<?page no="0"?> Saskia Kersten The Mental Lexicon and Vocabulary Learning Implications for the foreign language classroom Language in Performance LiP <?page no="1"?> The Mental Lexicon and Vocabulary Learning <?page no="2"?> 43 Edited by Werner Hüllen† and Rainer Schulze Advisory Board: Thomas Herbst (Erlangen), Andreas Jucker (Zürich), Manfred Krug (Bamberg), Christian Mair (Freiburg i. Br.), Ute Römer (Hannover), Andrea Sand (Trier), Hans-Jörg Schmid (München), Josef Schmied (Chemnitz) and Edgar W. Schneider (Regensburg) <?page no="3"?> Saskia Kersten The Mental Lexicon and Vocabulary Learning Implications for the foreign language classroom <?page no="4"?> Bibliografische Information der Deutschen Nationalbibliothek Die Deutsche Nationalbibliothek verzeichnet diese Publikation in der Deutschen Nationalbibliografie; detaillierte bibliografische Daten sind im Internet über <http: / / dnb.d-nb.de> abrufbar. Hil 2 © 2010 · Narr Francke Attempto Verlag GmbH + Co. KG Dischingerweg 5 · D-72070 Tübingen Das Werk einschließlich aller seiner Teile ist urheberrechtlich geschützt. Jede Verwertung außerhalb der engen Grenzen des Urheberrechtsgesetzes ist ohne Zustimmung des Verlages unzulässig und strafbar. Das gilt insbesondere für Vervielfältigungen, Übersetzungen, Mikroverfilmungen und die Einspeicherung und Verarbeitung in elektronischen Systemen. Gedruckt auf säurefreiem und alterungsbeständigem Werkdruckpapier. Internet: www.narr.de E-Mail: info@narr.de Printed in Germany ISSN 0939-9399 ISBN 978-3-8233-6586-0 <?page no="5"?> Contents 1 Introduction 1 The Mental Lexicon 5 2 The L1 and L2 Mental Lexicon 7 2.1 Defining the Content of the Mental Lexicon . . . . . . . . . . 8 2.2 Methodological Approaches . . . . . . . . . . . . . . . . . . . 9 2.2.1 Psycholinguistic Methods . . . . . . . . . . . . . . . . . . . . 9 2.2.2 Neurolinguistic Methods . . . . . . . . . . . . . . . . . . . . . 11 2.3 Modelling the Mental Lexicon . . . . . . . . . . . . . . . . . . 13 2.3.1 Models of Lexical Processing . . . . . . . . . . . . . . . . . . 13 2.3.2 Modelling the Bilingual Mental Lexicon . . . . . . . . . . . . 20 2.4 Internal Organisation of the Bilingual Mental Lexicon . . . . . . . . . . . . . . . . . . . . . . . . . . 23 2.4.1 L1 and L2 Lexicon - Same or Di fferent? . . . . . . . . . . . . . 29 2.4.2 The Influence of Proficiency . . . . . . . . . . . . . . . . . . . 32 2.5 The Neurocognitive Foundation . . . . . . . . . . . . . . . . . 33 2.6 Interim Summary . . . . . . . . . . . . . . . . . . . . . . . . . 34 3 Dynamic Systems Theory 35 3.1 Growth in Dynamic Systems Theory . . . . . . . . . . . . . . 36 3.2 Dynamic Systems Theory and Second Language Acquisition . . . . . . . . . . . . . . . . . . . . . . 37 3.3 Interim Summary . . . . . . . . . . . . . . . . . . . . . . . . . 38 4 Cognitive Linguistics and Foreign Language Teaching 39 4.1 Defining Cognitive Linguistics . . . . . . . . . . . . . . . . . 39 v <?page no="6"?> 4.2 Implications of Cognitive Linguistics for Vocabulary Learning . . . . . . . . . . . . . . . . . . . . . . . 42 4.3 Interim Summary . . . . . . . . . . . . . . . . . . . . . . . . . 46 Vocabulary Learning and Teaching 47 5 Current Issues in Vocabulary Research 49 5.1 What is it to Know a Word? . . . . . . . . . . . . . . . . . . . 50 5.1.1 What is a Word? . . . . . . . . . . . . . . . . . . . . . . . . . . 50 5.1.2 What is Involved in Knowing a Word . . . . . . . . . . . . . 52 5.2 Breadth and Depth of Vocabulary Knowledge . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54 5.3 Active vs Passive Vocabulary Knowledge . . . . . . . . . . . 55 5.4 Interim Summary . . . . . . . . . . . . . . . . . . . . . . . . . 56 6 Learning New Vocabulary 57 6.1 The Basis for Learning . . . . . . . . . . . . . . . . . . . . . . 58 6.1.1 Levels of Memory . . . . . . . . . . . . . . . . . . . . . . . . . 58 6.1.2 Long-Term Memory . . . . . . . . . . . . . . . . . . . . . . . . 60 6.2 Language Learning . . . . . . . . . . . . . . . . . . . . . . . . 62 6.2.1 Vocabulary in Language Learning . . . . . . . . . . . . . . . 63 6.2.2 Depth of Processing Hypothesis . . . . . . . . . . . . . . . . . 64 6.2.3 Incidental Vocabulary Learning . . . . . . . . . . . . . . . . . 66 6.2.4 Implicit versus Explicit . . . . . . . . . . . . . . . . . . . . . . 68 6.2.5 Item versus System Knowledge . . . . . . . . . . . . . . . . . 69 6.2.6 Rote Learning . . . . . . . . . . . . . . . . . . . . . . . . . . . 70 6.3 General Processes which aid Vocabulary Retention . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71 6.3.1 Noticing, Retrieving & Using Vocabulary . . . . . . . . . . . 71 6.3.2 Involvement . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73 6.4 Interim Summary . . . . . . . . . . . . . . . . . . . . . . . . . 74 7 Teaching Vocabulary 75 7.1 The Beginner ’s Paradox . . . . . . . . . . . . . . . . . . . . . 75 7.2 Presenting New Vocabulary . . . . . . . . . . . . . . . . . . . 76 7.3 Teaching to Promote Long-Term Vocabulary Retention . . . 78 7.3.1 Repeated Exposure to an Item . . . . . . . . . . . . . . . . . . 78 7.3.2 Richness Activities . . . . . . . . . . . . . . . . . . . . . . . . 78 7.3.3 Information Gap Activities . . . . . . . . . . . . . . . . . . . . 79 7.3.4 Structuring New Vocabulary . . . . . . . . . . . . . . . . . . . 80 vi <?page no="7"?> 7.3.5 Integration of Old and New Knowledge . . . . . . . . . . . . 80 7.4 Interim Summary . . . . . . . . . . . . . . . . . . . . . . . . . 81 8 SLA and the Young Learner 83 8.1 Knowing a Word in Primary School . . . . . . . . . . . . . . . 83 8.1.1 Games in Early Second Language Acquisition . . . . . . . . . 85 8.1.2 Tasks in Foreign Language Learning . . . . . . . . . . . . . . 86 8.1.3 Sca ffolding . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87 8.1.4 The Importance of Chunks . . . . . . . . . . . . . . . . . . . . 88 8.1.5 The Role of Writing . . . . . . . . . . . . . . . . . . . . . . . . 90 8.2 Enhancing Vocabulary Activities in Primary School . . . . . 91 8.3 Interim Summary . . . . . . . . . . . . . . . . . . . . . . . . . 92 9 Measuring Vocabulary Knowledge 93 9.1 Developing Assessment Tasks for Younger Learners . . . . . 94 9.2 Types of Vocabulary Tests . . . . . . . . . . . . . . . . . . . . 95 9.2.1 Selected-response Tasks . . . . . . . . . . . . . . . . . . . . . 97 9.3 Interim Summary . . . . . . . . . . . . . . . . . . . . . . . . . 99 Intervention Study in Primary School Classrooms 101 10 From Theory to Practice 103 10.1 Implications for the Classroom . . . . . . . . . . . . . . . . . 103 10.2 Choosing Vocabulary Items . . . . . . . . . . . . . . . . . . . 104 10.2.1 Vocabulary Items Chosen for Grade 3 . . . . . . . . . . . . . 105 10.2.2 Vocabulary Items Chosen for Grade 4 . . . . . . . . . . . . . 106 11 Outline of the Study 107 11.1 Procedure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107 11.1.1 Example Lesson Plan - Grade 3, 2nd Lesson . . . . . . . . . . 110 11.1.2 Example Lesson Plan - Grade 3, 5th Lesson . . . . . . . . . . 113 11.1.3 Summary of the Di fferences between the Groups . . . . . . . 116 11.2 Participants . . . . . . . . . . . . . . . . . . . . . . . . . . . . 117 11.3 Test Instruments Used . . . . . . . . . . . . . . . . . . . . . . 119 11.4 Questionnaires . . . . . . . . . . . . . . . . . . . . . . . . . . . 121 11.5 Exploring the Data for Grade 3 . . . . . . . . . . . . . . . . . 125 11.5.1 Test 1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 125 11.5.2 Test 2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 127 11.5.3 Comparing Test 1 and Test 2 in Grade 3 . . . . . . . . . . . . 129 11.5.4 Chi-Square Tests . . . . . . . . . . . . . . . . . . . . . . . . . . 129 vii <?page no="8"?> 11.6 Exploring the Data for Grade 4 . . . . . . . . . . . . . . . . . 132 11.6.1 Test 1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 132 11.6.2 Test 2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 134 11.6.3 Comparing Test 1 and Test 2 . . . . . . . . . . . . . . . . . . . 136 11.6.4 Chi-Square Tests . . . . . . . . . . . . . . . . . . . . . . . . . . 136 11.7 Discussion of Results . . . . . . . . . . . . . . . . . . . . . . . 137 11.8 Discussion of Classroom Observation . . . . . . . . . . . . . 138 12 Conclusion 141 Bibliography 144 Sources of Illustrations 158 Appendix 169 A Lesson plans 171 viii <?page no="9"?> List of Figures 2.1 Levelt’s Blueprint for Speech Production and Comprehension (De Bot et al. 2005, 40) . . . . . . . . . . . . . . . . . . . . 15 2.2 A Spreading Activation Model (Randall 2007, 115) . . . . . . 19 2.3 Lexical Comprehension / Production Model for Oral and Written Modalities (De Bot et al. 1997, 315) . . . . . . . . . . . . . 22 2.4 A Network Illustrating a Bilingual Spreading Activation Model (see also Randall 2007, 116) . . . . . . . . . . . . . . . . . . . . 24 2.5 Possible Types Organisation of the Mental Lexicon (adapted from Obler & Gjerlow 1999, 129) . . . . . . . . . . . . . . . . . 25 2.6 Schematic Representation of Various Models of the Organisation of the Mental Lexicon (Paradis 2004, 111) . . . . . . . . 27 2.7 Revised Hierarchical Model (Kroll 1993, 54) . . . . . . . . . . 28 4.1 A Conceptual Map of Cognitive Linguistics (Geeraerts 2006, 19) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40 6.1 A Model of Working Memory (Lefrançois 2006, 305) . . . . . 60 6.2 Tasks Involved in Learning the Meaning of Words (adapted from Aitchison 2003, 189) . . . . . . . . . . . . . . . . . . . . . 62 11.1 Worksheets for Intervention Group (left) and Control Group (right) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113 11.2 Worksheets for Intervention Group (left) and Control Group (right) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 114 11.3 Test used in Grade 3 . . . . . . . . . . . . . . . . . . . . . . . . 120 11.4 Test used in Grade 4 . . . . . . . . . . . . . . . . . . . . . . . . 121 11.5 Questionnaires 1 (bottom) and 2 (top) for Grade 3 . . . . . . . 122 11.6 Questionnaire 1 for Grade 4 . . . . . . . . . . . . . . . . . . . 123 11.7 Questionnaire 2 for Grade 4 . . . . . . . . . . . . . . . . . . . 124 ix <?page no="10"?> 11.8 Score Frequencies for Test 1 in Grade 3 . . . . . . . . . . . . . 126 11.9 Percentages for Test 1 in Grade 3 . . . . . . . . . . . . . . . . . 126 11.10 Score Frequencies for Test 2 in Grade 3 . . . . . . . . . . . . . 128 11.11 Percentages for Test 2 in Grade 3 . . . . . . . . . . . . . . . . . 128 11.12 Score Frequencies for Test 1 in Grade 4 . . . . . . . . . . . . . 133 11.13 Percentages for Test 1 in Grade 4 . . . . . . . . . . . . . . . . . 133 11.14 Score Frequencies for Test 2 in Grade 4 . . . . . . . . . . . . . 135 11.15 Percentages for Test 2 in Grade 4 . . . . . . . . . . . . . . . . . 135 x <?page no="11"?> List of Tables 5.1 What is Involved in Knowing a Word (Nation 2001, 23) . . . 53 6.1 Three Levels of Memory (Lefrançois 2006, 309) . . . . . . . . 59 6.2 Ways of Learning and Teaching Vocabulary (Nation 2001, 16) 64 11.1 Overview of Di fferences between Groups in Grade 3 . . . . . 116 11.2 Overview of Di fferences between the Groups in Grade 4 . . . 117 11.3 How the Groups were Coded . . . . . . . . . . . . . . . . . . 117 11.4 N for Each Class - Test 1 . . . . . . . . . . . . . . . . . . . . . . 118 11.5 N for Each Class - Test 2 . . . . . . . . . . . . . . . . . . . . . . 118 11.6 Pooled N for Each Group . . . . . . . . . . . . . . . . . . . . . 119 11.7 Mean Results for Test 1 in Grade 3 . . . . . . . . . . . . . . . . 125 11.8 Mean Results for Test 2 in Grade 3 . . . . . . . . . . . . . . . . 127 11.9 Comparison of Mean Results in Grade 3 . . . . . . . . . . . . 129 11.10 Vocabulary Knowledge Self-Evaluation for Test 1 . . . . . . . 130 11.11 Vocabulary Knowledge Self-Evaluation for Test 2 . . . . . . . 130 11.12 Games Rating Crosstabulation . . . . . . . . . . . . . . . . . . 131 11.13 Mean Results for Test 1 in Grade 4 . . . . . . . . . . . . . . . . 132 11.14 Mean Results for Test 2 in Grade 4 . . . . . . . . . . . . . . . . 134 11.15 Comparison of Mean Results in Grade 4 . . . . . . . . . . . . 136 xi <?page no="13"?> List of Abbreviations BIA Bilingual Interactive Activation CEF Common European Framework of Reference for Languages CL Cognitive Linguistics DST Dynamic Systems Theory EEG electroencephalography EFL Early Foreign Language EFLT Early Foreign Language Teaching ERP event-related potential FLA first language acquisition FLT foreign language teaching fMRI functional magnetic resonance imaging IAM Interactive Activation Model K-S test Kolmogorov-Smirnov test LL learners L1 native language L2 second or foreign language LTM long-term memory PET positron emission tomography SLA second language acquisition STM short-term memory T teacher TOT tip of the tongue xiii <?page no="15"?> Acknowledgements This book would not have been possible without the help of many people who encouraged me along the way and were always quick to o ffer advice, helpful comments, ideas and the occasional piece of chocolate to keep me going. It is based on a Ph.D. doctorate for Hildesheim University. I would like to thank my supervisor, Prof. Dr. Friedrich Lenz, for the inspiration for this study and his support and encouragement along the way. I am greatly indebted to all my colleagues, in particular to Janet McLaughlin, for greatly improving my English and for giving me valuable feedback during the final stages of writing this book, and to Carsten John, for providing much needed help with the methodology and statistics for the empirical research. I learned a lot from both of them. The empirical research would not have been possible without the support and participation of the schools, the learners and their teachers, who made room in their busy schedule for me to test my ideas in real life. Thank you for your time and commitment. Thanks also to Jana Winnefeld and Anika Wüstefeld for the excellent teaching during the intervention and for challenging my ideas about teaching from a professional perspective. I am also greatly indebted to the editor of Language in Performance, Prof. Dr. Rainer Schulze, who made valuable comments on the final draft of this book. And finally, thank you to Thomas Epping for his continual love and support. This is for you. <?page no="17"?> Chapter 1 Introduction Lexis was for a long time swept under the carpet in foreign language teaching (FLT), but over the last 20 years, this has changed. In the literature concerning vocabulary acquisition, there are accounts of what it means to know a word (see Chapter 5.1 and Table 5.1). However, much less attention has been paid to the way learners acquire such knowledge. Di fferent conditions thought to promote or impede vocabulary acquisition have been studied (as is partly done in this book), but the process itself has seldom been the focus of study (De Bot et al. 1997). Unlike grammar, which is fairly much acquired to serve most communicative needs by the time most learners leave school, new words are continually being acquired throughout a person’s life. Vocabulary learning never stops, as “a person is unlikely to ever run out of words to learn" (Schmitt 2000, 4). Many people have at least some experience of learning a foreign language and, in the process, learn new vocabulary. Interestingly, learning vocabulary is still strongly associated with rote learning in many parts of the world, that is, the repetition of items, usually using lists. For many learners, this is the only activity they associate with vocabulary learning. Over the past 20 years strong arguments for a focus on words instead of ‘language’ rules have been put forward, but this has not necessarily filtered down to the schools. In German, Wortschatz (vocabulary, often in relation to vocabulary size) expresses well what it means to know words: the more we know, the more abundantly our treasure chest of language is filled, enabling us to communicate on a variety of levels (Stork 2003). Vokabeln (a cognate of vocabulary, albeit with a narrower meaning) fill most learners with dread (Aßbeck 2002). Vokabeln are generally associated with having to 1 <?page no="18"?> learn long lists of pair associations (generally consisting of the target language word and a native language translation) (Aßbeck 1996; Knapp- Pottho ff 2000). This can be very demotivating as the mindless repetition of translation pairs are often perceived as being pointless and boring. These lists could and can still be found at the back of the course books used in the English language classroom. Even in primary school teaching of English, a fairly recent development in Germany and therefore more open to modern approaches and methods than other types of formal instruction, this focus on vocabulary is often absent. This is despite the fact that current research has shown that time spent teaching vocabulary is time well spent (Read 2004b). In a German school context, learning vocabulary in class is generally referred to as Wortschatzarbeit (vocabulary work), suggesting that this is fundamentally di fferent from learning Vokabeln (Knapp-Pottho ff 2000; Stork 2003). The practical implementation of theoretical considerations in formal language instruction, whether these considerations are linguistic or cognitive, is still rare, although many course books, for example, Discovery (Behrendt et al. 2004a,b) and Sally (Bredenbröcker et al. 2005a), claim that they base their methodology and curriculum on findings of linguistics, especially Cognitive Linguistics. The actual content of the course books, however, often does not reflect research findings in this area. The Common European Framework of Reference for Languages (CEF), when describing the linguistic competence a learner has to reach, does not only relate this competence to the range and quality of knowledge (e.g. in terms of phonetic distinctions made or the extent and precision of vocabulary). It also relates the acquisition to the cognitive organisation and the way this knowledge is stored (e.g. the various associative networks in which the speaker places a lexical item) and to its accessibility, that is, activation, recall and availability (Council of Europe 2001, 13). The CEF does not, however, state how these aims can be achieved. This issue is addressed in Cognitive Linguistic research and outlined in this book, as its objective is to give a description of the theoretical background of the lexical knowledge humans possess and the organisation of this knowledge in the monoand bilingual mind. The processes of second language acquisition and learning 1 will be discussed, with particular attention paid to Cognitive Linguistic models, including usage-based 1 The delineation between acquisition and learning in a second or foreign language is not always clear. I will use the terms interchangeably, although acquisition is generally associated with unconscious processes (see e.g. De Bot et al. 2005; Lightbown & Spada 2006; Johnson 2008). 2 <?page no="19"?> approaches to language acquisition (see e.g. Kemmer & Barlow 2000; Tomasello 2003). By connecting this research to studies conducted in the area of vocabulary learning and teaching, suggestions for an implementation of the theoretical assumptions in Early Foreign Language (EFL) classrooms will be put forward. At a time when research in the areas of the bilingual mental lexicon was still in its infancy, Meara (1993, 295) held the view that the "study of the bilingual lexicon is just not well enough developed for it to be able to tell practitioners what to do in classrooms". He continues by saying that it might be regarded as "esoteric and di ffi cult" (ibid.) by hands-on language teachers. Over subsequent years, however, further research has made it somewhat easier to access the research on the bilingual lexicon and use it to design informed lessons. In the intervention study outlined in this book, the focus is to do exactly this: to adapt laboratory findings and other research results in order to devise a way to teach vocabulary that might prove to be beneficial for foreign language learners. The Cognitive Linguistic perspective on foreign language learning in general and the acquisition of vocabulary in particular is a growing research area. The implications of Cognitive Linguistics for FLT are manyfold and promising, o ffering both ease of learning and more profound knowledge of the target language (see e.g. Niemeier 2008). These suggestions form the basis for the intervention study carried out in two German primary schools. The study investigates whether lessons enabling learners to elaborate on words and thereby process the vocabulary more deeply lead to better long-term retention of these items. Young learners were chosen as the subjects of this intervention because the main focus of empirical studies in vocabulary acquisition, learning and teaching still lies in the analyses of adult learners at intermediate or advanced levels, although some studies regarding SLA in young learners have been published. It has been shown that young learners may also benefit from teaching that is based on Cognitive Linguistic models (see e.g. Piquer Píriz 2008). In this context, the issue of measuring vocabulary knowledge will also be discussed, since no standardized vocabulary tests for young EFL learners have been proposed yet, although suggestions on how this can be done have been put forward (Becker et al. 2003, Niedersächsisches Kultusministerium [Lower Saxony Ministry of Education] 2006. The results of this empirical study, and their implications for future research, form the basis for evaluating the relevance and benefits of the theoretical implications of vocabulary research for primary school learners of English. 3 <?page no="21"?> The Mental Lexicon <?page no="23"?> Chapter 2 The L1 and L2 Mental Lexicon When discussing the mental lexicon, metaphors are frequently employed to give the reader an idea of what this lexicon might be compared to. One comparison that many people might associate with the term lexicon is that of a dictionary. However, this might bring about assumptions not in keeping with the actual organisation of the mental lexicon. The term mental lexicon is often used to refer to the “human wordstore" (Aitchison 2003, 10), which every speaker carries around “inside his / her head" (Singleton 2000, 161). It is said to be a part of long-term memory (Möhle 1997, 39) and, more specifically, “the memory system in which knowledge of a vast number of words, accumulated in the course of time, has been stored" (Hulstijn 1997, 210). The mental lexicon discussed in detail in the seminal book Words in the Mind by Aitchison (2003) is for the most part that of a monolingual speaker of English (Aitchison 2003, 4). She does, however, acknowledge that this is a situation “somewhat unusual in the world at large, where it is common for humans to use more than one language" (ibid., 255). The more usual bior even multilingual 2 mental lexicon, however, poses additional questions when exploring how it is organised (see Section 2.4). In order to devise a model of the mental lexicon, researchers use data collected from a number of psycholinguistic experiments. These experiments used an array of settings and methods. A short description of these is given in Section 2.2. (For a more detailed discussion see, for example, Aitchison (2003, chapter 2), Dong et al. (2005), Penke (2006)). 2 Some researchers pay special attention to a lexicon that incorporates more than two languages. For a description, see Cenoz et al. (2003). 7 <?page no="24"?> 2.1 Defining the Content of the Mental Lexicon One important question, when defining the mental lexicon, is: ‘What exactly is a word? ’ This straightforward question has, unfortunately, no easy answer, even if it is felt that we intuitively know exactly what a word is (cf. Aitchison 2003, chapter 3). Lutjeharms (2004) calls this lack of an easy answer a fundamental problem of delineation. For example, does the English expression test score consist of two words? And the German translation equivalent Testergebnis? The term lexeme, used to describe an abstract unit of the lexicon, is often used in the literature. Other terms such as sememe, morpheme, lexical entry and lexical unit, are used with the intention of solving the definition problem (Lutjeharms 2004, 10). The term word expressions, where the concrete representative of the abstract unit is referred to as a word form (Singleton 1999, 10) can also be found in the literature. To complicate matters even further, the term lexeme is used to denote something di fferent in the field of cognitive psychology. In this field, a lexeme is the phonological form of a word, whereas a basic abstract unit of the lexicon is called a lemma (for a bilingual adaptation of his model, see Section 2.3.2). Lutjeharms (2004) additionally ascribes the di ffi culty of fully grasping what a word is to the complexity of the words themselves. Among other things, all the layers of meaning and use have to be mastered and acquired. (For a more comprehensive description of what is involved in knowing a word, especially in the area of teaching lexical knowledge, see Chapter 5.1.) Most researchers agree that the lexeme, at least, contains information about phonology, morphology, syntax and semantics (e.g. Handke 1997; Wol ff 2002). Others also include graphemic information (e.g. Singleton 1999; Aitchison 2003). De Bot et al. (2005, 51) state that the lexicon contains “lemmas with all kinds of conceptual, semantic, syntactic and pragmatic information with their corresponding lexemes". These, however, are not necessarily the only aspects of information a speaker may have. For example, emotions might correspond with certain concepts or the speaker “may possess concepts for which there is no lemma" (ibid.). According to Börner & Vogel (1997, 3), knowledge stored in the L1 mental lexicon is organised, encompassing declarative knowledge (representation) and procedural knowledge (cognitive processes). 8 <?page no="25"?> 2.2 Methodological Approaches There are a number of psychoand neurolinguistic experiments that have helped to gain insight into the mental lexicon. The results of these help to evaluate existing models in terms of their ability to explain observed phenomena, to adapt the models, or to develop new models that better fit findings. In this section, a short description of the most common methods will be given. This description is, however, by no means exhaustive. (For a more comprehensive discussion, especially with regard to neurolinguistic and neurobiological methods, see, for example, Penke 2006, chapter 2.) 