

Humanities Non-Literary

English Non-Literary

English Literary

Science Non-Literary

7

23

2

1

6

4

3

3

270

103

English Non-Literary

Humanities Non-Literary

4

4

4

159

6 7

217

English Literary

Science Non-Literary

93

238

242

English Literary

English Non-Literary

Humanities Non-Literary

Max

171

767

34

19

110

45

41

40

162

217

87

78

132

25

225

13

8

60

8

24

16

27

6

9

11

20

384

282

332

244

345.5

357

203

264

304.5

362

212

203

188

100

100

101

124

100

133

102

102

100

109

100

100

100

2494

2494

945

1465

1222

1096

682

1075

2114

534

737

590

1087

10

12

11

12

17

11.5

9

9.5

12

7

8

6

12

a

The corpus includes multiple texts written in response to a single class task. ‘Title’ refers to the name we assigned to each task.

Year 11

Year 9

Year 6

Min

11

17

11

14

12

9

11

11

12

9

6

7

14

Verb–Object

Adjective–Noun

Titlesa

Median

Writers

Texts

Schools

Median combinations per text

Words per text

Number of

Table 7.1b  Corpus makeup – contributors, word counts, combination counts

21

30

23

26

30

20

20

21

24

16

14

13

25.5

All

106  Part 2: Formulaic Language

The Development of Academic Collocations in Children’s Writing  107

Analysis

Both the study corpus and the reference corpus were tokenised, lemmatised, tagged for part-of-speech and parsed for syntactic dependencies using the Stanford CoreNLP pipeline (Manning et al., 2014). A custombuilt R script was then used to extract word combinations from each corpus. Specifically, we extracted all lemmatised examples of: • adjectives with a modifying (amod) dependency on a common noun, e.g. high*level; important*role • common nouns with an object (iobj/dobj) dependency on a verb, e.g. have*effect; play*role These combination types were chosen because they are both common at all levels of the study corpus, so allowing robust analyses, and because they have been a frequent focus of previous research (e.g. Granger & Bestgen, 2014; Paquot, 2018, 2019), so allowing for comparison with existing literature. As part of a wider study, a sample of 180 texts was hand-coded for syntactic dependencies within noun phrases. This included marking the external dependency relation of each noun and all internal dependencies within the noun phrase. Each text was annotated independently by two trained raters. Interrater reliability was high, with mean correlation between raters for part-of-speech tag of r = 0.94 and for dependency relationship of r = 0.89. All discrepancies between raters were identified and adjudicated by this chapter’s second author. These hand-coded texts were used to check the reliability of the Stanford parser, both in terms of recall (i.e. the percentage of true cases which were identified) and precision (i.e. the percentage of identified items which were true cases). As can be seen from Table 7.2, precision and recall were high across all parts of the Table 7.2  Reliability of automated identification of syntactic relations amod N Year 6

obj recall

precision

N

recall

precision

Literary

278

86%

71%

522

18%

88%

Non-literary

206

84%

80%

262

32%

80%

Literary

359

83%

83%

412

21%

74%

Non-literary

305

86%

75%

370

30%

75%

Year 11

Literary

458

85%

81%

539

16%

76%

All year groups

Literary

Year 9

Non-literary

Non-literary All texts

313

88%

75%

467

40%

75%

1219

85%

82%

1643

17%

79%

921

85%

77%

1131

34%

75%

2140

85%

80%

2774

24%

76%

108  Part 2: Formulaic Language

corpus for the amod (adjective-noun premodification) dependency. For the iobj/dobj dependencies (verb-object structures), precision was high, but recall was relatively poor. In the following analyses, therefore, it should be borne in mind that the verb-object structures retrieved will have been only a subset of those which were actually present. Items from the reference corpus were tabulated so as to record, for each combination: frequency; MI; MI2; t-score; Delta-P (for a discussion of the calculation and use of these measures, see Gries & Durrant, 2021). Delta-P was measured in both directions. Henceforth, the strength of association from dependent word to the word it is dependent on (i.e. from subject noun to verb or from adjective to noun) will be labelled Delta-P Coll-Node. Delta-P in the opposite direction will be labelled Delta-P Node-Coll. Association measures are known to be unreliable for lowfrequency items so were recorded only for items occurring at least five times in the reference corpus. Using this dataset, frequency and association measures were retrieved from the reference corpus for each relevant combination in each text. Table 7.3 shows an abbreviated example of the resultant data for a Year 11 non-literary text. Using these data, the means of each of the frequency and association measures were calculated separately for each text. These means served as the basis for the analysis of variation across year groups and genres. Based on these data, three main analyses were conducted: (1) mean reference-corpus frequency of combinations across year × genre groups, i.e. the extent to which the combinations used in texts are also frequently attested combinations in the reference corpus; (2) mean reference-corpus association measures of combinations, i.e. the extent to which the combinations used are strongly associated word pairings in the reference corpus; (3) the proportion of combinations in each text which qualified as ‘collocations’ in the reference corpus, in the sense that they appeared with a frequency of more than one per million words and a MI of more than five. Extensive piloting showed that the results for the key variables of year group and genre were similar regardless of whether collocations were lemmatised or not and of whether they were based on type or token counts. For simplicity of presentation, and to avoid inflated error rates, the Table 7.3  Example of word combination data Item Teenage tearaway Considerable bias Own opinion Many effect Bad behaviour