2.2.1 Psycholinguistic Methods In order to gain insight into the inner workings of our minds or, more specifically, our mental lexicon(s), a large quantity of psycholinguistic methods are available that examine the cogs (Aitchison 2003, 18). These methods can be divided into two major groups: those that look at the results of language production or judgements about language, for example, whether a sentence is grammatical or not, and those that look at the underlying processes while they are being carried out (see e.g. Müller 2003; Gullberg & Indefrey 2006; Penke 2006). O ff -line Experiments Data collections of speech errors tell us something about the underlying mechanisms of language, in this case by looking at malfunctions (Aitchison 2003, 18f). By looking at slips of the tongue, we can observe di fferent types of errors, and, in particular assemblage errors and selection errors. The former means that the right item is chosen, but assembled incorrectly, for example, par cark, and not car park. The latter are the result of incorrect selection of an item. Aitchison (2003, 19) gives an example of this when she refers to a slip of the tongue Prince Edward made when talking about his schooling: ‘Corporal punishment is a last resort. It is di ffi cult to use capital punishment in any institution. A beating is very valuable: it shows people you have come to the end of your tether [italics mine].’ In the second sentence, the word corporal has been substituted by capital. Obviously, in British schools, capital punishment is never used. Another way to collect data is to elicit words or structures. During elicitation, a context is created that leads to the production of a certain form. One frequently cited example is the Wug test (Gleason 1958). The 9 <?page no="26"?> subjects are given a nonsense word to test whether they can form the regular plural form, for example, ‘One wug, two . . . ? ’. The answers given by participants are then analyzed (Penke 2006). In addition, there is the tip of the tongue (TOT) phenomenon. It can be quite di ffi cult to gather information on these in natural speech, because they are quite rare. Psychologists have, therefore, devised a way to deliberately induce a TOT state by giving subjects a definition of an uncommon word. Subjects are then asked about the word they cannot quite remember, thus providing data on word searches (Aitchison 2003, 24). Another way to gather production data is by using word association tasks. In these, subjects are presented with a word and are asked to provide the first association that springs to mind, either verbally or in writing. The associations provided can be analyzed according to the relation they show with the stimulus word (see for example Ellis 1996; Schmitt 2000; Wolter 2001). On-line Experiments The largest group of on-line experiments are reaction time experiments. In these experiments, subjects are presented with tasks that they have to complete. One common experimental design is the priming experiment, and, in particular, semantic priming. The participants are asked to judge whether or not a string of letters appearing on a screen, or the string of sounds heard, constitutes a word. In these lexical decision tasks, the time taken to make this decision is recorded (Aitchison 2003). During a priming experiment, the subject is additionally presented with a prime word (sometimes for only a few milliseconds, so that subjects are not even consciously aware that there is such a thing as a prime), followed by a target word. If there is a semantic relationship between the prime word and target word, reaction times are shorter than when the prime and target are semantically unrelated. In order to exclude other variables, primes can be cross-modal, that is, the way of presenting items is varied. For example, the prime might be presented aurally and the target word is given visually or vice versa. In priming experiments, the participant is thus prepared for a subsequent item, and the attention is pre-activated, as it were. The word winter, for example, primes the word snow, speeding up the recognition of this word in lexical decision tasks (Aitchison 2003, 25). The e ffects of 10 <?page no="27"?> priming have been found in a wide area of studies, not only in respect to semantic relations. There is also evidence of morphological priming, among others (Penke 2006). Furthermore, reaction time experiments can be used to explore the effects of word frequency (for example by studying whether more frequent words are processed more rapidly than less frequent words), the impact of word length on speed of processing, grammaticality judgements and in a number of other areas in psycholinguistics. 2.2.2 Neurolinguistic Methods Neuroimaging techniques are another way of investigating the processes involved in language use. There are several of these techniques available. However, only the most common ones will be described. (For a more comprehensive description, see Ahlsén (2006), Lefrançois (2006) and Randall (2007). See Gullberg & Indefrey (2006) for a collection of articles on the neuroscience of second language acquisition (SLA).) Electrical Activity The electric activity of brain cells can be measured in di fferent ways. The most familiar way of measuring this is by using electroencephalography (EEG). “It represents the potential di fferences between two points on the scalp surface caused by brain waves" (Ahlsén 2006, 163). For this, an even number of metal electrodes (usually 8-16) are attached to the scalp. The brain waves are then recorded while the subject carries out the experimental task. Similar to this, evoked or event-related potentials (ERPs) can be recorded using the same type of electrodes as in an EEG. However, an ERP is a electrical activity generated by reactions to stimulation and has only a small amplitude. Therefore, an averaging technique is applied, that is, the same test is done a great number of times. This is necessary, because blinking or frowning may interfere with the measurement. ERPs are categorized by their peaks, making them either positive or negative, representing stages of information processing. ERPs are named thus: N for a negative peak and P for a positive one. In addition, the number of milliseconds it peaks from the stimulus onset is measured, leading to names like N400 or P600 (Ahlsén 2006; for a detailed discussion see Penke 2006). These methods are used in a variety of psycholinguistic experiments, for example, looking at the e ffect ungrammatical sentences have on lan- 11 <?page no="28"?> guage processing. The experiments are non-invasive and fairly inexpensive, but on the other hand a large number of runs have to be done for each test because of the averaging technique. Imaging The two most commonly employed methods of producing images of the (working) brain are positron emission tomography (PET) and functional magnetic resonance imaging (fMRI). Both methods, although using different techniques, produce image slices of the brain. These show areas of increased blood flow at a given moment. In order to do this, PET uses a radioactive dye and fMRI measures increased blood flow through the magnetic properties of blood and water. “The argument is that areas of the brain which are actively engaged in a task will have increased blood flow. Thus, ‘hot spots’ with greater blood flow are areas which are processing the information" (Randall 2007, 24). Both methods allow us to localize activities. However, as the brain is constantly active and carries out a lot of simultaneous stimulus processing, it is quite di ffi cult to separate the activity caused by the subject carrying out the task while lying in the scanner, and the activity caused by the subject blinking, and so forth. When investigating language tasks, this is poses a serious problem and is addressed in two ways: [E]ither one uses a natural (“ecological") task such as the comprehension of a short story, in which case one must determine which component of the task is responsible for triggering which of the activated areas - and, supposedly, why. Alternatively, one attempts to target only one component of a complex task (e.g., single word comprehension, syllable detection, etc.), in which case one runs to not-so-unlikely risk of tapping into something other than what is used in a natural linguistic task. (Paradis 2004, 153) Although there is empirical evidence indicating that the languages of bilinguals are represented as independent subsystems, there are still many unanswered questions. This is because the spatial resolution of the normal 2 to 3 mm is still insu ffi cient (Paradis 2004, 127; 159). (For a more detailed discussion of the problems of neuroimaging studies, see Paradis 2004, chapter 6.) 12 <?page no="29"?> 2.3 Modelling the Mental Lexicon 2.3.1 Models of Lexical Processing The question of how lexical knowledge is organised has led to the development of several competing models. All models have to take into account the e ffects (listed below) that have been discovered using various psycholinguistic tests: • word frequency (frequent words are processed more rapidly) • lexicality (real words are processed more rapidly than nonwords) • word length (long words are processed more slowly) • context (the context has an important impact on word recognition) (see Handke 1997, 97f) Singleton (2000) makes a distinction between direct and indirect models of lexical processing. The indirect models have a two-stage access, a search procedure and a retrieval procedure. The direct models only need one stage to access a word. Singleton (2000, 170) uses two metaphors to characterize the distinction between the indirect and direct models: The indirect type of model assumes that the processing of lexical knowledge follows the same kind of pattern as looking up a word in a dictionary [...]. This kind of model sees lexical access as involving more than one component or step. Direct models, on the other hand, portray accessing lexical knowledge as a one-step process; a metaphor which has been used in this connection is that of a computer software package which allows items stored by name to be accessed simply by typing in of as many letters as are su ffi cient to distinguish the relevant name from all other stored names. De Bot (1992), on the other hand, distinguishes between active and passive models. Active models are defined by an active retrieval process, that is, the lexicon is scanned for words that comply with the defined characteristics. This process is time-consuming, as the entire lexicon has to be scanned. In order to account for such variables as the frequency e ffect, active models have been modified, for example, by stipulating that the items are ordered according to frequency of occurrence. 13 <?page no="30"?> In contrast, passive models seem more promising with regard to time consumption. Lexical elements have a certain number of characteristics which have to be stimulated, to a certain extent, in order to become activated. Each item has a detector that monitors the input for the relevant characteristics. An activated item then presents itself as a possible candidate for a given slot. The lexicon is, therefore, not actively searched. These models, for example, the logogen model (see Section 2.3.1), have the advantage of being extremely fast, as the number of possible candidates is narrowed down very quickly (De Bot 1992, 432). In the following, a concise description of the most recent influential models will be given. Levelt’s Blueprint Uttering a word or sentence is a highly complex event. To explain the complexity involved in speech processing, Levelt’s Speaking Blueprint (Levelt 1989, 1993) is one of the most widely accepted language production models for the monolingual speaker, as it is the most complete, empirically based model available (De Bot et al. 1997, 2005; see also Figure 2.1). The model consists of three stages: the Conceptualiser, the Formulator, and the Articulator. Speech production starts in the Conceptualiser, which generates a preverbal message that contains meaning intentions. These have to be put into words in the stages that follow. This message contains conceptual characteristics that lead to a selection of a set of lexical items called lemmas (words to represent a concept) in the Formulator. Once the lemmas have been selected, they are grammatically encoded, that is, placed into a well-formed sentence, which in turn has to be phonologically encoded. In speech comprehension, the same steps are involved, but in reverse order (De Bot et al. 2005, 40f). In Levelt’s model, the mental lexicon consists of two elements, the lemma and the lexeme. The lemma contains all the conceptual, semantic, syntactic and pragmatic information, whereas the lexeme is the phonological form associated with this lemma (for a discussion of the di ffering uses of these words in other linguistic areas, see Section 2.1). The lexicon plays a central role in this model, as Levelt believes that the formulation processes are lexically driven so the lexicon plays a central role in this model. The model places the lexicon as mediator between conceptualization and grammatical and phonological formulation (Levelt 1989, 181; Singleton 2000, 175). However, lexical knowledge, in this case, is purely declarative, since procedural knowledge is removed from the 14 <?page no="31"?> Figure 2.1: Levelt’s Blueprint for Speech Production and Comprehension (De Bot et al. 2005, 40) 15 <?page no="32"?> domain of the lexicon. Although this model is widely accepted, it is not without drawbacks, among them the strict modularity, lacking direct feedback mechanisms. De Bot et al. (2005, 41) sum this up in the question “How, for instance, can exactly the right words be selected from the lexicon when the Conceptualiser has no knowledge of which lemmas the lexicon contains? " As mentioned above, Levelt’s model was developed for a monolingual speaker only. The adjustments made to accommodate multilingual speakers will be discussed in detail in Section 2.3.2. The Logogen Model The logogen model, a direct model that attempts to explain language comprehension, was proposed by Morton (1969) in order to account for the relation between lexical responses participants gave in sentence completion tasks and the time it took them to recognize items in a context (Singleton 1999, 2000). The model derives its name from the logogen 3 , an “information gathering device" (Randall 2007, 57), one logogen for each item in the given mental lexicon. Its task is to gather perceptual information and semantic evidence for the presence of the word to which the logogen corresponds. When the information points towards a certain word, the corresponding logogen fires (Singleton 2000). In most contexts, more than one word could be used, therefore many logogens could fire. The logogen does not, however, fire when only one or two attributes pointing to the corresponding word are detected, but rather when a critical threshold is reached. This threshold level is influenced by many factors, for example, word frequency (Randall 2007). Each logogen has a certain activation level, depending on how often it has been activated in the past. The logogen is activated by a variety of input clues, for example, auditory and visual stimuli. This model can explain the e ffects of context that occur in priming experiments. If there is more than one language present, it is possible that a given language set is selected, meaning that all logogens for the other language do not fire in this context (Raupach 1997, see also Section 2.4). The Cohort Model One problem of the logogen model is the impreciseness of the notions ‘threshold’ and ‘activation level’. The cohort model (Marslen-Wilson & 3 The term logogen is a blend of the Greek term lógos, meaning ‘word’ and the Latin root gen, meaning ‘to bring to life’ (Singleton 2000, 171). 16 <?page no="33"?> Tyler 1980) supplies a solution to this problem because it states for each word where the critical activation level occurs (Singleton 1999, 2000). This model is named after a division of the ancient Roman army, an apt description as “[a] whole army of words, it seems, marches up for consideration each time a word begins" (Aitchison 2003, 235). This model, also primarily concerned with word recognition, stipulates a set of word detectors that are activated whenever a word is heard. As soon as the first sound (the ‘word initial cohort’) is heard, all detectors for words beginning with this sound are activated. These detectors then go on to monitor further input. Initially, a large amount of possible words are activated and, subsequently, narrowed down as new information becomes available until there is only one possible word left. This is determined by the ‘uniqueness point’, the precise point when a word is recognized (Singleton 2000, 172), thus predicting the critical activation level for every item in the lexicon (Singleton 1999, 91). One problem that arises, especially with the earliest version of this model, is that it could not cope with any ambiguous acoustic input. If the first sound is wrongly chosen, an incorrect cohort will be activated (Aitchison 2003, 236). Although this problem has been addressed in more recent versions of this model, it has led to an overlapping with other models, for example, the interactive activation model (see also Section 2.3.1). Connectionist Models Connectionist models, sometimes also called associationism, derive their name from an analogy between them and neuronal interaction in neural networks (Lefrançois 2006). This analogy is based on the assumption that any time the brain is active, some brain cells are active. They then send out neural signals to other cells. These signals can either cause arousal (excitatory signals) or suppression (inhibitory signals). This interaction results in a network of interconnected units, where repeated arousal reinforces the connections, while inhibitory signals lead to the deterioration of a connection (Aitchison 1992; Singleton 1999). Another term sometimes used for these models is that of ‘parallel distributed processing’, which “refers to the claim made by connectionists that di fferent portions of information are processed independently of one another (‘in parallel’) on di fferent levels (‘distributed’)" (Singleton 1999, 121). In language processing, a connectionist view states that each item in 17 <?page no="34"?> the mental lexicon is connected to one or many other entries, just as the di fferent areas of the brain are interconnected (Singleton 2000; De Bot et al. 2005). Researchers pursuing this line of thinking believe that the whole brain contributes to linguistic processing, not just specific language areas. The interaction between all these areas and processes is studied in order to discover more about language behaviour (Obler & Gjerlow 1999). Almost all models of the mental lexicon today are based on the principle of connectionism (De Bot et al. 2005), often combined with the activation metaphor (see below). Spreading Activation Models In spreading activation models, also known as Interactive Activation Models (IAMs) (NB: there is slight di fference between these two terms - see below), the items of the mental lexicon are represented as nodes which are connected to each other. These connections are not hierarchical, they are determined by the strength of the association. Categories are not necessarily determined by features. Category membership and typicality are not determined by logical relationship, but by strength of association (Randall 2007, 116). The IAM specifically was first proposed by McClelland & Rumelhart (1981). In this connectionist model, one of the principal elements is “the concept of spreading / interactive activation, the idea that in language processing a multiplicity of nodes are excited by the arousal of a node to which they are connected" (Singleton 1999, 125). According to spreading activation theory, the activation flows in one direction only, that is, outwards to more and more nodes, whereas the interactive version allows activation to move “backwards and forwards between the activated nodes" (ibid., 126). In Figure 2.2, an possible network for firetruck is shown. The strength of associations is represented by the length of the connecting lines. It can be seen that the item, firetruck, is not only connected to all other terms for vehicles. It is also directly and indirectly connected to those things that have the same colour, that occur in the same context (e.g., heat) or that have a connection depending on individual experience, not represented in the schematic drawing, but nevertheless possible, for example, hero. In this model, any activation will lead to connected nodes also receiving a rise in level of activation, which in turn prepares them for recognition. Therefore, this model can explain the priming e ffect (see Section 2.2), since all the items connected to the prime will have a raised level of activation, which results in faster word recognition (Randall 2007). 18 <?page no="35"?> Figure 2.2: A Spreading Activation Model (Randall 2007, 115) The spreading or interactive activation models are models “in which an initial impetus progressively fans out and activates more and more words as it spreads along various connections" (Aitchison 2003, 224). These activated links are then inspected, the relevant ones become more and more excited and the irrelevant ones get less excited, or, as Aitchison (ibid.) describes it: “the rich get richer, and the poor get poorer". In these models, one of the principal elements is “the concept of spreading / interactive activation, the idea that in language processing a multiplicity of nodes are excited by the arousal of a node to which they are connected" (Singleton 1999, 125). As mentioned above, in the spreading activation view, the activation flows in one direction only, because an initial impetus progressively fans out and activates more and more words as it spreads along the various connections. As the activated links are inspected, those that are relevant get more and more excited, while those that are unwanted fade away. (Aitchison 2003, 224) The IAM has the advantage of being highly individual, thus allowing for personal di fferences that can always be found, for example in word association tasks or other psycholinguistic experiments. An alternative way of modelling mental lexicons has been proposed by Meara (2006). He does not attempt to account for every detail of how a mental lexicon works, but rather employs only the essential features of a lexical network (ibid., 638). This allows an exploration of the fundamentals of the mental lexicon, for example, link constraints, which leads to 19 <?page no="36"?> certain consequences for bilingual lexical models. Furthermore, it raises the question of entanglement. Traditionally, bilingual lexical storage is only described in vague terms (see also Section 2.4). However, a formal model, like the one used by Meara (2006) does not work in vague terms. Rather, it needs clear underlying properties. Then “[e]mergent properties - features which are not built in to the underlying assumptions, but just happen anyway - are much easier to identify" (ibid., 640). Although research in this area is not directly applicable to real mental lexicons or language teaching, formal modelling and simulation work can “open up very important new avenues of research in applied linguistics" (ibid., 641). 2.3.2 Modelling the Bilingual Mental Lexicon Levelt’s Lexical Processing Model for the Bilingual Speaker In order to determine how the bilingual mental lexicon might be organised, models have to be developed or adapted that go beyond the description of the monolingual language user. Levelt’s Blueprint (see Section 2.3.1) is a widely accepted model that is consistent with experimental data, therefore an adaptation for production and comprehension, when more than one language is present, must ideally explain certain phenomena of bilingualism. De Bot (1992) addresses these phenomena, stating that an adaption of Levelt’s model must not only be able to account for aspects such as code-switching and cross-linguistic influence, but also o ffer the possibility to incorporate the fact that most bilinguals do not master both language systems to the same degree. Furthermore, the model should be able to incorporate an unlimited number of languages. Accordingly, the elements of Levelt’s original model (Levelt 1989, 1993; see also Section 2.3.1) must be examined in the light of these requirements. One of the first questions that arises is whether a bilingual model requires the reduplication of all or only certain components of the original model. De Bot (1992, 7) states that his aim is to “keep the original model intact as much as possible and to revert to adaptations only if empirical findings on language production cannot be explained with the existing model". A model striving to explain language comprehension and production, when more than one language is present, must ideally explain certain phenomena. For example, it should explain the fact that any bilingual speakers are able to use their two languages separately or to code switch, depending, among other factors, on the situational context. Addition- 20 <?page no="37"?> ally, psycholinguistic tests have revealed that there is, for example, a crosslinguistic influence, as evident in crosslingual priming e ffects. In addition, the model should be able to mirror the fact that most bior multilinguals are not equally competent in their di fferent languages. The level of proficiency should, therefore, be mirrored in such a model. One additional point is that the proposed model should be able to be extended to incorporate any number of languages (De Bot 1992). For each component of the original model (see Figure 2.1), the question arises whether it can function for bilingual speech without any fundamental changes (ibid.). The knowledge component is not language specific, therefore a single system su ffi ces (De Bot 2003). Although communication in di fferent languages relies on specific conventions, this is part of the same component. Levelt talks about choosing di fferent registers or varieties. Following this line of thinking, there is not much di fference between choosing di fferent registers or di fferent languages (De Bot 1992, 2003). Information on the register to be used, like the information on the chosen language is thus part of the preverbal message. Levelt (1989) assumes that the conceptualizer is language-specific and that these distinctions have to be defined in the preverbal message. The most plausible assumption is that at the macroplanning level, “the language to be used is selected on the basis of information from the discourse model and that accordingly language-specific encoding takes place in microplanning" (De Bot 1992, 8). One unsolved problem relates to how bilinguals deal with the problem that sometimes a concept, which needs to be expressed in a particular language, lacks a lexical item, leading to problems in the formulator during grammatical encoding, that is, when a (non-existent) lemma is selected from the lexicon. In principle, “the conceptualizer should ‘know’ that a given concept cannot be lexicalized properly, but it is absolutely unclear how this takes place" (ibid.). The formulator itself is not languagespecific and stores possible sounds of the languages. A major drawback of Levelt’s model is, however, that it is a state (that is, fixed) model and does not try to explain language acquisition or learning (De Bot 1992). A further adaption for the purpose of SLA has been proposed by De Bot et al. (1997). When reading, for example, the comprehension system tries to match a word form and a meaning. This process involves a number of steps. Firstly the string of letters must be matched to a lexeme. When this match is made, the lexeme activates a lemma, which in turn has to be matched 21 <?page no="38"?> Figure 2.3: Lexical Comprehension / Production Model for Oral and Written Modalities (De Bot et al. 1997, 315) with a concept (see Figure 2.3 and De Bot et al. 1997). In production, the information takes the reverse path, from the top down, whereas in comprehension there is interaction between top-down information and bottom-up information. The process of inferring word meaning within this framework “can be viewed [...] as an attempt to fill in an empty lemma structure" (ibid., 317). The Bilingual Interactive Activation Model The Bilingual Interactive Activation (BIA) model is based on the spreading activation approaches. In order to be able to distinguish between the di fferent languages, as users of more than one language are obviously able to do, a language node has to be included to allow for information regarding the language an item belongs to (Lowie 2000; De Bot 2003; Dijkstra 2003; Lutjeharms 2004). This node can act both as an inhibitor or an activator, depending on the situational context. This would favour the subset hypothesis (see Section 2.4). A helpful concept is the idea of language subsets. “Subsets are groups of lexical items that are clustered due to some shared characteristic. The shared characteristic may, for instance, be a register (formal, informal, etc.) or a language" (De Bot et al. 2005, 46). As the BIA is multi- 22 <?page no="39"?> dimensional, a word can be a member of several subsets. It is suggested that the activation level of entire language subsets can be increased or decreased, leading to the possibility that at one point in language processing a choice about language is made. This “language selection must occur at lemma stage before it is matched with a lexeme" (ibid., 47). The language nodes help to control language processing by activating a particular language subset and inhibiting others (De Bot et al. 2005). This, however, is not an ‘either or ’ decision: Activation, and in particular inhibition will never be like an on / o ffswitch. It is more like holding down ping-pong balls in a bucket filled with water: With your hands you can hold down most of the balls, but occasionally one or two will escape and jump to the surface. Likewise, complete suppression of a language, particularly one with a high level of activation may be impossible. (De Bot 2004, 26) A related proposal was put forward by Bock & Levelt (1994), who added a language level to the spreading activation model (see also Randall 2007). This model incorporates three levels: the conceptual, the language (or lexeme) level and the lemma level (see also Sections 2.1 and 2.3.2). Here, the conceptual level operates as a semantic network, with concepts connected to each other in networks. [...] In addition to the conceptual and language levels, they added a ‘hidden’ abstract level of knowledge which they called the lemma level. Nodes at this level contain the syntactical information about a word, which act as an intermediate level between the concept and the actual word itself, the lexeme. (Randall 2007, 115) This allows for the storage of more than one language and suggests that both languages are stored in one place. This, however, is not the only possible organisation of the mental lexicon (see below). 2.4 Internal Organisation of the Bilingual Mental Lexicon Over 50 years ago, researchers had already raised the question as to how the languages of a bior multilingual person are organized, as “[t]he 23 <?page no="40"?> Figure 2.4: A Network Illustrating a Bilingual Spreading Activation Model (see also Randall 2007, 116) question of ‘two systems or one’ resurfaces at each level of analysis of bilingual language behavior" (Obler & Gjerlow 1999, 128). One of the first models of the organisation of bilingualism was proposed by Weinreich (1953). He describes three possible relationships between the lexicons of the languages: coordinate, compound or subordinate (also called subordinative) 4 . Weinreich (1953) attributed these di fferent types of bilingualism to di fferent types of learning experience: • learning an L2 through the L1, according to him, resulted in a subordinative bilingualism; • compound bilingualism could be achieved if two languages were used interchangeably at home; • if the L2 learning was school-based, the di fferent contexts in which languages are learned would result in coordination. 4 The terms compound and coordinate are not the only ones used. The distinction between compound and coordinate lexical organisation has been referred to as common or shared storage vs. separate storage, interdependence vs. independence and single code vs. dual code (for an overview see De Groot (1993, 1995)). 24 <?page no="41"?> However, Weinreich (1953, 10) acknowledges that a person’s bilingualism need not be of a single type (see also De Groot 1993, 27f; Singleton 1999, 173; Ahlsén 2006, 122). Figure 2.5: Possible Types Organisation of the Mental Lexicon (adapted from Obler & Gjerlow 1999, 129) As illustrated in Figure 2.5, the proposed organisations di ffer in the way they are connected to the conceptual knowledge of a speaker. Whereas the L1 and L2 lexicons are connected in the compound organisation and thus form one system and access the conceptual store as one, the coordinate perspective postulates two separate systems that access the conceptual store independently. The subordinate organisation also assumes separate systems, but in this case “L2 word forms are connected to L1 meanings via primary connection to L1 forms" (Singleton 2003, 169). Paradis (2004, 187) classifies these di fferent organisational patterns in terms of the possibility of borrowed elements of di fferent kinds, so-called interferences. The coordinate model shows no interference at all, the compound organisation allows bidirectional interference and the subordinate model only allows unidirectional interference. Interference does not only apply to lexis and grammar, but also to the cue strength in the interpretation of sentences and pragmatic relations (Paradis 2004, 187f). One argument for the separation of the two lexicons is outlined by Singleton (1999, 167f). When analysing an unfamiliar word, for example, we look for morphemes that may only be permissible in one of our languages but not in the others. These analogizing tactics (or ‘gang e ffects’, as Singleton (1999, 168) calls them) run through each language separately. Arguments in favour of integration come from cross-linguistic influ- 25 <?page no="42"?> ence and code-switching (Singleton 1999, 168), because any speaker can use an L2 word in an L1 sentences, sometimes even modifying the L2 word according to L1 syntax or morphology. In accordance with research findings, a definite dichotomy seems to be unsustainable. Singleton (1999, 190f) summarizes the conundrum thus: It appears from the evidence reviewed that L1 and L2 lexis are separately stored, but that the two systems are in communication with each other - whether via direct connections between individual L1 and L2 lexical nodes, or via a common conceptual store (or both). Therefore, and in the light of recent findings, the possible models have changed somewhat. Instead of the fixed categorisation of compound, coordinate or subordinate organisation, a mixed system has been proposed. One deciding factor may be word type. Drawing on research that discovered very strong priming e ffects for cognates 5 , a compound system is suggested for these (De Groot 1993; Meara 1993; Nation 1993; Lutjeharms 2004; Dong et al. 2005). Bickes (2004) reviews other evidence that seems to indicate that cognates are more easily processed than non-cognates. Concrete nouns might also be processed more rapidly because they might not only be stored in the semantic memory system, but also linked to an image. These vocabulary items are also learned more easily than abstract nouns (Ellis & Beaton 1993). This could be because “[c]oncrete words and cognates may share more conceptual nodes than abstract words and noncognate words"(Dong et al. 2005, 222), thus facilitating activation. Hulstijn (1997) and Paradis (2004), among others, go beyond the distinction of compound, coordinate or subordinate relations and describe four hypotheses that have been proposed. The extended system hypothesis assumes a single, undi fferentiated store. According to this hypothesis, all words share the same store, but each entry carries a little tag, as it were, that distinctly labels the word as L1, L2, L3, and so forth. In light of the spreading activation, it has been argued that L1 words automatically inhibit those with an L2 label and vice versa. If this were true, however, it would be much more di ffi cult for bilingual speakers to switch between languages. However, speakers frequently do so with ease, sometimes even within the same sentence. The dual system hypothesis stipulates separate, independent stores for each language. This hypothesis, at the other end of the cline, seems to be too simplistic, as word association tests and other experiments suggest 5 Cognates are translation-equivalent words similar in form, whereas non-cognates are dissimilar translation equivalents (De Groot & Poot 1997). 26 <?page no="43"?> at least some interference occurs when an individual has some knowledge of more than one language. Figure 2.6: Schematic Representation of Various Models of the Organisation of the Mental Lexicon (Paradis 2004, 111) The tripartite hypothesis might explain the phenomenon of interference, as it assumes an overlap between stores. According to this hypothesis, similar words (cognates) are thought to share a common store, whereas language-specific words are housed in separate stores. The fourth hypothesis, the subset hypothesis, is based on the idea that L1 words are strongly connected, as are L2 words, in a way that could be described as separate families within a community (Hulstijn 1997, 211). This fourth hypothesis is also favoured by De Bot (2004, see also Section 2.3.2). A schematic representation of these models can be found in Figure 2.6. In this figure, the symbol A represents elements of language A, B symbolizes language B and A / B represents the fact that these elements are equivalent (Paradis 2004, 111). De Groot (1993) shares the view that di fferent lexical attributes lead to di fferences in storage. According to her interpretation of the data available, she advocates not one single representational system, but rather a mixed system “where concrete words and words perceived as cognates 27 <?page no="44"?> across the two languages are stored in a ‘compound’ manner, whereas abstract words and non-cognates in the respective languages are stored in a ‘co-ordinate’ manner" (Singleton 1999, 172). Drawing on a range of evidence from cross-language priming, and other tests, Kroll (1993) argues for a model of lexical and conceptual representation that allows for both lexical and conceptual links between the two languages (see Figure 2.7). The strength of these connections di ffer, however, depending on factors like proficiency and age of acquisition. Kroll (1993) also argues for a developmental shift from lexical mediation to concept mediation as proficiency increases. Studies suggest that this shift appears quite early in second language acquisition (Kroll 1993, 75). Figure 2.7: Revised Hierarchical Model (Kroll 1993, 54) Dong et al. (2005) review several models that have been proposed in recent years. They found that the “general consensus seems to be that bilinguals rely on a single shared conceptual store" (Dong et al. 2005, 223). Lutjeharms (2004) suggests that the organisation of a multilingual mental lexicon may be influenced by a variety of factors, including, but not limited to, word properties and language proficiency. Another factor influencing the activation of an item is its frequency, with the level of proficiency also having an influence here because it leads to a form of subjective frequency (Lutjeharms 2004, 15). These and other factors all interact with one another. Scherfer (1997) proposes that monolingual speakers have a subsystem of lexical concepts within their conceptual store. Bior multilingual speakers, therefore, have two or more lexical subsystems. Every lexical concept carries a language index, marking it L1, L2, L3, and so forth. The Three-Store Hypothesis also proposes a separate nonlinguistic cognitive store. It postulates that a bilingual speaker possesses two language sys- 28 <?page no="45"?> tems (one grammar for Lx and one for Lz, including a set of lexical meanings for words in each language) and one non-linguistic cognitive system which interacts with the two grammars in every act of comprehension or expression. (Paradis 2004, 196) Evidence for this separation between the conceptual store and the linguistic systems can be found in paroxysmal aphasia, where patients show normal behaviour with bouts of aphasia while preserving non-verbal intelligence (Paradis 2004, 196). As discussed above, it posits two separate neurofunctional entities: The lexicon, which is part of the language system, and the conceptual system, which is ontogenetically prior and builds concepts through experience and, later, groups together conceptual features so as to match the semantic constraints of lexical items. The result is that the cognitive system eventually builds linguistically and culturally defined concepts and classification procedures in addition to experientially derived ones. (Paradis 2004, 198) The grounding in experience is also important for the Cognitive Linguistic theory outlined in Chapter 4. 2.4.1 L1 and L2 Lexicon - Same or Di fferent? As will be discussed in more detail in Chapter 5, words are related to each other in a large number of ways. “It seems logical to assume that these relationships are not just quirks, but reflect some type of underlying mental relationship in the mind" (Schmitt 2000, 38). In order to investigate these mental relationships, psycholinguistic tests, for example, word association tests, are used (see also Section 2.2). The relationship between the cue word and the associations provided by the subjects are commonly categorised as syntagmatic, paradigmatic, and clang relations (Singleton 1999; Schmitt 2000; Wolter 2001). As an illustration, consider the stimulus word dog. If a subject answers bite, this is seen as a sequential relationship, for example a collocation, and is usually a member of a di fferent word class. These responses often occur in close proximity in a sentence. This kind of relation is called a syntagmatic relationship. 29 <?page no="46"?> A paradigmatic response to the same stimulus would be cat, which belongs to the same word class. Paradigmatic relations are more semantic in nature, often exhibiting sense relations including near synonymy. The last category, clang association, is solely based on phonological similarity. For example, the response fog to the stimulus dog is not a semantic relation. Syntagmatic responses are also referred to as combinatory associations and paradigmatic responses as substitutional associations (Singleton 2000, 170). It has long been maintained that the L1 and L2 mental lexicon, regardless of the relationship between them, have fundamentally di fferent internal structures (for an summary of this discussion, see e.g. Singleton 2007). Wolter (2001, 42) summarizes the proposed di fferences as follows: (a) the connections between words in the second language learner ’s mental lexicon are less stable than the connections of native speakers, (b) phonology appears to play a much more prominent organizing role in the L2 mental lexicon than it does for native speakers, and (c) the semantic links between words tend to di ffer in a systematic way from those of native speakers. Analysing several studies, Meara (1980) concludes that, although syntagmatic and paradigmatic relations between L2 words developed in a similar way to those during L1 acquisition, the answers given during word association tasks in an L2 are less stereotypical, that is, a broader variety of answers can be observed. Furthermore, the associational links connecting L2 items seem to be less stable, less numerous and less easily accessible during experimental tests (Meara 1980). The reasons for this lie in the fact that second or foreign language (L2) learners usually have a smaller vocabulary than native (L1) speakers, therefore they do not have the same amount of associations and connections between nodes. The nodes themselves are not as numerous as for native language (L1) speakers. This may lead to the L2 learners’ mental lexicon being less advanced (Schmitt 2000). Another reason for the di fferences in responses during word association tasks might be found in the fact that L1 children also produce more syntagmatic responses, whereas L1 adult speakers produce paradigmatic responses. The di fference, therefore, might lie in the stage of organisation, that is, L2 learners could go through similar stages when learning a new language. 30 <?page no="47"?> Lastly, although the responses of L2 learners are highly unstable, as proficiency increases, the patterns of the responses stabilize. This supports the theory that the di fference in word association tasks reflect the development of proficiency rather than a fundamental di fference in quality (Schmitt 2000). In support of this theory, Söderman (1993) found that second language learners’ clang and syntagmatic responses decreased as the exposure to the L2 increased. Thus, native speaker-like associations in an experimental setting seem to be linked to greater language proficiency (Schmitt 2000). The same seems to hold for the number of associations provided. The higher the level of proficiency subjects have, the more associations they produce. This seemingly suggests that “a greater number of responses indicates more words are connected to the stimulus word in the lexicon" and the lexicon itself has “a greater level of organization" (Schmitt 2000, 42; see also Wolter 2006). This research indicates that not only should paradigmatic relations play a role in L2 teaching, but also syntagmatic relations, in particular collocations, as these seem to be a natural way in which learners connect words during early stages of SLA (Schmitt 2000). Some researchers have also voiced doubt about the conclusions that can be drawn from word association experiments using decontextualised, isolated words. Language is a highly complex system depending on a large number of factors and variables, therefore the evidence from psycholinguistic tasks may be less enlightening than previously thought (Paradis 2004). 31 <?page no="48"?> 2.4.2 The Influence of Proficiency De Groot (1995) lists eight factors that she believes influence the organisation and reorganisation of a bilingual’s mental lexicon: • The level of proficiency a bilingual has reached in her / his second language; • specific characteristics of words; • the strategy used when words in a new language are being learned; • the context in which the two languages are used; • the age at which the new language is first acquired; • the script relation between the bilingual’s two languages; • the family relation between the two languages. (De Groot 1995, 152) The level of proficiency seems to be one of the most important influencing factors in the organisation of the mental lexicon. In the early stages of foreign language learning, in a formal instruction setting, the beginner ’s mental lexicon seems to be of the subordinate type (Kroll & Stewart 1994; De Bot et al. 2005). Newly acquired L2 lexical items need to go through L1 equivalents to access the conceptual level (see e.g. Näf 2004; Lutjeharms 2004). One possible explanation for this might be that beginner language learners are traditionally taught new L2 items by direct wordto-word translations (Snodgrass 1993, 87). This view has been supported by psycholinguistic translation experiments, although these findings are controversial (Kroll 1993; Snodgrass 1993). De Groot & Poot (1997) investigated the e ffect of L2 proficiency e ffect in translation tasks, both in forward (L1 to L2) and backward (L2 to L1) translation. They found that the translation processes di ffer at di fferent levels of L2 proficiency, “suggesting di fferent types of underlying memory structures" (De Groot & Poot 1997, 218; see also O’Gorman 1996). The di fference is seen to lie in the type and strength of the connections between memory stores (see Figure 2.7). Whereas high proficiency leads to strong connections between representations in L2 word-form memory and the corresponding representations in conceptual memory, low proficiency shows the opposite pattern. Here, the connections between the L1 and L2 word-form stores are the stronger ones. “As a consequence of these di fferences in strength of the various types of connections, the 32 <?page no="49"?> researchers assume that word translation proceeds along di fferent routes in bilinguals of di fferent fluency levels" (De Groot & Poot 1997, 218). Scherfer (1997) suggests that learners with a low level of proficiency store new L2 knowledge according to formal criteria first, predominantly relying on phonological similarities. This would explain why beginners use a larger number of clang responses in word association tasks than compared to more proficient L2 and L1 subjects. 2.5 The Neurocognitive Foundation Hagoort (2006, 93) states: “The classical view among neuroanatomists is that [...] architectural di fferences in brain structure are indicative of functional di fferences and, conversely, that functional di fferences demand di fferences in architecture". This theory, however, is not necessarily supported by the data. Some researchers hold the view that functional di fferences in brain areas are due to the form of the input signal. These signals then lead to functional specialization. The nerve cells, the neurons, are interconnected. If one nerve cell is excited, its axon sends signals to other neurons. In a tree-like fashion the axon of a cell can be in contact with 1,000 to 10,000 other neurons (Penke 2006, 77). The neurolinguistic theory of bilingualism proposed by Paradis (2004) “has a considerable impact on applied linguistic theory (linguistics applied to language teaching)"(Paradis 2004, 33). The brain is one large processing unit, in which everything is directly or indirectly interconnected. Some parts of the network may be dedicated to specific tasks and thus may form di fferent modules. These can be selectively impaired by injury, trauma and / or other pathologies (Paradis 2004, 191f). Paradis (2004, 189) states the null hypothesis for this theory of modularity, that is that there is nothing in the bilingual brain that di ffers in nature from anything in the unilingual brain. Hence the mechanisms for comprehension and production in the bilingual brain do not need any additional component that is not already present in the unilingual speaker-hearer ’s brain. Nor is there any di fferent component in the mechanism for a fluent as compared to a nonfluent bilingual. The only di fference is the extent to which they make use of parts of the verbal communication system. 33 <?page no="50"?> Based on the work of Fodor (1983), proposals have been made that linguistic knowledge is modular. A module is an autonomous system that only allows specific processing procedures (Lutjeharms 2004). Each of these modules “has its own internal structure and modus operandi which are una ffected by those of the other components" (Paradis 2004, 242). Fodor ’s view of modularity is di fferent from that of Chomsky (see e.g. 1980, 1988) because it applies to a larger area of language, not only language acquisition. It is also primarily concerned with language processing (Singleton 2000, 176). The research data, so far, are fairly inconclusive with regard to the modularity hypothesis, or as Singleton (1999, 113) more cautiously expresses it: “such indications as are forthcoming in this area are not generally seen as o ffering unambiguous support". Although the term module is widely debated, Paradis (2004) uses this term to pertain to neurofunctional modularity, which he argues is at a di fferent level of abstraction. Here, a module is a “dedicated, structurally independent system, whose internal structure is impervious to the influence of other systems, from which it may nevertheless receive inputs, and to which it may provide outputs" (Paradis 2004, 120). He acknowledges that the term carries a lot of historical baggage and suggests possible substitutes, such as “ ‘isolable subcomponents’, or a combination of mega-system, system, subsystem and sub-subsystem" (Paradis 2004, 120). 2.6 Interim Summary The mental lexicon is a highly complex system of interconnected information. Its organisation, especially if more than one language is present, is equally complex and is influenced by a number of factors, for example language proficiency, the nature of language acquisition and personal experience. All these factors lead to changes over time, making the mental lexicon a dynamic system. Such a dynamic model of the mental lexicon has certain implications for SLA, with the key points being the associations and the activation of given items (De Bot et al. 2005). The nature of a dynamic system and the implications for foreign language teaching will be discussed in the following chapter. 34 <?page no="51"?> Chapter 3 Dynamic Systems Theory In many SLA studies, the learner ’s development is viewed as a linear development (for a review, see Lightbown & Spada 2006). The development is seen to take place in consecutive steps that gradually lead from no knowledge of the L2 to native speaker like proficiency. In contrast to this view, De Bot et al. (2005, 2007), however, have suggested applying the Dynamic Systems Theory (DST) to second language acquisition (SLA) because a wide range of factors that are intricately interrelated and interacting a ffect the process and outcome of SLA for each individual learner. DST was originally developed in biology, but has since been applied to many other disciplines (Larsen-Freeman & Cameron 2008, 2f). In linguistics, it was first used to describe first language acquisition and is currently also used to describe SLA (De Bot et al. 2005, 14). Although the theory is naturally still vague in places, it nevertheless is of great interest to SLA, especially in the Cognitive Lingustic approach (see Chapter 4). DST sees systems as ‘nested’, that is, larger systems contain subsystems. For language, therefore, there is language in general, consisting of separate languages, dialects or varieties spoken by groups of people, and within these language systems we can see sub-systems such as the phonemic, lexical and grammar systems. Then each individual has his own language system, which may contain di fferent sub-systems, such as di fferent languages, varieties, registers, each again with its own phonemic, lexical and grammar systems. (De Bot et al. 2005, 16) 35 <?page no="52"?