Frequency

MI

MI2

t-score

Delta-P Coll-Node

Delta-P Node-Coll

0

NA

NA

NA

NA

NA

0

NA

NA

NA

NA

NA

3.4738

0.0036

0.0132

13

4.7747

8.4751

5

0.1166

2.4386

0.1737

0.0001

0.0001

92

6.2737

12.7973

9.4677

0.0517

0.0192

The Development of Academic Collocations in Children’s Writing  109

analyses that follow will therefore focus only on lemmatised collocation tokens. Like many corpora, the texts that form the data points in our analyses are not independent: for example, multiple texts are written by individual writers, and multiple writers are sampled from individual schools. As Gries (2015) has argued, data of this sort violate the assumption of independence on which standard statistical methods are based. Separate texts written by a single writer or to a single title, or produced within a single school or subject area, are clearly more closely related to each other than they are to those produced by another writer, to another title, or in another school or subject area. Moreover, it is plausible that each of these grouping variables (i.e. writers, titles, disciplines, schools) has the potential to exert its own influence on vocabulary use. To address these issues, our analyses used mixed-effects models. For each analysis, we adopted the three-stage stepwise procedure detailed in Gries (2015) and outlined below. 3 Stage 1 identified the maximal fixed effects structure and the maximal random effects structure of interest. For all analyses, the maximal fixed effects structure comprised the main effects of year group and genre plus their interaction. The maximal random effects structure comprised schools; disciplines; writers as nested within schools and titles as nested within disciplines. The two nested structures are crossed because individual titles were written by multiple writers, whilst individual writers wrote on multiple titles. Titles also cut across schools as students from multiple schools wrote on common titles, reflecting the influence of a national curriculum with shared public examinations. For Stage 2, we combined the maximal fixed effects structure with the maximal random effects structure. We then determined the optimal random effects structure relative to this combination by (a) removing each random effect in turn and (b) comparing the overall quality of the model when the effect is present versus when it is absent. In each case, particular random effects were retained only if their removal made the model quality significantly worse; otherwise, the effect was eliminated from the final Stage 2 model altogether. For Stage 3, we determined the optimal fixed-effects structure relative to the optimal random effects structure identified in Stage 2. This involved sequentially removing any fixed effects which were neither significant in themselves nor participated in any higher order interactions. As with the Stage 2 procedure, a particular fixed effect was retained only if removing it made the model quality significantly worse; otherwise, the effect was eliminated in order to derive the final models reported below. In both stages, model quality was determined with reference to the Akaike information criterion (AIC) score for each model iteration. Mixed-effects models need to meet certain assumptions to be accurate and generalisable (Tabachnick & Fidell, 2014; Zuur et al., 2009). These

110  Part 2: Formulaic Language

were checked as follows: histograms of residuals were checked to identify significant outliers; residuals vs. observed values were checked to confirm the linearity of the data; Q-Q plots were checked to confirm the normal distribution of residuals and random effects; plots of standardised residuals vs. fitted values were checked to confirm homoscedacity of residuals. All analyses met the necessary assumptions. Findings Quantitative variation across genres and year groups

The first of our main analyses is based on the mean reference-corpus frequency of combinations in each text. Figure 7.1 shows how this varies across year groups and genres for each combination type. As the results for the two combination types are in close parallel, the data from both types were combined to produce a single mixed-effects model, shown in Table 7.4. As we would expect, given the academic nature of the reference corpus, non-literary texts had substantially, and significantly, higher means than literary texts. The measure also increased significantly across year groups. Figure 7.1 appears to show that development across year groups is restricted to non-literary writing, suggesting that children are learning to discriminate more clearly between the two genres. Although pilot analyses showed this year × genre interaction to be highly significant in a linear model, it narrowly missed significance in our mixed-effects model (p = 0.07). More research is therefore needed to determine whether the interaction is generalisable. It is also notable that, though the conditional R 2 for the mixed-effects model is high (0.34), the portion of this attributable to fixed effects (0.09) is relatively modest. This suggests that

Figure 7.1  Mean combination frequency

The Development of Academic Collocations in Children’s Writing  111

Table 7.4  Mixed effects model for mean log frequency for both combination types

Value Intercept

Std. Error

DF

t-value

p value

−6.2622

3.6833

136.66

−1.700

0.0913

Year group

1.5131

.4091

130.70

3.699

0.0003

Genre

9.7896

123.62

4.492