> The result is a complex system where all subsystems and sub-subsystems are connected by direct or indirect links. A change in one system is likely to e ffect change in another. It will have a direct e ffect when the second system uses the output of the first, and an indirect e ffect if the connection goes via other systems. A complex system, for example a learner, has a set of interacting variables. Due to this complete interconnectedness, it is not possible to exactly calculate the outcome over time. Because of this complexity, describing the system as a whole is di ffi cult. It is highly variable, as described above, and therefore ever changing. It may also undergo times of internal self-reorganisation (De Bot et al. 2007). There are, however, also periods of stability, when only an extreme outside force will be able to change the system. These periods, also called attractor states, are configurations in subsystems that lead to overall stability. An example of this is the morphological formation of the past tense. It goes through various stages, which are attractor states in themselves (De Bot et al. 2005, 17). Another, non-linguistic, example given by De Bot et al. (2007, 8) is that of a horse’s gait: it either gallops or trots, but there is no in-between stage. Attractor states are generally preferred by the systems, but are not necessarily predictable. The opposite, not preferred, states are sometimes referred to as repellent states. In SLA, attractor states can be formed within subsystems before the target form is attained. This is the same e ffect known as ‘fossilisation’ (see e.g. Selinker 1972; Lightbown & Spada 2006) outside DST. Fossilisation displays all the main characteristics of an attractor state. It is a stable state that the system (in this case the learner) seems to prefer over other states. It is di ffi cult for the system to overcome this state and it usually takes a strong (external) force to achieve this (De Bot et al. 2005). Dynamic systems seem to be highly dependent on their initial states, that is, slight di fferences in the first configuration may lead to dramatic consequences later. This is sometimes also referred to as the butterfly e ffect: a small local e ffect having a global impact (De Bot et al. 2007). 3.1 Growth in Dynamic Systems Theory Researchers have for a long time tried to model language learning, many of which can be placed under the umbrella of DST (De Bot et al. 2007, 11). One important role is given to resources that aid growth. Growth in (L1) language learning is a process that is “concerned with the increase or decrease (i.e. negative increase) of one or more properties" (Van Geert 1995, 314). In order to facilitate the process of growth, there must be 36 <?page no="53"?> resources. These resources can be divided into internal and external resources: • internal resources, resources within the learning individual: the capacity to learn, time to learn, internal informational resources such as conceptual knowledge, and motivational resources; and • external resources, resources outside the learning individual: spatial environments to explore, time invested by the environment to support learning, external informational resources such as the language used by the environment, motivational resources such as reinforcement by the environment and material resources such as books, and TV’s. (De Bot et al. 2007, 11) Both internal and external resources are limited, but there are also compensatory relations between di fferent types of resources (De Bot et al. 2007, 12). Within the learning system there are subsystems that draw upon resources to a varying degree. Some of the subsystems even support each other ’s growth, for example, increased listening comprehension stimulates the development of lexical skills, while at the same time knowing more words helps to understand language more easily (De Bot et al. 2007, 12). The di fference in initial states and di fferent growth levels, depending on the level of available resources at a given time, might lead to di fferent growth patterns. This is important for language learning, as the same learning operation can have very di fferent outcomes depending on the starting point and learning rates. Therefore, similar learning procedures could lead to diverging patterns of development, which might also be gradual or be sudden changes, for example due to a restructuring of the system (De Bot et al. 2007, 13f). 3.2 Dynamic Systems Theory and Second Language Acquisition For studies in SLA it is very di ffi cult to give insight into the interacting factors a ffecting the process and it is verging on impossible to predict the outcome (De Bot et al. 2007). One example of the intricate way factors may be interrelated is the e ffect of phonemic coding ability. First of all, this has an e ffect on the acquisition of the first language, as phonological 37 <?page no="54"?> awareness is a good predictor of reading acquisition in the L1 (see Sparks et al. 2000; De Bot et al. 2007). Similarly, problems with the perception of sounds may lead to L2 fossilisation. If we assume some similarities between the L1 and L2 acquisition processes, it is very well possible that in the initial stages of L2 learning, the L2 learner is apt to store misperceived sounds, which quickly become entrenched in his or her speech and reach an attractor state that is di ffi cult to get out of. (De Bot et al. 2007, 15) In the light of data on fossilisation and other phenomena, it seems likely that the di fferent languages an individual knows are part of one dynamic system only and therefore interact. Not only is the L2 a ffected by the L1, for example, in pronunciation, but the L1 may also show signs of interaction with the L2 subsystem. One study summarized by De Bot et al. (2005, 23) showed that bilingual speakers tend to slow down in the speed with which they could name objects when compared to monolingual speakers. One of the most important paradigms of DST in SLA is that “causal factors need not be mutually exclusive" (De Bot et al. 2007, 17). A more complex way of studying L2 development has, therefore, to be developed. 3.3 Interim Summary Dynamic Systems Theory o ffers a new perspective on language and first and second language acquisition. As outlined above, the mental lexicon in itself is a complex system nested within a larger system. These systems are shaped and changed by the use of language, either through input from the outside or international reorganisation processes triggered by language use. 38 <?page no="55"?> Chapter 4 Cognitive Linguistics and Foreign Language Teaching 4.1 Defining Cognitive Linguistics Cognitive Linguistics is a flexible framework rather than a single theory of language. [...] It constitutes a cluster of many partially overlapping approaches rather than a single well-defined theory that identifies in an all-or-none fashion whether something belongs to Cognitive Linguistics or not. (Geeraerts 2006, 2; see also Figure 4.1) The term cognitive in Cognitive Linguistics stresses that “language is a psychologically real phenomenon" and that “the processing and storage of information is a crucial design feature of language" (Geeraerts 2006, 3). Cognitive Linguistics can be described by four specific characteristics (ibid., 4f): 1. Linguistic meaning is perspectival 2. Linguistic meaning is dynamic and flexible 3. Linguistic meaning is encyclopedic and non-autonomous 4. Linguistic meaning is based on usage and experience The first point suggests that the world is construed in a particular way and that language embodies a perspective on the world (ibid., 4). Secondly, 39 <?page no="56"?> Figure 4.1: A Conceptual Map of Cognitive Linguistics (Geeraerts 2006, 19) meaning is connected to our shaping of the world, we adapt, for example, our semantic categories according to circumstance and leave room for nuance (ibid.). Thirdly, the meaning we construct through language is not separate from the rest of our mind, but rather reflects our overall experience as human beings (ibid., 5) and fourthly, meaning is essentially grounded, which means that it is rooted in experience (see also Langacker 2008). Cognitive Linguistics has a relevance for SLA, as it incorporates a usage-based model of language structure (Barlow & Kemmer 2000; Langacker 2008). Every language consists of an inventory of linguistic units abstracted from usage events (Langacker 2008). Kemmer & Barlow (2000, viii) describe this as the “intimate relation between linguistic structure and instances of use". The term linguistic 40 <?page no="57"?> structure is in itself ambiguous, as it can refer to both the external linguistic system derived by analysing linguistic data and to structures an analyst posits for the internal linguistic system (Kemmer & Barlow 2000). From a cognitively oriented point of view, the emphasis will always be on the second position. In usage-based models it is assumed that the linguistic system of a speaker is fundamentally grounded in usage events (for a collection of papers on usage-based models see e.g. Barlow & Kemmer 2000). Usage events are instances in which a person produces or understands language. Linguistic representation is linked to usage events in three ways: • Usage events form the basis of a speaker ’s linguistic system; • There is a direct relationship between the abstract representation of a person’s grammar and the usage events experienced by this person; • Usage events are necessary for the ongoing operation and structuring of the linguistic system. (cf. Kemmer & Barlow 2000, viii-ix) Therefore, usage events can be described as “the specific communicative events in which people learn and use language" (Tomasello 2000, 61) and “instances of language use in all their complexity and specifity" (Langacker 2008, 457). These events do not have a particular size, it depends on the analytical purposes as to how they are segmented (Langacker 2008). The relationship between the abstract representation and the actual instance of use is particularly important, as units of language are not fixed but dynamic (see Chapter 3). They are therefore perpetually being reshaped and restructured, depending on use in both comprehension and production. “Thus, usage events play a double role in the system: they both result from, and also shape, the linguistic system in a kind of feedback loop" (Kemmer & Barlow 2000, ix). In usage-based models, grammar is only derivative, arising from language use only, without the need for an underlying, possible innate grammar (Langacker 2000; Tomasello 2000, 2003). The availability of an item depends on the frequency and recentness of its activation (Paradis 2004, 28). Consequently, the more often an item is used, the more likely it is to be retained over a longer period of time. “A language needs to be used in order to keep its activation threshold 41 <?page no="58"?> su ffi ciently low to prevent accessibility problems." (Paradis 2004, 31). This notion is also acknowledged in traditional approaches to vocabulary learning, where it is assumed that an item needs to be encountered a number of times to be fully acquired (see Section 7.3.1). Thus, usage events are important for vocabulary acquisition because ideally they also constitute acquisition events (Meara 1997). Many researchers maintain that the mental lexicon only contains declarative knowledge (Levelt 1989; Scherfer 1997). However, in accordance with the usage-based approach, procedural knowledge is also seen as playing an important role. Following the argumentation of Mc- Neill (1979), Möhle (1997, 48f) holds the view that mental elements, that is, concepts, are linked in a procedural fashion. It is therefore not sufficient to know how a language works (declarative knowledge). A language learner also needs to acquire how language is used (procedural knowledge). Although McNeill’s theory is one relating to first language acquisition, Möhle (1997, 49) suggests that second language acquisition may be enhanced the use of complex expressions that serve to convey the learner ’s intention in a real communicative situation. In order to develop an L2 mental lexicon, it is no longer su ffi cient to learn vocabulary lists and focus solely on declarative knowledge. 4.2 Implications of Cognitive Linguistics for Vocabulary Learning Evidence that language consists of prefabricated formulaic sequences (also called chunks, see Section 8.1.4 for a detailed discussion) leads to a believe, in Cognitive Linguistics (CL), that language in use is based largely on memorised symbolic items ranging continuum-fashion (a) in size, between bound morphemes (like -ette), words (like cat), phrasal expressions (like once in a blue moon) and complete sentences (like It takes one to know one) and (b) in “substantiveness" (Croft 2001), between fully specified forms (like -ette, cat, once in a blue moon and It takes one to know one) and schematic ones like the -er, the -er and would you mind -ing? ) with syntax being, developmentally, an emergent phenomenon (Tomasello 2003). (Boers & Lindstromberg 2008c, 7) 42 <?page no="59"?> Additionally, the implications of CL that language is an integral part of cognition and not a separate faculty (see Tomasello 2000; Geeraerts 2006; Langacker 2008) lead to the conclusion that “general theories of cognitive processing will pertain to learning a foreign language, in particular, theories which concern memory" (Boers & Lindstromberg 2008c, 11; see also Chapter 6). Furthermore, if a “word’s meaning is all the events, contexts, uses, and so on that it can be associated with" (Verspoor 2008, 265) and the linguistics system is seen as interconnected with “other cognitive systems such as perception, cognition and emotion, this means that the meaning of a word also involves perceptual, cognitive and emotional association" (ibid.). Approaches to vocabulary learning inspired by CL are, in particular, good for encouraging deep processing (see Section 6.2.2), thereby heightening the probability of words and phrases being committed to memory (Boers & Lindstromberg 2008c). In terms of vocabulary acquisition from a CL perspective, “deep processing is believed to take place when the learner actively performs a relatively complex mental operation with regard to the lexical information" (ibid., 12). These elaborations can be semantic and / or structural in nature. Semantic elaborations are mainly concerned with connecting new items to already existing knowledge or encountering an item in a meaningful context. Structural elaborations, on the other hand, refer to the formal properties of a word (ibid.). In adopting a CL approach to FLT, the focus is on the study of linguistic motivation, as it “is commonplace for the meanings of linguistic forms to be motivated by language users’ experience of their physical, social and cultural surroundings" (ibid., 18), that is, extra-linguistic motivation, and by the properties of a language as such, that is, intra-linguistic motivation. The pervasiveness of linguistic motivation leads to at least two consequences for SLA and FLT: (a) Learners who are aware that an L2 is much more than a system of arbitrary form-meaning connections may be relatively likely to adopt mnemonically fruitful practices of insightful learning rather than less e ffective ones associated with blind memorisation. (b) Students who appreciate that language has considerable rhyme and reason may enjoy a ffective benefits such as heightened selfconfidence. (Boers & Lindstromberg 2008c, 18) 43 <?page no="60"?> Furthermore, drawing learners’ attention to the motivation and systematic nature of language will help learners to develop rich connotations and associations in the L2. This in turn will aid retention by creating a profound understanding of the target language (ibid., 26f). The pedagogical e ffectiveness of CL-inspired approaches to language learning has been explored in a number of studies in recent years (Achard & Niemeier 2004; Tyler et al. 2005; Boers & Lindstromberg 2008a; see also Verspoor & Lowie 2003; Boers et al. 2007). It could be shown that CL is “e ffective at teaching (high-frequency) polysemous words (mostly prepositions and particles) and figurative phrases (mostly idioms)" (Boers & Lindstromberg 2008c, 41). Beréndi et al. (2008, 87), for example, could show that explicit instruction of “the metaphorical motivation behind the meaning extension of polysemous word ... and the meaning of figurative idioms may aid comprehension and retention". Skoufaki (2008) investigated under which conditions conceputal metaphors can aid the semantisation of idioms. This piece of research draws on earlier studies that investigated the e ffectiveness of grouping new vocabulary by conceptual metaphors (see e.g. Boers 2000). This line of research suggests that a categorisation, based on conceptual metaphors, is very e ffective for language learning (Skoufaki 2008, 104). Condon (2008, 153) discusses a study on the teaching of phrasal verbs that suggests that: 1. spending time on CL insights can be of benefit in the classroom for the purposes of learning English phrasal verbs; 2. a CL approach can successfully be integrated with a pre-existing language learning programme so as to yield superior results for longer-term learning rather than short-term learning; and 3. a CL approach seems particulary to suit lexis that instantiates fairly concrete and transparent CL motivations. Boers & Lindstromberg (2008b) give a description of a series of classroom activities that implement research findings of CL. These mainly consider learners of intermediate or high proficiency and focus on idioms, but nevertheless give an interesting insight into how to translate research into classroom methodology (this also being the objective of this study). Verspoor (2008, 264) focuses on the implications of bilingual word associations, arguing that “at the initial stages of L2 acquisition, the learners assume more or less full overlap between the conceptual content of the L1 word and the L2 word" (see also Lowie & Verspoor 2004). However, 44 <?page no="61"?> meaning and conceptual realizations are highly complex. It could, therefore, be di ffi cult for learners to identify to what extent there is indeed an overlap of cognates or whether these are significantly di fferent. Direct instruction, combined with awareness raising tasks, may help learners to explore this. A similar approach is advocated by Niemeier (2005, 2008) who also outlines awareness-raising activities which help learners comprehend grammatical phenomena. Piquer Píriz (2008, 221) focuses on young learners and argues that if speakers master the core meanings of words that are frequently used in English and are aware of the regular principles of meaning extension, mainly metaphor and metonymy, they will tend to be able to understand and perhaps produce these words with other, related meanings. This goes beyond the argument of Kemmer & Barlow (2000), outlined above (see Section 4.1), who maintain that a language learner does not need to master all the meanings of words, but requires usage events for particular combination of words from which they abstract usage patterns. In first language acquisition (FLA), children develop linguistic abstraction by using what is known as pivot schemas: “Often, there is one word or phrase that seems to structure the utterance in the sense that it determines the speech act function of the utterance as a whole ..., with the other linguistic item(s) simply filling in variable slot(s)" (Tomasello 2003, 114). These pivot schemas are used to generate new sentences, for example when children use new object labels in existing schemas. These pivot schemas also o ffer possibilities for the implementation in the EFL classroom. Research suggests that FLA and SLA may have similar underlying processes (Cameron 2001; Brewster et al. 2002; Lightbown & Spada 2006). According to Boers & Lindstromberg (2008c, 27), a CL approach to vocabulary teaching hypothesizes that it will help learners attain a more profound understanding of the target language, better remember words and phrases (owing to greater depth of processing in general and to a dual coding in particular), appreciate the link between language and culture, and become more confident [...]. All of the issues will be addressed in the research outlined in Chapter 11. 45 <?page no="62"?> 4.3 Interim Summary The Cognitive Linguistic perspective of foreign language learning in general and the acquisition of vocabulary in particular is a growing research area. The implications of CL for FLT are manifold and promising, o ffering both ease of learning and a more profound knowledge of the target language. However, this has not yet been adopted by course books currently used in schools. Very few of these implications have had an impact that goes beyond a general discussion in language teaching methodology. There are, however, some books available that, for example, o ffer a systematic treatment of phrasal verbs (Rudzka-Ostyn 2003) or are aimed at raising learners’ metaphor awareness (Lazar 2003). 46 <?page no="63"?> Vocabulary Learning and Teaching <?page no="65"?> Chapter 5 Current Issues in Vocabulary Research Although a much cited anecdote is that a traveller would not take grammar books to a country where the language was unknown, but would bring a dictionary, the teaching, learning and acquiring of vocabulary has for a long time been treated like the ugly duckling of language learning. It was regarded as unfashionable and many were of the opinion that vocabulary acquisition would take care of itself (Nation 1990; Moir & Nation 2008). In recent years, however, it has been recognized that vocabulary plays an important role when learning a language (Schmitt 2000; Nation 2001; Meara 2002; Thornbury 2002; Folse 2004; Nation & Gu 2007). It is especially interesting to note highly diverging views on the importance of vocabulary held by researchers, teachers and learners. Regarding the latter two, Coady (1997a, 274) compiled the following list: In general, students feel that words are very important and are eager to learn them. In contrast, teachers tend to feel that words are easy to learn - grammar is the challenge. Many teachers and scholars feel that teaching vocabulary is a low-level intellectual activity unworthy of their full attention. [...] [M]any teachers [...] seem to conclude that words are going to be learned naturally from reading and do not need to be taught. 49 <?page no="66"?> Therefore, it becomes much clearer why teachers do not see the need for vocabulary teaching in spite of the students’ request for it. Or, if the teachers do see the need for some vocabulary instruction, it is of a temporary, bridging nature until the students can do it on their own. Vocabulary learning raises a number of question that will be addressed in more detail below, among these that “learners need to know what vocabulary to learn, how to go about learning it, and how to assess and monitor their progress" (Moir & Nation 2008, 159). 5.1 What is it to Know a Word? 5.1.1 What is a Word? The term word, although most people intuitively know what is and what is not a word, is hard to define. One possibility would be to base a definition on orthography, defining a word as “any sequence of letters (and a limited number of other characteristics such as hyphen and apostrophe) bound on either side by a space or a punctuation mark" (Carter 1998, 4). This definition, frequently used when working with electronic corpora, is suitable under some circumstances, for example while playing Scrabble. For the purpose of this study, however, it does not su ffi ce, as it has a number of theoretical and practical problems (Carter 1998; Folse 2004) because it does not take into account larger units of language. These units, however, form an important part of the vocabulary of the language being learned. In vocabulary teaching, the units are often referred to as vocabulary items 6 instead, as it might be composed of more than a single word. Other terms used to count the items of a language are tokens, types, lemmas (see also Section 2.1) and word families (Nation 2001 and Nation & Gu 2007; see also Schmitt 2000, Folse 2004). A token is a counting unit. When all the words in a given text are counted, even if the same word occurs more than once, it is the tokens being counted (Nation & Gu 2007, 18). Tokens are used to estimate how many words are in a given line, usually together with another counting unit, the type. 6 In the following, the terms vocabulary unit or vocabulary item will be preferred, although it is not always possible to stick to this terminology. 50 <?page no="67"?> Type counts do not count the same word twice, but only di fferent words are counted, for example to answer the question how many words you would need to know to understand a given book (Nation 2001, 7). A number of problems arise when counting types, one of which is that the same word form may have more than one meaning, if it is a polysemous and / or polyfunctional word. “For example, well can mean healthy, good, a place to draw water from, an exclamation and so on" (Nation & Gu 2007, 19). Another term often used in the literature, when considering vocabulary, is that of word families. A word family is usually defined as consisting of a headword, its inflected forms and closely related derived forms (Thornbury 2002; Nation & Gu 2007). One problem when using the unit word families is immediately obvious: how to decide “what should be included in a word family and what should not" (Nation 2001, 8). According to Thornbury (2002, 5), research suggests that the brain groups di fferent forms of the same word together and he therefore concludes that, despite the problem of delineating one word family from another, “it makes more sense to talk about [...] word families" than using other terms. Similar to word families is the lemma, a term predominantly used in lexicography, which consists of “a headword and some of its inflected and reduced [..] forms" (Nation 2001, 7). Generally, all items included in a lemma are the same part of speech, therefore walk as a noun and walk as a verb are not the same lemma, whereas variant spelling is often included in the same lemma (Nation 2001). The question whether to include irregular forms in the lemma is one of the problems with this term (Nation 2001). As already mentioned, vocabulary items often consist of more than one word, known as multi-word items, formulaic sequences or chunks (see also Section 8.1.4). They have subcategories such as collocations, phrasal verbs, fixed (sometimes also called set) phrases and semi-fixed phrases and idioms (Folse 2004). These multi-word items behave like a single word (Thornbury 2002). In order to incorporate this, the term lexeme is used. A lexeme is “a word or group of words that functions as a single meaning unit" (Thornbury 2002, 7). The di fferent uses of the terms lexeme and lemma are also discussed in more detail in Section 2.1. When children acquire words in their first language (L1), they map words onto concepts (Thornbury 2002). 51 <?page no="68"?> 5.1.2 What is Involved in Knowing a Word What is actually involved in knowing a word is another highly complex issue. There are a number of aspects that are involved in knowing a word completely. However, not every individual knows every word of his or her L1 in every detail. In this section, an attempt will be made to describe the most important aspects of lexical knowledge, applicable to the L1 and L2. On the most basic level, knowing a word involves knowing: • its form, and • its meaning. (Thornbury 2002, 15) As “[w]ords are not isolated units of language, but fit into many interlocked systems and levels" (Nation 2001, 23), it is important to have a closer look what it takes to know a particular word. It has to be borne in mind, as mentioned above, that a vocabulary item does not always consist of only one word. In his book, Nation (2001, 27) gives a description of what is involved in knowing a word (see Table 5.1). Not all of these aspects are fully learned on the very first encounter of a word, even if it is deliberately taught. Nation (2001, 82) gives three reasons for this: 1. There are numerous things to know about a word and its use (see Table 5.1); 2. The knowledge of a word needs to develop by di fferent activities; 3. Learners are not able to deal with all the aspects of a word at once; too much information confuses them. Active vocabulary teaching today is not necessarily confined to the foreign language classroom. There is a growing awareness that explicit instruction may also be beneficial in an L1 classroom (see e.g. Van de Guchte & Vermeer 2003). When learners encounter a new vocabulary item in a classroom setting, it is assumed that all these aspects are new to them (Singleton 2000). These di fferent aspects also have an impact on the kind of presentation of vocabulary items (Nation 2001, 33). Di fferent aspects call for di fferent ways of introducing and learning a word. Some aspects may be deliberately excluded, as they might hinder the retention of vocabulary items (see e.g. Folse 2004). 52 <?page no="69"?> Form Spoken Form r a What does the word sound like? p How is the word pronounced? Written Form r What does the word look like? p How is the word written and spelt? Position Grammatical position r In what patterns does the word occur? p In what patterns must we use the word? Collocations r What words or types of words can be expected before or after the word? p What words or types of words must we use with this word? Function Frequency r How common is the word? p How often should the word be used? Appropriateness r Where would we expect to meet this word? p Where can this word be used? Meaning Concept r What does this word mean? p What word should be used to express this meaning? Associations r What other words does this word make us think of? p What other words could be used instead of this one? a r = receptive knowledge, p = productive knowledge Table 5.1: What is Involved in Knowing a Word (Nation 2001, 23) 53 <?page no="70"?> 5.2 Breadth and Depth of Vocabulary Knowledge The question of the breadth and depth of vocabulary knowledge (Nation & Gu 2007) is directly linked to how many words an L2 learner should learn and whether all items have to be acquired in their entirety, that is, in every aspect of knowledge outlined in Table 5.1. The number of items known in a target language are often given in terms of the number of word families a learner should know. This is seen to be a good way to measure the breadth of target language vocabulary knowledge. For example, a learner who already knows the word sensible is thought not to have any great di ffi culty in understanding the derived form sensibly (ibid.). The concept of word families also is dependent on the proficiency of the learner in question, as “what might be a sensible word family for one learner may be beyond another learner ’s present level of proficiency" (ibid.). A good starting point, however, seems to be the most elementary and transparent members. Nation (2001) states that while it might ultimately be the goal of a foreign language learner to know as many words as a native speaker 7 it is neither an essential nor a short-term goal. Breadth of vocabulary knowledge alone is not the only parameter that should be explored when discussing vocabulary knowledge. As vocabulary knowledge is not an all-or-nothing phenomenon, the depth of knowledge should also be taken into account. A possible way of categorizing knowledge is outlined in Table 5.1. However, not all elements described are necessary to use a word successfully in a given situation or to understand it in context. Schoonen & Verhallen (2008, 212) argue that depth of vocabulary knowledge is important because “[l]exical knowledge is more than matching a word form with the right picture or the right synonym". Depth of lexical knowledge is often seen to pertain primarily to individual words and not the lexicon as a whole. This view is criticized by Meara & Wolter (2004) who point out that another feature of the mental lexicon has to be considered, namely the interconnections between items. Di fferences in lexical knowledge can also be seen as di fferent connections in the number and type of relations in the lexicon (Meara 1997). 7 Although how many words or word families a native speaker knows on average is also open to discussion, see for example Nation (2001). 54 <?page no="71"?> 5.3 Active vs Passive Vocabulary Knowledge Meara (1997) suggests that one of the crucial distinctions between active and passive vocabulary lies in the directionality of the connection. If a word is only connected to the rest of the lexicon in a uni-directional way, that is without information flowing back, then it would be a passive item. Active vocabulary items, on the other hand, “are connected to their parent lexicons by more than one type of connection" (Meara 1997, 119). The active vocabulary, therefore, can be easily activated by other words “because it has many incoming and outgoing links with other words" (Nation 2001, 25). Passive vocabulary items might only be activated by external stimuli, “[t]hat is, they are activated by hearing or seeing the form" (ibid.). This distinction is also often called the receptive / productive distinction, referring to the receptive skills of listening and reading and the productive skills of speaking and writing (Thornbury 2002). Again, the terminology is not without problems, as some researchers do not see why reading and listening should be subsumed under the heading passive, as this suggests that the learner does not bring anything to these tasks (Nation 2001). Furthermore, the use of the terms active and passive suggests a clear delineation between the two kinds of vocabulary knowledge that does, in fact, not exist. Rather, there are degrees of knowing a word and this can be treated as a scale of knowledge (Nation 2001, 25). Corson (1995), for example, holds a view, which is strongly based on language use, that passive vocabulary includes the active vocabulary and at least three other kinds of vocabulary knowledge, one for words that are only partly known, words that are for some reason avoided in active use (perhaps because their pronunciation is di ffi cult), and low frequency words that are almost never used in production (see also Nation 2001). The more often an item is encountered by a learner, the more likely it is to form new connections from the old to new knowledge gleaned from these repeated encounters. In this way, the word is moved along the scale away from the passive towards active vocabulary. Another aspect that is interesting in terms of vocabulary knowledge is the degree of control over word knowledge a learner has regarding a given item. The significant factors here are whether a learner can easily access the item or whether s / he has to have a prompt or a particularly rich context to remember a given word (Thornbury 2002). This again might be due to the number of connections a given item has made with 55 <?page no="72"?> other units. The more widely connected an item is, the more likely is its activation. The activation threshold will then be low due to frequent prior activation. A high degree of control is sometimes equated with automaticity, meaning “the extent to which learners can automatically retrieve and produce words in language use" (Nation & Gu 2007, 84). 5.4 Interim Summary The words a learner acquires in a foreign language has to meet many needs. It is hard for a teacher to predict which words will be especially useful and interesting for a given learner and whether these items will then be known in all the aspects useful for the learner, that is, whether the knowledge is deep enough. Lastly, learners should have a high degree of control over the most important items, resulting in easy access and rapid understanding. How this can be achieved will be the topic of the following chapter. 56 <?page no="73"?> Chapter 6 Learning New Vocabulary In order to be able to remember and use all the aspects of knowledge outlined in Section 5.1, the information must be successfully stored in the learner ’s memory. Language learning is part of a larger dynamic learning system, the basics of which will be discussed below (6.1; see also Chapter 3). Hypotheses on how language in general, and vocabulary in particular, can be learned successfully will also be outlined (see Sections 6.2 and 6). The implications of these hypotheses for actual language teaching will be addressed in Chapter 7. As outlined in Chapter 2, the models reviewed propose lemmas that contain a large amount of information. In addition, in network models, these lemmas interact and are a ffected by activation. When learning new vocabulary items, information has to be incorporated in the system, associations between items have be established and / or strengthened (De Bot et al. 2005, 51). These points are addressed by language learning theories discussed in the following sections. When new words are learned, regardless of whether this is in school or in other settings, the lexical knowledge is expanded in two directions, breadth and depth (see Section 5.2). Breadth refers to the quantitative knowledge, while depth describes the gain in meaning and meaning relations, enriching the understanding of already familiar words (Schoonen & Verhallen 2008). 57 <?page no="74"?> 6.1 The Basis for Learning “At the root of learning is the process of making meaning out of participation in the social world" (Cameron 2001, 241). This definition is frequently used in connection with language learning. From a psychological perspective, learning is defined more broadly as “[a]ll relatively permanent changes in behavior that result from experience, but that are not caused by fatigue, maturation, drugs, injury, or disease" (Lefrançois 2006, 419). At the heart of learning lies the committing of information to memory. 6.1.1 Levels of Memory In psychology, a three component model of memory, also known as the modal model, is widely accepted (see e.g. Lefrançois 2006, Chapter 9). It distinguishes between three components, sensory memory, relating to sensations, short-term memory (STM) and long-term memory (LTM), both of which are of importance for language acquisition, in both the L1 and L2. Table 6.1 shows an outline of this model. Information first enters the sensory memory component, from which it can go to STM. Some of the information might then go on to be stored in LTM, where it can take the form of concepts and from where it is available for retrieval by the short-term memory component (Lefrançois 2006, 300). In order to acquire a new word, it needs to be stored in the LTM, a “kind of filing system" (Thornbury 2002, 24), which has an enormous capacity and is durable, although this does not mean that information cannot be lost. In terms of vocabulary learning, this loss is called attrition (see e.g. Thornbury 2002, De Bot et al. 2005). Memory and its subcomponents play an important role in language learning (Schmitt 2000). Ellis (1996) found evidence that shows that short-term memory capacity is a good predictor of eventual vocabulary achievement. Research into what is remembered and what is forgotten, especially with regard to vocabulary acquisition, suggests a number of principles that may aid long-term retention. The most important principles are mentioned below (see Section 6.3, a more comprehensive list can be found in Thornbury 2002). 58 <?page no="75"?> Sensory Short-Term Long-Term Alternate Label Echoic or iconic Primary or working Secondary Duration Less than 1 second Temporary, less than 20 seconds Permanent, indefinite Stability Fleeting Easily disrupted Not easily disrupted Capacity Limited Limited (7 ± 2 items) Unlimited General Characteristics Momentary, unconscious impression; a passing sensation or association What we are paying attention to; immediate consciousness; active; maintained by rehearsal All our knowledge; passive; the result of encoding, storage and retrieval of information Table 6.1: Three Levels of Memory (Lefrançois 2006, 309) Baddeley’s Model of Working Memory The model of the working memory (a synonym for STM, see Table 6.1) proposed by Baddeley & Hitch (1974) has had an impact on the theory of language learning (Ellis 2001). Within the working memory, Baddeley & Hitch identified three subcomponents: the central executive, a system that oversees the entire process, and two so-called slave systems, the phonological loop and the visual-spatial sketchpad 8 (Lefrançois 2006, 304). The slave systems “maintain (as a sort of loop) the e ffects of sensory stimulation so that the central executive might have access to them" (Lefrançois 2006, 304). In language processing, the phonological loop is responsible for verbal material and the visual-spatial loop deals with visual input, that is, writing (see Figure 6.1). The phonological loop itself consists of two components, the phonological store, which represents material in a phonological code, and the articulatory rehearsal process, which can be used to refresh decaying phonological representations to maintain memory items (Ellis 2001, 33f). 8 Sometimes also referred to as visuo-spatial sketchpad, see for example Ellis (2001) or Randall (2007). 59 <?page no="76"?> The second slave system, the visual-spatial sketchpad, generates images, thus maintaining information, and manipulates information with visual or spatial dimensions, for example by rotating perspective or orienting through (mental) space (Ellis 2001, 34). Figure 6.1: A Model of Working Memory (Lefrançois 2006, 305) Empirical evidence, particularly from studies using this dual task paradigm, support the existence of separate systems, showing that there is only minor interference when subjects are asked to learn a list of words (a task involving the executive system) while retaining a sequence of numbers at the same time (a task for one of the slave systems) or studies that look at brain activity during tasks engaging STM (Lefrançois 2006; for a more detailed discussion see Randall 2007). In this model, the central executive is both responsible for the transfer into long-term storage and retrieval from it. 6.1.2 Long-Term Memory According to Lefrançois (2006, 307f), LTM has four particularly important characteristics: (1) Long-term memory is highly stable. (2) Long-term memory is generative. (3) Understanding influences long-term memory. (4) Some things are more easily remembered. 60 <?page no="77"?> LTM is the place where all educational experiences and knowledge of language is stored, therefore these four characteristics have an impact on L2 learning. Every piece of information that has been transferred into LTM will be remembered for a long time (see point 1). Studies show that many memories remain stored for longer than any tested intervals (Thornbury 2002; Lefrançois 2006). However, our long-term memory does not merely store information, it also constructs new information (point 2) based on preconceived notions and beliefs, for example, schemas. A schema is a unit in cognitive structure (Lefrançois 2006, 424). In Cognitive Linguistics, a prominent type is the image schema, “schematized patterns of activity abstracted from everyday bodily experience, especially pertaining to vision, space, motion, and force" (Langacker 2008, 32). Thus, even if a certain piece of information was not present in the input, it may be supplied by schemas from past experiences. Furthermore, our memories are rarely a reproduction of linear events, rather, the main points or most important facts are remembered. Our personal memory of a given situation is generated, based on our understanding of it (see point 3). This memory is based on scenes, frames, and scripts (Geeraerts 2006). Lastly, some memories, often called flashbulb memories (point 4), seem to be more prominent than others, often because they represent a very emotional situation. One example of this: most people alive at the time have a clear recollection of what they were doing when they heard that Kennedy had been shot (Lefrançois 2006). Unlike retrieval from STM, retrieval from LTM requires a search and therefore is slower (Thornbury 2002, 131; Lefrançois 2006, 309). Most current models of LTM are associationistic ones, using a variety of abstract concepts to metaphorically explain how the mind is structured, usually stipulating interconnections between representations (Lefrançois 2006, 314f; see also Section 2.3.1 and Figure 2.2). One objective of vocabulary learning and teaching must, therefore, be to transfer lexical information from STM to the permanent LTM (Thornbury 2002, 131). The Depth of Processing Hypothesis, which addresses precisely this objective, will be discussed in Section 6.2.2. The term chunking in connection with memory refers to “the development of permanent sets of associative connections in long-term storage and is the process that underlies the attainment of automaticity and fluency in language" (Ellis 2001, 38). A chunk itself is a unit of memory, that may be connected to other already stored chunks, thus building struc- 61 <?page no="78"?> tures. This process may lead to hierarchical organisation of memory and is ubiquitous. The higher a chunk is situated in the hierarchy of memory, the more sub-patterns are connected to it (Ellis 2001). For a more detailed discussion of the relevance of chunks in Early Foreign Language Teaching (EFLT), see Section 8.1.4. 6.2 Language Learning Language learning involves a number of di fferent learning processes. Aitchison (2003) states that in order to acquire word forms and the associated concepts in the L1 learners have to perform at least three tasks: a labelling task, a packaging task, and a network-building task (see Figure 6.2). This also applies to learning L2 words (see Verhallen & Schoonen 1998). Figure 6.2: Tasks Involved in Learning the Meaning of Words (adapted from Aitchison 2003, 189) In the labelling task, learners have to discover what sequences of sound are used to name things. During the packaging - or categorization task, as most researchers prefer to call it (see e.g. Verhallen & Schoonen 1998) - , they must find out which things can be grouped under one label. This task is essential in developing a concept. In the network-building task, relations between words are established (Aitchison 2003), leading to networks in which categorisation and abstraction become more and more important (Verhallen & Schoonen 1998; Cameron 2001). 62 <?page no="79"?> These tasks are also important for L2 acquisition, although the already present L1 lexicon may lead to di fferent network-building processes (see Chapter 2). For example, foreign language learners are more likely to master basic level terms before superordinate and subordinate terms because these terms represent the most frequently used words (Cameron 2001). In conceptual terms, basic level objects represent the highest level in which objects are similar in terms of shape and use. A single mental image can then be used to denote the category (Lako ff 1987). Therefore, when teaching new vocabulary for a certain topic, it is advisable to begin with the basic level items before moving up to more superordinate and general words or down to more specific, subordinate terms (Cameron 2001). 6.2.1 Vocabulary in Language Learning Learners do not usually commit a new item to memory after the very first encounter. Rather, it requires a process that takes place over time. Nation & Gu (2007, 85) subdivide this process into five stages: 1. encountering new words; 2. getting the word form; 3. getting the word meaning; 4. consolidating word form and meaning in memory; 5. using the word. Along similar lines, Schmitt (1997) developed a taxonomy of vocabulary learning strategies, consisting of discovery strategies and consolidation strategies, which in turn can be subdivided in a number of ways, for example memory strategies, cognitive strategies and metacognitive strategies. He details which strategy is used most or least and by whom, as well as discussing cultural di fferences in strategy use. Meara (1997, 118) suggests that L2 language acquisition consists of acquisition events, which he defines as consisting of “building of a connection between a newly encountered word, and a word that already exists in the learner ’s lexicon" (Meara 1997, 118). The advantage of these acquisition events is that they allow for the cumulative nature of vocabulary acquisition. As outlined above in Section 5.1, knowing a word is a process that usually involves many facets. Additionally, not every vocabulary item 63 <?page no="80"?> Direct teaching: Teacher explanation Peer teaching Direct learning: Study from word cards Dictionary use Incidental learning: Guessing from context in extensive reading Use in communication activities Planned encounters: Graded reading Vocabulary exercises Table 6.2: Ways of Learning and Teaching Vocabulary (Nation 2001, 16) will be acquired in its entirety. In accordance with Meara’s (ibid.) suggestions, unknown words do not have any connections to previous knowledge, whereas known words do have these connections. The number of these connections di fferentiates between poorly-known words and better-known words, having few and many connections respectively. Every encounter with an item is an acquisition event which adds up to a rich vocabulary, although Meara (1997) also concedes that individual acquisition events may be di ffi cult to identify. Nation (2001) provides a description of how new vocabulary items can be taught (see Table 6.2). These di fferent approaches and their application will be discussed in more detail below (Chapter 7). Furthermore, Coady (1997a, 282, drawing on research done by Stahl & Fairbanks 1986) states three principles that underlie any e ffective vocabulary teaching: 1. Providing learners with both definitional and contextual information about words. 2. Encouraging learners to process information about words at a deeper level. 3. Providing learners with multiple exposures to a word. 6.2.2 Depth of Processing Hypothesis The depth of processing hypothesis, first proposed by Craik & Lockhart (1972), states that the main di fference between STM and LTM lies in the way the input is processed. Depending on the level of this processing, information will be lost from STM or will be transferred to LTM. In the sensory register [...], no processing occurs. At the STM level, a “shallow" level of processing occurs, consisting mainly of the recog- 64 <?page no="81"?> nition of stimuli through perceptual analysis. With deeper processing (involving activities such as analysis, organization, and recognition of meaning), material is transferred to long-term memory and hence is not lost immediately. (Lefrançois 2006, 306) The depth of processing model for language learning uses the model suggested by Craik & Lockhart (1972), proposing that the durability of memory traces is a ffected by the depth of processing. Craik & Tulving (1975) later specified that retention is not merely determined by the presence or absence of encoding, but by the richness of encoding. A more elaborate encoding results in a more durable connection, that is, the deeper the processing, the better the learning (see Sökmen 1997; Laufer & Hulstijn 2001; Folse 2004; Lefrançois 2006). These elaborations and processes indicate that a word needs to be manipulated and thought about by the learner, preferably using already existing mental information to which it can be linked. The more a learner engages with a word, the better it will be retained (Schmitt 2000). This does not only mean repetition, because new information needs to be thought about which helps to make associations (Randall 2007, 169). Sökmen (1997) states that although repetition will also lead to retention, richer levels of encoding will result in better learning. These richer levels can be achieved, for example, when “students are asked to manipulate words, relate them to other words and to their own experiences, and then to justify their choices, these word associations are reinforced" (Sökmen 1997, 242). Classroom activities facilitating deep processing are time-consuming and hard work for the students, but these provide the foundation for truly acquiring a given item (Ellis 1997a). Other researchers, for example Thornbury (2002, 25), state that cognitive depth lies in the number of decisions a learner has to make about a word. The more cognitively demanding and the more numerous these decisions are, the better the word is remembered. The depth of processing concept is not without problems. As Randall (2007, 135) points out, it lacks clarity and there are contradictory results from studies which employ di fferent levels of processing. In a similar vein, Laufer & Hulstijn (2001, 6) argue that all research in this area is faced with the need for “an unambiguous, operationalisable definition of any notion proposed as a replacement for depth of processing". In order to remedy this situation to some extent, they propose a “motivational-cognitive construct of involvement, consisting of 65 <?page no="82"?> three basic components: need, search, and evaluation" (Laufer & Hulstijn 2001, 17). This construct can be implemented by devising tasks that vary in regard to the degree of need, search, and evaluation (ibid.). The need component of this hypothesis refers to the “need to achieve" (ibid, 14), and, therefore, the motivational dimension of involvement. Search and evaluation are the two cognitive, that is, information processing, dimensions of involvement, referring to the search for the meaning of an unknown word and the comparison of a word or meaning with other words or meanings, respectively (ibid). Looking at these three dimensions of involvement, the involvement load of a given task can be stated. The Involvement Load Hypothesis predicts that tasks with a higher load will result in better retention than those having a lower involvement load. This hypothesis, however, is also in need of “more precise definitions of the involvement components and a more thorough theoretical link between them and theories of information processing and a ffective components of cognition" (ibid, 22f, see also Nation 2001). In an article investigating consciousness-raising tasks, Eckerth (2008) argues that, as “language acquisition can be speeded by explicit instruction" (Ellis 2002, 175), this explicit knowledge about language has an e ffect on “input perception, language processing, and output monitoring which can be conducive to second language acquisition" (Eckerth 2008, 120). Ellis (1997b, 160) defines a consciousness-raising task as: a pedagogic activity where the learners are provided with L2 data in some form and required to perform some operation on or with it, the purpose of which is to arrive at an explicit understanding of some linguistic property or properties of the target language. This definition takes the depth of processing hypothesis further, as the elaborations are not only meant to result in long-term memorization but also provide explicit knowledge about the target language. This in turn facilitates better retention. The tasks are “designed to encourage learners to reflect upon the L2 and to discover aspects of L2 form, meaning, and function" (Eckerth 2008, 121f). 6.2.3 Incidental Vocabulary Learning Many authors make a distinction between incidental vocabulary acquisition and intentional (or explicit) vocabulary acquisition (see e.g. Schmitt 2000; Singleton 2000; Hulstijn 2001; Nation 2001). These two main processes are assumed to be di fferent in the following way: “explicit learning 66 <?page no="83"?> through the focused study of words and incidental learning through exposure when one’s attention is focused on the use of language, rather than the learning itself" (Schmitt 2000, 116). In a research context, vocabulary acquisition is regarded as being incidental if the participants are not informed that they have to take a test on vocabulary knowledge at a later time (Laufer & Hulstijn 2001). Incidental learning, sometimes also referred to as ‘guessing from context’, was for a long time favoured by foreign language teachers, as it requires only minimal attention from the teacher. Nation (2001, 232), for example, calls it “the most important of all sources of vocabulary learning". Basically, all the learners have to do, and therefore the only thing a good teacher has to ensure, is that learners read extensively in the target language and are provided with appropriate texts. One of the main reasons that this approach to vocabulary learning was favoured lies in the fact that many people argue that, as L1 acquisition of lexical knowledge is largely based on ‘picking up words’ from context, it should be the same for L2 learners. Evidence from recent research, however, seems to indicate that relying on incidental vocabulary learning as an integral part of L2 instruction is highly problematic. Folse (2004) reviews a number of studies, looking mainly at adult learners of intermediate or higher L2 proficiency, and concludes that this type of learning is very limited. First of all, learners tend to ignore unfamiliar words altogether, unless the unknown word seems to contribute a great deal of information to the reading task. Secondly, the correctness of the guesses leaves a lot to be desired, leading at worst to faulty learning (see Folse 2004). The individual studies vary in design and variables such as level of proficiency and presentation of target vocabulary items (e.g., in some studies, the words for one group were given an L1 translation equivalent, while the other group of subjects were given explanatory glosses in the L2), or trained in the use of dictionaries and other strategies. Folse (2004, 76f) additionally points out that context in a real world setting “is often not very clear in terms of revealing the meaning of the word if the reader really does not know the word". Context is very important for SLA, but it should be of a kind that o ffers enough clues for the learner to base an educated guess on. Nation (2001, 237) refers to a meta-analysis of studies on incidental vocabulary learning in native speakers which found that an average of 15% of unknown words are incidentally learned while reading a text that has 3%, or less, of unknown words in it. Another problem with incidental vocabulary learning is that, in order 67 <?page no="84"?> to be able to do this, a learner must have a good knowledge of the target language already, otherwise he or she cannot guess at all. In order to overcome this problem, in lessons for beginners the context provided must be especially rich and geared towards the students’ needs (Nation 2001). Otherwise, it leads to no learning at all. Folse (2004) concludes that guessing from context has a place in L2 instruction, as it is a valuable strategy for improving reading comprehension. However, it should be taught as a strategy for reading comprehension and not seen as a way to broaden pupils’ vocabulary knowledge because learning rates can be increased by deliberate attention being focused on vocabulary as opposed to only inferring this knowledge from context alone (Nation 2001). For a detailed discussion of what can be learned from context, see for example Laufer & Hulstijn (2001); Nation (2001); Folse (2004). 6.2.4 Implicit versus Explicit As outlined above, it has been argued that, as children acquire their mother tongue without any instruction, that is, language is acquired without paying special attention to vocabulary, then it is the best, perhaps even the most natural and most important mechanism in regard to learning an L2. Many studies have been conducted on incidental vocabulary acquisition through reading (for a review see Singleton 1999; Nagy & Herman 1985; Folse 2004). The distinction between implicit and explicit vocabulary learning is not to be regarded as an either-or relation, but rather a continuum (Carter 1998). Ellis (1995) describes four hypotheses that can also be seen as points along this continuum: • A strong implicit learning hypothesis holds that words are acquired largely by unconscious means. • A weak implicit learning hypothesis holds that words cannot be learned without at least some noticing or consciousness that it is a new word which is being learned. • A weak explicit learning hypothesis holds that learners are basically active processors of information and that a range of strategies are used to infer the meaning of a word, usually with reference to the context in which it appears. • A strong explicit learning hypothesis holds that a range of metacognitive strategies are necessary for vocabulary learning. In particular, the greater the depth of processing involved in 68 <?page no="85"?> the learning, the more secure and long term the learning is likely to be. (see also Carter 1998, 2001) The fourth hypothesis has a direct link to the Depth of Processing Hypothesis by Craik & Lockhart (1972) outlined in Section 6.2.2. The opposite view that implicit learning is the fundamental way of knowledge acquisition is maintained by Krashen (1981), who makes a clear distinction between learning and acquisition. For language learner beginners, most researchers are of the opinion that the very first vocabulary needs to be explicitly taught, as it is only after the learners have a certain repertoire of L2 words at their disposal that they will be able to acquire new words incidentally (Carter 1998). 6.2.5 Item versus System Knowledge Nation (2001, 23) di fferentiates between item knowledge and system knowledge. Item knowledge, for example, involves memorizing the form of a word. On the other hand, system knowledge means that the underlying rules, for example, sound-spelling correspondences, are learned instead of the individual spelling of a given word. Item learning allows the learner to "build up their resources without having to pay much attention to how the items [...] are organized" (Ellis 1999, 474) and occurs at all levels of L2 development. System learning entails abstraction, as rules are constructed and interrelationships between these rules are established. Both item and system knowledge are processes necessary for L2 acquisition (Ellis 1999), but they should be carefully balanced (Nation 2001). “There is no doubt that attention to form and rules must be supported and prepared for by experience with the items in use" (Nation 2001, 59). In terms of vocabulary learning, this can be translated into di fferent approaches to SLA: the item-based approach, where each vocabulary unit is learned as an individual item, no matter how complex it may be. This is, at first at least, a memory task, but as the learners’ lexicons expand, the items are internally reorganized. From this reorganisation, rules emerge (Thornbury 2002). This approach is probably closest to the way in which words are acquired naturally (Thornbury 2002; Tomasello 2003). Similarly, Ellis (1994, 1995) argues for less attention to formal features. Instead, the meaning should be the main focus, rules will then 69 <?page no="86"?> develop through the experiences that arise when learners use a linguistic item (see also Nation 2001). The rule-based approach, which corresponds to system learning, isolates and highlights any relevant patterns or regularities. The rules underlying the target language are explicitly taught. This leads to a number of problems. For example, the scope of a given rule is not always clear (Thornbury 2002, 108). 6.2.6 Rote Learning In conventional approaches to vocabulary learning, words are learned in lists of paired words, also called pair associates (Carter 1998, 193). These word lists can often be found at the back of a course book, usually consisting of a target item on the left side and an L1 translation equivalent on the right. A short example sentence using the item is sometimes provided, too, but this is quite rare. These lists are what many people associate with vocabulary learning (and teaching, at least to a certain extent). A common task for homework is to memorize the new words of a unit discussed in class. There are several computer programmes available to make memorisation easier. A popular one in German schools is, for example, Phase 6 (see www.phase-6.de). The type of learning involved here is also called rote learning. Folse (2004, 39) states that for beginners, the huge amount of vocabulary learned at the beginning “would invariably feature learning words from lists". Unfortunately, he does not explain why he has come to this conclusion. Using word lists is an atomistic technique (Singleton 1999, 50) that leads to the isolation of items. They are decontextualized and thus left without the important information that is required to know a word in depth (see Section 5.2). Schmitt (2000, 24) states that simply reciting an item seems to have little e ffect. If items are repeated in a more complex fashion, for example by providing an organising principle for the items to be learned, lexical items are retained more successfully. This view is directly linked to the Depth of Processing Hypothesis, as again the more deeply an item is processed, the better it will be remembered (see Section 6.2.2). Randall (2007, 169f) has a more positive view of rote learning, as it may involve more than just reciting. Rather, especially in the Confucian tradition, recitation involves thinking it over. Randall argues that it does not only rely on “mindless repetition" (ibid., 169) but also incorporates two additional principles, depth of processing and practice. 70 <?page no="87"?> However, “learning new words is more than the acquisition of isolated lexical units: new words are embedded in a lexical network which means that all kinds of connections with related words have to be established" (Schoonen & Verhallen 2008; see also Meara 1997; Aitchison 2003; Meara & Wolter 2004). Therefore, pure rote learning may be successful up to a point, but it should not be the sole learning principle on which vocabulary learning is based. 6.3 General Processes which aid Vocabulary Retention 6.3.1 Noticing, Retrieving & Using Vocabulary Nation (2001) lists three major processes that help to reach the learning goal, that is, the retention the words in long term memory: Noticing, retrieval and creative use. Other processes that help learners to learn a vocabulary item successfully will be addressed in Chapter 7. Noticing The first process, noticing, involves paying attention to an item, a factor that has been studied in research on incidental vocabulary learning, especially for reading comprehension. The salience of the word plays an important role. More generally speaking, noticing can be facilitated by formal instruction, the easiest way being that the teacher deliberately spends time discussing a certain item in more detail. Nation (2001, 63) also cites motivation and interest as important factors, as these have an immediate impact on the learner ’s attention. If a certain word is not perceived as useful, it is more likely to be quickly forgotten than words that are of importance or interest to the target audience. Nation (2001) states that one negative feature of noticing, especially in a formal setting, is that it often requires decontextualisation because the focus shifts away from the message and towards the language item as a part of the language system. Decontextualisation in this case does not mean that there is no context at all, but rather that an item is removed temporarily from the context in order to focus on its function, form and / or meaning. One way of doing this is giving a definition of a word or negotiating the meaning of a given item (Nation 2001; Folse 2004). Schmidt (2000) even proposes a noticing hypothesis, stating that nothing can be learned unless it has been noticed. Sometimes, vocabulary 71 <?page no="88"?> items have to be brought to learners’ attention to make them salient and thus leading to their learning (Lightbown & Spada 2006). This can be achieved by a number of vocabulary exercises, for example an ‘odd-man out’ activity where learners have to find the word, in a group of words, that does not belong to the group (Folse 2004). Noticing can also be promoted by highlighting certain words in a given text or by writing them on the board (ibid.). In a beginners’ course, the items chosen for further activities are automatically noticed. If a story is told, certain words can be made more salient by preor post-teaching activities (Cameron 2001; Brewster et al. 2002). Retrieving a Vocabulary Item Retrieving a word is one of the major processes that may lead to an item being remembered because “if [a] word is subsequently retrieved during the task then the memory of that word will be strengthened" (Nation 2001, 67). Retrieval, like vocabulary knowledge, can be divided into receptive retrieval, which involves perceiving a word and retrieving its meaning, and active retrieval when the learner wants to communicate the meaning of a given item of vocabulary. It has been suggested that each retrieval strengthens the connection between form and meaning (Nation 2001, 66f). Taking this argument one step further, we can also argue that repeated retrieval also strengthens the connections between words in a spreading activation network, therefore aiding the deep processing (see Section 6.2.2) and raising the level of activation (see Section 2.3.1). Creative Use The third process that leads to vocabulary retention is creative use, sometimes also referred to as generative use (Nation 2001, 68). This involves using words that a learner has previously encountered in a di fferent way. “At its most striking, the new meeting with the word forces learners to reconceptualise their knowledge of that word" (Nation 2001, 68f). A word may be used in its metaphorical extension, but also other variations in terms of inflection or collocations, for example. These ways are seen as beneficial. Joe (1995) shows that the knowledge of a word is related to the degree of generative use (see also Nation 2001). Generative use can lead to elaborations of a word, which in turn leads to new meaning connections and deepens the level of processing. For example, while negotiating meaning, a vocabulary item is not only noticed but also generatively used, because it is occurs “in a variety of 72 <?page no="89"?> grammatical contexts, often in a variety of inflected or derived forms, and often with reference to a variety of instances" (Nation 2001, 69). Generative use seems to be especially beneficial to long-term retention, but only if it does not occur in the form of close repetition of the previously encountered context, but rather is used in a creative and new way (ibid.). The processes of noticing, retrieving, and generating are also regarded by some researchers as being the variables that influence depth of processing. Therefore, each item a learner needs to remember should be go through all of these stages (ibid.). 6.3.2 Involvement Laufer & Hulstijn (2001) suggest that three components play an important part in long-term retention of a word: need, search, and elaboration. While the first component of the three is within the a ffective domain, the latter two refer to information processing. In order to achieve a high involvement load as described in the Involvement Load Hypothesis (see Section 6.2.2), the tasks employed during language learning have to generate a genuine need to find the meaning of a given word as well as encourage learners to search for the meaning. This can be done, for example, by looking it up in a dictionary, and by evaluating the word, for example, by comparing it with other possible words that could be used in this task. The examples Laufer & Hulstijn (2001) use, in their article on the Involvement Load Hypothesis, however, refer primarily to the incidental learning 9 of vocabulary through reading. This is only partly transferable to the teaching of beginners who have no or little literacy in the L2. However, the Involvement Load Hypothesis is a valuable theoretical framework to classify tasks used in language teaching in order to predict their e ffect on vocabulary retention. In view of the needs of beginners, the motivational aspect plays an important role. The learners have to feel a genuine need to learn a given item, for example because it is needed in a task they have to complete. Furthermore, the comparison of words can be instigated by additional vocabulary exercises that encourage learners to make decisions about a word and therefore evaluate it. A number of exercises that aim at exactly this are outlined in Chapter 11. 9 In this context, incidental learning means that participants had no foreknowledge of the fact that their recall was tested afterwards. 73 <?page no="90"?> 6.4 Interim Summary The knowledge of vocabulary has many dimensions. Apart from the depth and breadth of knowledge it is also necessary to take the degree of control and involvement of the learners in a given task into account. Additionally, the various types of learning, for example, incidental learning and explicit learning have an impact on the way a learner commits a given item to memory. The implications of the issues raised in this chapter for the teaching of vocabulary will be discussed in more depth in the following chapter. 74 <?page no="91"?> Chapter 7 Teaching Vocabulary 7.1 The Beginner’s Paradox One striking finding when reviewing the relevant literature in the field of vocabulary acquisition is that many of the recommended ways to broaden L2 vocabulary are only relevant for intermediate or proficient second language learners. This is especially true in the area of incidental vocabulary acquisition through extensive reading. Coady (1997b, 229) puts it rather succinctly: “How can they learn enough vocabulary through extensive reading when they do not know enough words to read well? " The same argument applies to other areas of vocabulary acquisition. Laufer (1997, 23) states that a learner needs a vocabulary size of about 3,000 word families to be able to carry out reading comprehension tasks. A word family by her definition contains a base form plus its inflected and derived forms, leading to a total amount of 5,000 lexical items. During each term in a German school roughly 500 to 700 new lexical items are introduced and also supposedly learned by the pupils (Do ff & Klippel 2007, 51). Therefore, in order to reach the estimated threshold of 3,000 word families to read well enough to acquire vocabulary incidentally through reading, learners have to broaden to their lexical knowledge through other means beforehand. Schmitt (2000, 145) believes that “with rank beginners, it is probably necessary to explicitly teach all words until students have enough vocabulary to start making use of the unknown words they meet in context". The following sections will address the ways of teaching vocabulary to learners in more detail, paying special attention to young language 75 <?page no="92"?> learners 10 because they are the subjects of the intervention study conducted. 7.2 Presenting New Vocabulary Beginners are generally believed to require direct teaching, therefore the method of presentation of new vocabulary items is crucial. As outlined in Section 6.2, learners need to encounter a word, connect it to a meaning and have to integrate it with existing knowledge. These aspects have to be addressed in language instruction. The labelling task, for instance, requires defining a new vocabulary item. This can be done in a number of ways, for example, by guessing the meaning from the context provided. However, as beginners have a very limited knowledge of the L2, the teacher needs to provide them with additional clues. Nation (2001, 85) describes a number of ways in which meaning can be communicated: • by performing actions • by showing objects • by showing pictures or diagrams • by defining in the first language (translation) • by defining in the second language • by providing language context clues When choosing a way to present meaning, the teacher has to consider the reason for explaining and the degree to which the way of explaining represents the required meaning of a word (Nation 2001). For example, pictures used should be clear and prototypcial (see e.g. Rosch 1975; Aitchison 2003), so that the possibility for misinterpretation is kept to a minimum. “A picture is not necessarily worth a thousand words, but one which clearly represents the underlying concept of the word undoubtedly is" (Nation 2001, 85). 10 Young language learners are defined as children from age 5 to 12 (see e.g. Rea-Dickins 2000). 76 <?page no="93"?> Defining an item in the L1, that is, the translation of a given item, has been widely criticised, as this encourages the perception that the two languages are equivalent (ibid., see also Krumm 1995). Furthermore, although learners seem to go through a phase in which their lexical knowledge is subordinate to or lexically mediated by the L1, studies have shown that this changes over time (see Chapter 2.4). Relying on translation over a prolonged period of time might, therefore, hinder the reorganisation of the learner ’s mental lexicon. Giving definitions in the target language is an additional way used to convey meaning, in particular if the word is abstract and therefore di ffi cult to mime or draw. Verbal means of clarifying meaning include giving a definition using other words, for example synonyms, antonyms or superordinate terms, or by providing a sample situation or example sentences including the target word (Thornbury 2002). The above ways of conveying meaning are very important in the primary school English classroom (see e.g. Cameron 2001; Brewster et al. 2002). Another important way, although not strictly a presentation of a word, is to elicit the meaning by representing the meaning, using a picture or other means, and asking the learners to supply the form (Thornbury 2002). Using this elicitation technique actively involves the learners and provides them with speaking opportunities, both of which are thought to be beneficial to vocabulary acquisition. Vocabulary learned in primary school is usually quite easy to convey using the means described above, as it is generally concrete and thus has a high degree of demonstrability (Brewster et al. 2002). Furthermore, one of the principles of EFLT is that the language used in the classroom should be of immediate interest to the learner (Niedersächsisches Kultusministerium [Lower Saxony Ministry of Education] 2006), therefore negotiating meaning is generally easy as the target words pertain to the children’s immediate surroundings and the here and now. However, presenting vocabulary items once in a way that helps learners to understand the meaning immediately is not the only factor that influences whether a given word is learned or not. 77 <?page no="94"?> 7.3 Teaching to Promote Long-Term Vocabulary Retention 7.3.1 Repeated Exposure to an Item In many books on vocabulary learning and teaching (e.g. Schmitt 2000; Nation 2001; Thornbury 2002; Folse 2004) it is stressed that learners need to be exposed to a word numerous times before it can be regarded as fully acquired. Nation (2001, 81) reviews several studies investigating the number of times that vocabulary items need to be repeated, concluding that “most learners required five to seven repetitions". However, he also states that this number can vary considerably, as some learners have been known to require over twenty repetitions to learn a group of six paired associates (ibid.). Sökmen (1997) conducted an extensive literature review and concludes that several studies using a range of 5-16 encounters with a word showed that many encounters are needed to truly acquire a given item (see also Nation 1990, 43 ff ). When teaching a foreign language to young learners, these repetitions are often referred to as ‘vocabulary recycling’ (e.g. Cameron 2001, 87). Recycling vocabulary at regular intervals leads to the items remaining active and in use. This is also acknowledged in most course books used in German primary schools. Most of them mention the benefits of repeated exposure to words over the years and often refer to this as the Spiralcurriculum, stating that this ensures the repetition of certain chunks and vocabulary items over the entire time a specific course book is used (see e.g. Hollbrügge & Kraaz 2003b, 2004b). 7.3.2 Richness Activities Nation & Newton (1997) outline several tasks that promote depth of processing, among these being richness activities. These activities “aim to increase the number of associations attached to a word" (ibid., 249) and can be subdivided into those that promote syntagmatic relationships and those that establish paradigmatic relationships (see also Nattinger & DeCarrico 1992; Lewis 1997b). Richness activities can take many forms, from exercises where learners match collocations to given items, usually provided in two lists or focus on one word only, asking learners to provide as many collocations as possible (see e.g. Nation & Newton 1997, 250; Lewis 1997a, 92 ff ; Na- 78 <?page no="95"?> tion 2001, 106, 160; Thornbury 2002, 118 ff ). These activities promote syntagmatic relations. Paradigmatic relationships can be developed by “matching and classification activities" (Nation & Newton 1997, 251). Among these are semantic mappings, as these associate an item with others of related meaning. In her book “Teaching Languages to Young Learners", Cameron (2001) describes a number of ways to make these strong memory connections, most of them concerning the organisation of vocabulary, for example in thematic sets or hierarchical relations (see also Van der Rhoer & Vermeer 2005). This has been debated in current literature, as research seems to indicate that learning vocabulary items in thematic sets, near synonyms or antonyms hinders vocabulary retention (Folse 2004). These findings, however, were especially apparent in psycholinguistic laboratory studies in which learners were presented with lists of nonsense words and their L1 equivalents (see e.g. Waring 1997), therefore their implications for language teaching remain uncertain. Sökmen (1997) also describes several ways to provide rich instruction, for example by describing target words until the students have grasped the meaning. While doing this, the meaning should not be given away immediately, because this encourages the students to process all the input. Nation (1990, 55) calls this the ‘What is it’ technique. He states that the teacher has to choose the examples carefully to make sure that the learners have to listen to several sentences and hear the words a number of times. This “inductive approach allows repetition of the word and encourages the learner to make an e ffort to the meaning" (Nation 1990, 55f). 7.3.3 Information Gap Activities Information gap activities, sometimes also referred to as split information tasks (see e.g. Nation 2001) or information transfer tasks (see e.g. Brewster et al. 2002), are often done in pair work. Each learner has unique, essential information, for example, provided by two di fferent pictures. The learners then describe their respective pictures and try to exchange information (Nation 2001). A variation of this task is to ask learners to carry out a survey or fill in a questionnaire or quiz “where they ask each other questions and listen carefully for the answer" (Brewster et al. 2002, 104). When teaching young learners, the teacher has to ensure that the questions the learners are supposed to ask are prepared and rehearsed 79 <?page no="96"?> adequately before the task is done (Cameron 2001). It might be helpful to provide learners with the adequate chunks necessary for this task and a way of taking notes of the answers because the writing skill in the foreign language is not fully developed yet. Information gap activities help learners to communicate meaningfully and freely, promoting oral fluency (Cameron 2001). Additionally, the communication aids the processing of vocabulary items, as additional information is exchanged and connected to the words the activity focuses on. 7.3.4 Structuring New Vocabulary Providing principles of organisation of words is thought to facilitate the integration of new items in the “complex web of connections in the mind" (Cameron 2001, 81). These structuring principles can take the form of sense relations (also called semantic relations or lexical relations), types of these relations being, for example, antonymy, synonymy, hyponymy, or meronymy (ibid.). Many of these relations can be illustrated by using tables or grids (Nation 1990) or semantic mappings (Nation 2001; Neveling 2005). These illustrations provide additional visual support for the learners, which is important in the EFLT classroom because writing plays only a minor role (Brewster et al. 2002). Additionally, the distinction between function words and content words can be used to draw the learners’ attention to language structure (ibid.). Another way to facilitate strong memory connections is to arrange things that go together in thematic sets. The benefits of structuring vocabulary will be addressed in more detail in Section 8.2 7.3.5 Integration of Old and New Knowledge The mental lexicon is a network of associations, “a web-like structure of interconnected links" (Sökmen 1997, 241). Therefore, L2 learners need to establish links between new items and the already existing knowledge to store new vocabulary e ffectively and to facilitate speedy retrieval (Sökmen 1997). Connecting new knowledge to already existing knowledge leads to categorisation and organisation. Knowing words in “‘generalized’ groups of semantically related items" should be encouraged as an important factor of L2 vocabulary learning (Carter 1998, 192). 80 <?page no="97"?> Nation & Gu (2007, 90) list this way of aiding long-term retention under vocabulary consolidation strategies, calling it ‘encoding strategy’ and define it as creating “a meaningful connection between what the learner knows and what the learner is learning". 7.4 Interim Summary Beginners require a di fferent approach to vocabulary teaching and learning. The incidental vocabulary acquisition through reading, a field in which much vocabulary research has been conducted, is clearly not a feasible way for broadening and deepening their vocabulary knowledge. Rather, new vocabulary items need to be introduced using more than definitions in the native language. In order to ensure that new vocabulary is committed to long term memory, several prerequisites have to be met. The most important ones for this study are the structuring of vocabulary, richness activities, and integration of old and new knowledge. 81 <?page no="99"?> Chapter 8 SLA and the Young Learner Research into SLA often equates “‘second’ language development with ‘adult language development" (Nicholas & Lightbown 2008, 28) and “[a]lthough children feature as participants in this research, it is relatively rare to find reviews or overviews of SLA that deal specifically with child SLA" (Philp et al. 2008, 3). This is debatable, as children feature as participants in a number of studies (for current examples, see e.g. Paradis 2007; Schoonen & Verhallen 2008). There are, however, many possible reasons for this perceived lack of research. One might be that doing research with young learners is more di ffi cult and requires more resources than a study in which the participants can be recruited from a university. 8.1 Knowing a Word in Primary School Pupils learning a foreign language at primary school do not necessarily have to master everything detailed in Table 5.1 (see page 53). The written form of a word, for example, only plays a minor role in Early Foreign Language Learning in Lower Saxony, Germany. Instead, more importance is placed on oral communication and listening comprehension, restricting written input to the recognition of words and the copying of written phrases. Hatch & Brown (1995, 372) outline what they call essential steps in vocabulary learning (see also Section 7.2), based on findings from research on learners’ strategies: 83 <?page no="100"?> • having sources for encountering new words; • getting a clear image, whether visual or auditory or both, for the forms of the new words; • learning the meaning of the words; • making a strong memory connection between the forms and the meanings of the words; • using the words. These steps are essentially the same for young learners and adult learners. As described in Chapter 5.1, there are many details to be learned about a word to fully know it and “children’s capacities for learning change as they get older. So the learning of words is a process that continues, but that changes in nature as it continues" (Cameron 2001, 84). Cameron (2001) uses metaphors to illustrate this evolving process: learning words does not mean ticking o ffitems from a shopping list, but rather resembles cleaning a house, as it is not over after it has been done once only. Instead, vocabulary needs to be recycled at intervals, as recycling makes recall more probable (ibid). L1 learners acquire breadth and depth of knowledge about lexemes by experimenting with words in a large range of contexts. L2 learners, on the other hand, often have only very limited access to these opportunities (Carter 1998). Foreign language education should therefore aim at providing as many opportunities to experiment with language as possible. As concrete words are generally learned first (Carter 1998), they naturally form the basis of beginners’ vocabulary. For other reasons for choosing vocabulary for primary school settings, see Chapter 10.2. Selecting vocabulary items for the primary school L2 learners raises additional issues. Many factors influence how many words can be learned during the initial years, a realistic target seems to be around 500 words per year of instruction (Brewster et al. 2002, 81). As young learners have di fferent needs from the adult learner, the selection of vocabulary is influenced by factors such as demonstrability, brevity, regularity of form, centres of interest and learning load, among others (Brewster et al. 2002, 81f). Adult learners might not be interested in topics like dinosaurs, witches, or toys, but these play an important part in a child’s daily life and therefore feature prominently in coursebooks and curricula designed for EFLT. In some areas, the recommendations for selecting vocabulary items for young learners are in direct contrast to those for adult learners. The selection of vocabulary depends on the age of the learners and the methodology of teaching these depends on the level. Brewster et al. (2002), for 84 <?page no="101"?> example, maintain that vocabulary at primary school level must have a high level of demonstrability, meaning that it is easy to illustrate it using pictures, mimes or gestures, as these visual aids help learners to memorize these words. Furthermore, Cameron (2001) and Brewster et al. (2002) advocate the teaching in sets, be it lexical or rhyming / phonological sets, a practise that others argue leads to confusion and actually hinders memorization (see also Folse 2004). 8.1.1 Games in Early Second Language Acquisition Formal instruction in early foreign language learning “consists primarily of playful activities in which a few words, songs, and games are engaged in" (Nicholas & Lightbown 2008, 45), often for only a limited amount of time. In Germany, for example, a typical setting is two 45 minute-long periods of English a week 11 . In a survey outlined in Nicholas & Lightbown (2008, 45f) “[a]ctivities that involve listening, repetition, chanting, and singing were considered suitable for both ‘preschool’ and ‘primary’ learners", especially as they provide a link between home and school and thus between the ‘real world’ and the foreign language classroom (Brewster et al. 2002). Games are therefore often used in the primary English language classrooms, as children generally enjoy them and are motivated by taking part in them (Brewster et al. 2002). In order to prevent games from being mere time-fillers, the goals and focus of a game should be carefully thought through. The rules of a game can, for example, provide the background for practicing certain chunks or for recycling vocabulary. Furthermore, they can provide a template for meaningful communication that helps to develop the speaking skills of the learners and aids language processing and ultimately vocabulary retention (Brewster et al. 2002). Learners are involved in a group activity and can get help from their peers on the meaning of unknown language. Such group-based peer interaction provides good learning environment (Nation & Newton 1997). Games can be used to promote other skills, too, as learners can contribute to a game in di fferent ways. Some games mainly promote receptive skills, for example a Bingo game. Others require a number of skills, for example listening and speaking, communicative skills, and reading. Games in the foreign language classroom can be used to recycle already encountered words in new communicative situations (Niedersäch- 11 This depends on the Federal State, as schools fall under their jurisdiction. 85 <?page no="102"?> sisches Kultusministerium [Lower Saxony Ministry of Education] 1995), thereby promoting deep processing and long-term vocabulary retention. 8.1.2 Tasks in Foreign Language Learning A task in the context of instructed SLA can be defined “as an activity in which language is used meaningfully to achieve the goal of the activity, which is generally non-linguistic" (Oliver et al. 2008, 131). The CEF (Council of Europe 2001, 10) gives a slightly broader definition: “any purposeful action considered by an individual as necessary in order to achieve a given result in the context of a problem to be solved, an obligation to fulfil or an objective to be achieved". A good example of this, also used in the study outlined in Chapter 11, is the information gap activity (sometimes also referred to as split information tasks (Nation 2001)). Tasks are usually employed to provide a reason and motivation for the learners to communicate and therefore leads to the production of L2 utterances (Oliver et al. 2008). Tasks are often used for teaching young learners, as they “provide a range of di fferent types of interesting and engaging activities" (ibid, 131). They are not only intended to develop communicative skills, but also to contribute to the learners’ general language development, albeit often incidentally, depending on the task (Ellis 2003; Rubin & McCoy 2008). Task-based learning is generally believed to lead to higher motivation in learners, especially if the task is perceived to be relevant to their lives outside the classroom (Rubin & McCoy 2008). Tasks in themselves are highly complex and can be classified using a wide range of criteria (for an summary of this discussion, see Robinson 2007). The focus of task-based learning are tasks that “involve the specification not of a sequence of language items, but of a sequence of communicative tasks to be carried out in the target language" (Willis & Willis 2001, 173). The emphasis can also be the “real-ness in outcome" (Cameron 2001, 30), with learners working together to do things like working together to solve a problem, to play a game or to compare experiences, and so forth (Willis 1996). All this said, not every activity qualifies as a task. According to Cameron (2001, 31), the additional qualities that turn an activity into a task are unity and coherence. A task should have a clear beginning and end, a clear purpose and meaning and a clear language learning goal. She goes on to give an example that also has a particular relevance for 86 <?page no="103"?> the intervention study outlined in Part 3: [W]e can see that although any instance of singing is an activity, only more carefully planned and structured events using songs will be classed as ‘language learning task’. This perspective turns the song into a tool for language teaching and learning, that can be e ffectively planned, implemented and evaluated. (Cameron 2001, 31) In primary school, tasks have to be planned with extra care, because the attention span of the young learners and other needs of this age-group have to be kept in mind. 8.1.3 Sca ffolding In EFLT, tasks have to be sca ffolded, that is, the learners have to be supported by the teacher. The term sca ffolding was first used for the finedtuned verbal help caretakers provide during L1 acquisition (Cameron 2001). In experiments, it was found that parents do the following when sca ffolding e ffectively: • they made the children interested in the task; • they simplified the task, often by breaking it down into smaller steps; • they kept the children on track towards completing the task in reminding the child of what the goal was; • they pointed out what was important to do or showed the child other ways of doing parts of the tasks; • they controlled the child’s frustration during the task; • they demonstrated an idealised version of the task. (Cameron 2001, 8) These observations can be transferred to teacher-learner interaction in the classroom. One way of providing sca ffolding in the classroom is to provide routine, especially when it comes to classroom management. These routines can then be adapted over the course of time, keeping pace with the developmental changes, for example, longer attention spans as the children get older (Cameron 2001, 9 ff ). The familiarity of the routines helps learners to predict meaning and intention and these routines 87 <?page no="104"?> can again be extended to incorporate new elements and more complex language over time (Cameron 2001, 10). In order to sca ffold successfully in the EFLT classroom, it is important to be aware of the sub-goals because complex tasks need to be broken down into manageable steps. It is in the responsibility of the teacher to make sure that these steps add up to one complete, complex task (Cameron 2001, 28f). 8.1.4 The Importance of Chunks Prefabricated chunks of language 12 can be used by the learner to build a corpus of language to use in given situations (Randall 2007, 171). This corpus can then be used to generate underlying rules, without the need for explicit teaching. In the literature, there are many terms used for this linguistic phenomenon. Wray (2000, 465), for example, uses the term formulaic sequence and defines it as: a sequence, continuous or discontinuous, of words or other meaning elements, which is, or appears to be, prefabricated: that is, stored and retrieved whole from memory at the time of use, rather than being subject to generation or analysis by the language grammar. These sequences are treated like single “big words" in memory, as “the role of working memory in learning such structures is the same as for words" (Ellis 1996, 111). Conklin & Schmitt (2008) suggest that, as formulaic sequences play an important role in everyday language use, this component of the mental lexicon is enormous. Although there seems to be no study that has attempted to put a number to the quantity of formulaic sequences known by a proficient speaker, it is acknowledged by most that in order to become proficient, there must be knowledge and mastery of chunks at some level (Spöttl & McCarthy 2003; Conklin & Schmitt 2008). Wray (2000) discusses two main reasons for the existence of formulaicity (or chunks) in language, the first being the assumption that formulaic sequences save processing resources, as “there is less e ffort involved in this [i.e. slotting words into established chunks] than in creating the whole construction from scratch" (Wray 2000, 474). The second looks at formulaic sequences as acts of communication, classifying them by their socio-interactional function (see also Nattinger & DeCarrico 1992). 12 I will use the terms chunk, formulaic sequence, and multiword item interchangeably. For a discussion of the di fferent terms see Wray (1999, 2000, 2002). 88 <?page no="105"?> The socio-functional explanation of formulaic sequences explains their pervasiveness. They are very useful, as “they are often linked to a single meaning / pragmatic function, which gives them considerable semantic / pragmatic utility" (Conklin & Schmitt 2008, 73). Conklin & Schmitt (2008) go on to review a number of studies that approximately 20 percent to 58 percent of all discourse consists of chunks, variation depending on register and context. The psycholinguistic explanation, looking at processing e ffi ciency, is also of relevance for SLA. In e ffect, the mind uses an abundant resource (long-term memory) to store a number of prefabricated chunks of language that can be used ‘ready made’ in language production. This compensates for a limited resource (working memory), which can potentially be overloaded when generating language on-line from individual lexical items and syntactic / discourse rules. (Conklin & Schmitt 2008, 75) There is indirect support for this assumption, for example the fact that speech production poses a great cognitive challenge (De Bot 1992; Wray 2002; Kuiper 2004). In order to alleviate this load, the mind uses prefabricated chunks. This appears not only true for the L1, as Oppenheim (2000) found that the speech of her subjects using an L2 contained formulaic sequences with an overall mean of 66 percent. In situations where the speakers (regardless of whether they are using their L1 or L2) are under pressure, the reliance on chunks seems to be high (Conklin & Schmitt 2008). Empirical proof for the hypothesis that these sequences are stored as whole units is provided by a number of studies, the latest being the study by Conklin & Schmitt (2008). They also found that the processing advantage exists not only for native speakers, but also for proficient nonnative speakers, concluding that idiomatic expressions and other chunks, at a higher level of proficiency, are not more di ffi cult to understand than literal speech (ibid, 85). Wray & Perkins (2000) propose a way of accommodating the sociofunctional and psycholinguistic explanations within a single framework. When a speaker selects a formulaic sequence, one reason might be that he or she wants to ensure e ffi cient production by easing the pressure on processing. When, on the other hand, a speaker uses a chunk for interactional purposes, the main focus is often not on fluent and correct production, but on e ffi cient comprehension (Wray 2000). 89 <?page no="106"?> The relevance of chunks in EFL instruction is apparent. In both L1 and L2 acquisition, children seem to be able “to use word strings that they understand only at a holistic level, as input for a process of segmentation and analysis" (Wray 2000, 480). Additionally, Johnstone (2000) found substantial evidence of chunks of language in primary school learners. Lewis (1993, 1997b) outlined in his books both the theoretical basis for teaching a foreign language using chunks as well as providing suggestions of how this can be implemented into the language classroom. Chunks should also play an important role in early foreign language teaching. Prefabricated phrases form pivotal points that children can use to convey their communicative needs because they provide patterns with open slots that can be filled according to the communicative need (see also Chapter 4).They also help to ease the cognitive load during language processing, especially in situation where the learner is dealing with many cognitive tasks at once. “Young learners tend to rely on the formulaic system as they hear and see language in use; and through this they gain implicit knowledge of the language rules, and thus they are able to create their own structures and discourse" (McKay 2006, 38). Although learners will develop metacognitive awareness over the course of time, chunks play an important role throughout the process of SLA, as learners who “had little opportunity to draw on a formulaic system [...] quickly become tongue-tied and anxious as they try to construct a sentence in their head based on the rules they have learned" (McKay 2006, 37). Chunks are often presented to learners in “stories, songs, rhymes, dialogues, and through classroom language" (Cameron 2001, 50). In the initial stages of language learning, formulaic sequences play an important role for motivational reasons as well, as children memorize these chunks quickly and have the impression that they can understand and speak quite fluently from the start (Brewster et al. 2002). In consequence, simple greetings, classroom language, routines, and communication strategies (e.g. asking for further explanation) play an important role in the EFL classroom (Brewster et al. 2002, 105f). 8.1.5 The Role of Writing Learners in a primary school setting are still in the process of acquiring the writing system of their L1. Writing in the L2, therefore, plays a controversial role. While some researchers maintain that “literacy provides many substantial benefits for instructed learning, for example, allowing access to multiple kinds of input or providing a means for recording new 90 <?page no="107"?> vocabulary to be learned and practised" (Nicholas & Lightbown 2008), literacy in a second language “may lead learners to rely on the written form of the language as a guide to pronunciation rather than relying on aural input" (ibid.). This latter argument is one of the reasons why writing plays only a very small role of English language instruction in Germany primary schools. The written input is mostly limited to single words and simple sentences the children have to recognize but do not necessarily need to produce actively (Niedersächsisches Kultusministerium [Lower Saxony Ministry of Education] 2006). Learners are generally not asked to write productively without having sample sentences or words from which to copy the spelling. The course books used in the primary school classroom rely heavily on pictures, as a general approach is to “encourage speaking first and leave reading until later, especially if the pupils are still not fully literate in their L1" (Brewster et al. 2002, 110). On the other hand, writing should not be excluded entirely from the L2 classroom. One thing many learners might have already encountered outside the school is environmental print in English: print on T-shirts, food labels or advertisements, and so forth (Brewster et al. 2002). Examples of environmental print can be used to call attention to English orthography. Another approach that is often used is functional print. This refers to printed material displayed in the classroom, for example on posters, or labels that have been glued to equipment, and so on. This is one way to familiarizes learners with the written word without explicitly teaching writing skills (ibid.; Cameron 2001). According to Cameron (2001, 139), at approximately 10 years, the first language oracy and literacy are quite firmly established in children. Reading can then start to play a role in foreign language learning from this age onwards. Cameron (ibid.) states, however, that "the caveats about oral skills" must be remembered and only familiar vocabulary should be used in the initial stages 13 . 8.2 Enhancing Vocabulary Activities in Primary School In a number of methodological handbooks (e.g. Brewster et al. 2002), authors advocate a variety of possible activities that aid long-term voca- 13 A detailed overview on methodological approaches to teaching literacy skills can be found in Cameron (2001, Chapter 6). 91 <?page no="108"?> bulary retention. Brewster et al. (2002) mention classifying and sorting into categories as one way of practising vocabulary. For consolidating vocabulary knowledge, they take this one step further and suggest that teachers should “encourage pupils to build up their own word networks taking as their starting point a topic for example, Transport, and then writing down words they associate with this" (Brewster et al. 2002, 90). Cameron (2001, 87) mentions a number of possible vocabulary memorising activities which make central use of networks. As outlined in Part 1, lexical knowledge can be viewed as a network with many di fferent interconnections, the activation spread depending on the contextual clues. Activities and tasks that help to establish these connections therefore help learners to memorise an item. Therefore, words need to be personalised in a number of ways (Brewster et al. 2002; Thornbury 2002). This may be done by allowing the learners to make decisions about an item and providing activities where they can convey their decisions to other learners, thus creating authentic communicative situations. These communicative situation form the basis of usage-events, which in turn lead to acquisition events (see Chapter 4). If learners are provided with meaningful communication, they will eventually abstract underlying patterns inherent to the target language and gain a more profound knowledge of this language (Boers & Lindstromberg 2008a). 8.3 Interim Summary Although adult and young learners SLA may be similar in a number of ways, the specific needs of the young learners have to be taken into account when developing a series of lessons intended to facilitated vocabulary learning in primary school. The implications of Cognitive Linguistics and more traditional approaches have to be implemented in the EFLT classroom to facilitate acquisition of vocabulary items and enable learners to commit these items to long-term memory. 92 <?page no="109"?> Chapter 9 Measuring Vocabulary Knowledge Measuring the knowledge of learners is referred to as either evaluation, assessment, or testing. These terms are not necessarily synonymous, although they are sometimes used interchangeably (Iannou-Georgiou & Pavlou 2003). Whereas assessment and evaluation are general, global processes, testing is only one part of assessment (ibid.). Assessment in school can be a valuable tool for providing feedback on teaching and learning processes for the learners, teachers and parents (McKay 2006). Learners at least partially judge the importance of vocabulary items on whether it is part of a test or not (Schmitt 2000). This might not be as true for young learners, but even here this backwash (sometimes also referred to as the washback e ffect) can have a positive or negative impact on learning. If, for example, vocabulary is part of the language lessons but never commented upon or assessed, young learners might get the impression that vocabulary does not matter. When assessing language learning, teachers and assessors have to be aware of the role of knowledge in the social and cognitive processes during assessment. During the early stages of SLA, children rely on formulaic sequences (see Section 8.1.4), therefore the tasks chosen for assessment should enable children “to draw on their known formulae and vocabulary" (McKay 2006, 46) by containing familiar routines and repetitive elements, and so forth. As children’s abilities in the L2 increase, the tasks can go beyond the predictable. “Assessment and feedback need to evoke positive emotions in children about language learning, about 93 <?page no="110"?> themselves and about others" (ibid.). A skilled teacher assesses the development of his or her learners almost continually, using a number of di fferent assessment techniques, for example by observation, record-keeping or portfolio assessment (Cameron 2001). These methods, however, are not the main focus or relevant in the research described in Chapter 10, as the participants of the study described in Part 3 were only taught for six consecutive lessons. This left very little time for portfolio or other continuous assessment methods to be used. While observations were done using field notes, the main emphasis of the study lies in the testing of vocabulary knowledge directly after the intervention study and two weeks after the last class. This second test was carried out to measure long-term retention and produced unexpected results (see Sections 11.5.2 and 11.5.4). 9.1 Developing Assessment Tasks for Younger Learners Although EFLT has been widely implemented across the European Union for several years, the issues of assessment have not been addressed in any depth (Johnstone 2000). Although the CEF (Council of Europe 2001) provides a tool for developing assessment procedures, standardized tests are not currently widely available. This is partly due to the great variability of contexts among the relevant countries (and in Germany even within the country) and a lack of consensus of what EFLT proficiency might mean (Johnstone 2000, 123). As with any assessment, the following points need to be addressed when developing tests / assessment tools for language learners: • To what extent are results reliable? • To what extent is there construct validity? • To what extent is the assessment task authentic? • To what extent is assessment task interactive? • To what extent is the assessment practical? • To what extent is the impact of the assessment positive? (cf. McKay 2006, 113) McKay (2006, 113-117) then addresses these questions in turn. The question of reliability is important, as any test needs to be able to provide 94 <?page no="111"?> consistent results. Variables like the assessor or the teacher should have no influence on the outcome of the assessment task. A test, therefore, should yield the same results regardless of who marks it and give the same result for learners who have the same ability (Thornbury 2002). Construct validity addresses the issue of making sure that a test measures the ability it is supposed to measure. When constructing a test, the ability it measures has to be clearly defined and the scores of the tests have to be interpreted accordingly. In the light of a growing interest in authenticity in the classroom, tasks used for assessment should be as authentic as possible. The tasks have to be relevant and natural to the learners. Paper-and-pencil tests have little if any authenticity to the child’s world; writing an e-mail to pen-pals has some authenticity for upper elementary foreign language learners keen to find out about children their own age in the target country; playing a fun game in the foreign language has high authenticity for young learners. (McKay 2006, 115) The interactivity of tasks plays a role in assessment because interaction lies at the heart of any language knowledge and use. If a child is given the task to sort objects without having to use language in any way, either for comprehension or production, the task is not interactive and therefore not a suitable one to assess language skills. Another factor is the availability of resources. This can and does have an impact on the practicability of assessment tasks. A task that requires many resources may not be practical in a school setting, where time, space, and materials are often restricted. Finally, assessment should have a positive impact on the learners because their motivation and interest in the language should not be undermined by frequent and / or troublesome / di ffi cult tests. 9.2 Types of Vocabulary Tests Language tests can serve a variety of purposes: • to find out where learners are experiencing di ffi culty so that something can be done about it (diagnostic tests) • to see whether a recently studied group of words has been learned (short-term achievement tests) 95 <?page no="112"?> • to see whether a course has been successful in teaching particular words (long-term achievement tests) • to see how much vocabulary learners know (proficiency tests) (Nation 2001, 373) Short-term and long-term achievement tests are of particular interest here, as this was the test type used in the intervention study outlined in Part 3. These tests usually “comprise words that learners have been studying recently, usually within the last week or two" (Nation 2001, 374). The test items are, therefore, directly derived from the course material. The results of short-term achievement tests do not indicate the total vocabulary size of a learner, rather they focus on the success, or otherwise, of the recent study. Useful items to incorporate in such tests are word-meaning relations, translation equivalents (although these are not without problems, as discussed in Section 4.2), matching completion in sentence context and true / false items (ibid, 375). A more detailed discussion of the other test types can be found in Nation (2001), Iannou- Georgiou & Pavlou (2003) and McKay (2006). Long-term achievement tests were also used in the present study. These were administered two weeks after the first short-term achievement test. Whether this constitutes true long-term retention is open to discussion, as a longer delay might have yielded additional results. However, due to time and resource constraints, this could not be done. (For a more detailed discussion of the procedures of the study and the tests used, see Chapter 11.) Although beyond the scope of the present study, the aspects of breadth and depth of vocabulary can be measured using standardized tests, for example the Vocabulary Levels Test (Schmitt et al. 2001) and the Vocabulary Size Test (see e.g. Nation & Gu 2007). The Vocabulary Levels Test is often also employed to measure the total vocabulary size of non-native speakers of English, although it was not designed to do this (Nation & Gu 2007, 105). It is a test for receptive vocabulary knowledge only, as it requires subjects being tested to read the relevant items. A test that was specifically designed to measure depth of word knowledge is the WAT test developed by Schoonen & Verhallen (2008). Using this test, researchers are able to measure the depth of vocabulary knowledge in young L2 speakers. The target groups for this test are children of immigrant parents living in the Netherlands. It tests the knowledge of Dutch in these children, therefore this test, used for second language learners, cannot be directly transferred to a foreign language learning 96 <?page no="113"?> classroom, where the amount of input 14 is much more limited because learners do not live in the target culture. Johnstone (2000, 134) discusses a way out of this dilemma by outlining three data-collection procedures undertaken in schools in Scotland: 1. a systematic classroom observation instrument; 2. a paired-speaking task; and 3. a vocabulary retrieval test. In the context of the present study, the third procedure is of most interest. The learners in Scotland were asked to provide their associations in relation to topics with which they were familiar. These were entered into a computer by a researcher and shown to the learners, who were invited to add more associations. The rationale for this task rested on the assumption that there was a di fference between the language resource that was ‘available’ to pupils in their long-term memory and the language they would produce when under the constraints of engaging in meaningful realtime interaction. (Johnstone 2000, 135) This procedure is time consuming, as each learner has to be tested individually. As a tool for assessing range of vocabulary, however, and perhaps even vocabulary size and depth of vocabulary knowledge, the vocabulary retrieval test may be useful in future research (see also Read 2004a). In a similar fashion, it has been suggested that word webs compiled by learners are a useful tool for vocabulary assessment (Klippel 1995; Neveling 2004, 2005). Such word webs give researchers a valuable insight into the organisation of a learners’ vocabulary, but are again not suited to measure vocabulary knowledge in such a restricted study as the one outlined in Part 3. 9.2.1 Selected-response Tasks A common task in language assessment is the selected-response task, in which learners are expected to select a response from input. Typical 14 In the present case, English input in schools at the primary level in Lower Saxony is normally restricted to two 45-minute lessons a week. 97 <?page no="114"?> selected-response at primary school level tasks include “multiple-choice items, picture cloze and picture-matching vocabulary items, where children are given input [...] from which to select" (McKay 2006, 106). During this type of task, listening and reading skills can be assessed. Children may, for example, have to listen and select the appropriate picture, listen to a description and draw, colour, label or sequence pictures, listen to a story and tick items on a chart, and so forth. (Brewster et al. 2002). These tasks are very common in the assessment supplements of course books used in German primary schools, for example Making the Grade with Ginger (Bebermeier et al. 2004) and Show what you know (Becker et al. 2003) or course book independent tests, for example See what I can do in English (Ehlers et al. 2006). Based on this, they also form the foundation of the vocabulary tests used in the study outlined in Part 3. Selected-response tasks only assess parts of the knowledge that is involved in knowing a word (see Table 5.1), namely the receptive side of vocabulary knowledge, mainly limited to the form and meaning of an item. This is not unusual, as most “vocabulary tests target only one or two aspects of word knowledge" (Thornbury 2002, 130). With regard to measuring language competence in German primary schools, listening is seen as the most prominent skill to be developed and thus also assessed before all others because writing has only a minor role in primary English classrooms (Niedersächsisches Kultusministerium [Lower Saxony Ministry of Education] 2006, 6) and should therefore not form the basis for assessment (ibid., 12; 17). Additionally, receptive listening forms the basis for productive use of a language and should therefore be the focal point of EFLT, especially during the first year (in Lower Saxony Grade 3) (Niedersächsisches Kultusministerium [Lower Saxony Ministry of Education] 1995, Niedersächsisches Kultusministerium [Lower Saxony Ministry of Education] 2006). As mentioned above (see Part 1 and Schoonen & Verhallen 2008), words do not occur in isolation, they are part of a semantic network of related words. Therefore, selected-response tasks are not necessarily the most e ffective way to measure vocabulary knowledge, as they leave out the question of depth and the availability of said items in communicative situations. Vocabulary knowledge has many di fferent dimensions that have to be taken into account. Aside from sheer size, the depth of lexical knowledge is also very important and should be tested to obtain information on how well a word has been acquired and how many connections to other relevant knowledge has been established (Meara 2004; Meara & Wolter 2004; Schoonen & Verhallen 2008). However, in order 98 <?page no="115"?> to test vocabulary knowledge in these areas, new tests would have to be developed. This, however, was beyond the scope of this investigation. 9.3 Interim Summary Vocabulary knowledge can be assessed in many ways, but tests developed for young learners of English are still rare. Although several suggestions have been put forward, a standardized test that meets all requirements outlined above is not yet available. Since listening is the skill that is most prominently developed in German primary schools, selected response tests that measure listening comprehension were used in testing vocabulary knowledge in the study outlined in 11. 99 <?page no="117"?> Intervention Study in Primary School Classrooms <?page no="119"?> Chapter 10 From Theory to Practice 10.1 Implications for the Classroom At a time when research in the areas of the bilingual mental lexicon was still in its infancy, Meara (1993, 295) held the view that “[t]he study of the bilingual lexicon is just not well enough developed for it to be able to tell practitioners what to do in classrooms". He continued by saying that it might be regarded as “esoteric and di ffi cult" (ibid.) by hands-on language teachers. Over subsequent years, however, further research has made it a bit easier to use the research on the bilingual lexicon to develop lessons. In the intervention study outlined in Part 3, the objective was to do exactly this: to adapt laboratory findings and other research results in order to devise a way to teach vocabulary that might prove to be beneficial for foreign language learners. In the following I will summarize the implications that served as guiding principles while developing the lessons (see Chapters 11.1.1 and 11.1.2). Words are part of a structured and highly complex network (Meara 2004; Meara & Wolter 2004). Therefore, words should not be learned in isolation, but in the context of authentic texts / situations and learners should be encouraged to spend time on words and their elaborations, thereby taking part in usage-events (see Chapter 4). Furthermore, current research has gathered information on conditions that aid the comprehension and internalization of a new lexical item. One aspect that can be found throughout the literature is that in order to go beyond rudimentary knowledge, the learner needs to be exposed to the item in multiple ways, in varied and meaningful contexts (De Bot 103 <?page no="120"?> et al. 1997; Singleton 1999; Nation 2001). It may be the case that not all information posited in Chapter 5.1 can be extracted from the context during the first encounter with an item. This is perhaps the most common scenario. However, “[o]nce a lexical item has been added to the lexicon in a more or less primitive form [...], its continued use will gradually specify additional information" (De Bot et al. 1997, 317). As the two or more languages of a language learner are interconnected, sometimes very closely, at other times more distantly, all knowledge, whether linguistic or otherwise, is likewise interconnected. New items have to be integrated into the system, therefore old knowledge cannot be discounted (Scherfer 1997, 195). During the early learning process, subordinate organisation of the L2 mental lexicon seems to be the most likely type (see Chapter 2.4.1 and 2.4.2; Scherfer 1997), therefore the influence and role of the L1 needs to be taken into account. When introducing new vocabulary items, Scherfer (1997, 205) holds the view that the teacher should try to activate the already existing, underlying conceptual knowledge first, especially when the L1 and L2 equivalents are lexicalized di fferently, thereby ensuring that learners do not assume that there is a complete conceptual overlap in the two languages (Verspoor 2008). In an overview of vocabulary teaching in classrooms, McCarthy (1984, 14) describes an “obsession with items in isolation" and “entrenchment in word-and-definition" (McCarthy 1984, 21). This, however, is in the process of changing, as research in Applied and Cognitive Linguistics finds its way into textbooks and teacher education. 10.2 Choosing Vocabulary Items The words learned during the first two years of English language teaching in Germany are commonly defined by two entities. The first is the curriculum for the given Bundesland (Federal State) and the course book used, if any, which is based on the curriculum (Do ff & Klippel 2007). In Lower Saxony, the guideline published by the Ministry of Education is called the Core Curriculum (Kerncurriculum) (Niedersächsisches Kultusministerium [Lower Saxony Ministry of Education] 2006). Although it is not mandatory, in Lower Saxony, to follow one of the many books available, most schools prefer to do so, as there are still many teachers not specifically trained in the field of early foreign language education. These teachers, however dedicated and experienced in other subjects, seem to feel the need of a guideline, even if they are sometimes 104 <?page no="121"?> deviated from. There are many other reasons for choosing a course book for the primary school, for example, pressure from the parents who sometimes do not value a subject that does not require a course book, and many others too numerous and complex to discuss in this context. The first 1500-2000 words, regarded as being crucial in language teaching (Schmitt 2000) are taught over the course of the first three years of formal instruction, at a rate of about 500 items per school year (Brewster et al. 2002; Do ff & Klippel 2007). One of the schools taking part in the intervention study followed the course book Discovery (Behrendt et al. 2004a,b). The other school, while following the guidelines of the core curriculum for primary schools in Lower Saxony (Niedersächsisches Kultusministerium 2006), uses various materials, adapted whenever necessary, or teachers design their own material. 10.2.1 Vocabulary Items Chosen for Grade 3 The core curriculum states that at the end of Grade 4 learners should know words from the topics of family and friends, food and drink, shopping, school, and hobbies (Niedersächsisches Kultusministerium [Lower Saxony Ministry of Education] 2006, 15; see also Niedersächsisches Kultusministerium [Lower Saxony Ministry of Education] 1995, 38; 44). This is in keeping with attaining level A1 of the Common European Framework of Reference (Council of Europe 2001) at the end of primary school. Furthermore, the topic of likes and dislikes is one mentioned in the CEF as one topic that learners should be able to master at Level A2 (Council of Europe 2001, 58). Expressing likes and dislikes plays an important role in the life of young learners and because of this, tasks based on this were used in the study. Taking this into account, the topic food was chosen for the intervention study carried out in four Grade 3 classes. During the first three lessons, learners learned eight vocabulary items of di fferent fruit, consisting of the words apple, banana, grapefruit, lemon, orange, plum, pear, and strawberry. These specific fruit were chosen because they were in season at the time of the study and thus readily available for tasting. Second, these are probably the most common fruit encountered by learners in their daily lives. The second set of three lessons covered the names of toppings that can be put on a pizza, a dish most learners of this age are familiar with and like. This area was also chosen because not only vegetables could be 105 <?page no="122"?> introduced, but also other food often encountered in everyday life. The target vocabulary in these lessons comprised the following ten words: cheese, corn, ham, mushroom, onion, pepper, pineapple, salami, tomato, tuna. In total, the learners were taught and tested on 18 vocabulary items. In addition, they encountered other useful words related to the target vocabulary items. 10.2.2 Vocabulary Items Chosen for Grade 4 Learners in their second year of EFL instructions are more advanced than those in Grade 3 and are also already familiar with a large variety of items from the thematic sets stipulated by the Core Curriculum (Niedersächsisches Kultusministerium [Lower Saxony Ministry of Education] 2006). The vocabulary chosen, therefore, recycled some of the words the learners had already encountered. These items, for example, chocolate, orange juice, milk, eggs, and so forth, were part of the multiword items that made up the target vocabulary of this part of the study. The topic chosen for Grade 4 was that of shopping, a topic that lends itself well to developing listening and speaking skills because shopping normally involves a fair amount of communication. The lessons were divided into two sets of three lessons, the first three lessons dealing with the following items: a bar of [chocolate] 15 , a packet of [biscuits], a tin of baked beans, a tin of tuna, a carton of [eggs], a carton of [milk], a box of [cornflakes], a packet of fish fingers, a packet of [tea], a carton of [orange] juice, a jar of [honey], a jar of [jam]. The second set also used the shopping theme, but shifted the content from everyday food to items that are needed for planning a birthday party, another area of interest for primary school children. The items for this set of three lessons consisted of: a packet of [balloons], a bottle of coke, a tub of [ice cream], a bag of crisps, a bag of chips, a bag of sweets, a bag of wine gums, a roll of streamers. The items a bar of [chocolate] and a carton of [orange juice] were used in this part of the study as well. They are very common items at birthday parties and they therefore provided a link between the two parts of the lessons taught during the study. The items mentioned above reflect the fact that learners with a lower level of proficiency seem to rely a lot more on syntagmatic relations between words than more advanced learners (see Section 2.4.1). 15 Items in square brackets had been encountered by pupils prior to this study. 106 <?page no="123"?> Chapter 11 Outline of the Study The research was conducted in six primary school classes, four Grade 3 and two Grade 4. Each class was taught six lessons. During these lessons, a substitute teacher taught in order to control the teacher variable. Two di fferent substitute teachers had to be used for the two di fferent schools. Both substitute teachers were qualified primary school teachers. They were instructed to follow the outline of the lessons developed specifically for this study. (An example plan for these lessons can be found in Chapter 11.1.1). In the strictest sense, post-test results are only valid if some kind of pre-test has also been administered (Thornbury 2002). This, however, was not practicable in this study, as the learners knew very little L2 vocabulary to begin with and the range of topics they had been introduced to prior to the treatment di ffered profoundly. Therefore, pre-tests would have been almost impossible to design and interpret. 11.1 Procedure The classes were randomly divided into intervention classes and control classes. The intervention group received instructions that di ffered from those of the control group in many di fferent ways (see Tables 11.1 and 11.2). One main objective was to provide the intervention group with instructions and tasks that facilitate deep processing, as this is generally believed to aid long-term retention (see Chapter 6.2.2). To ensure this, activities, tasks and games were modified accordingly. 107 <?page no="124"?> To ensure that the results of the control group were not influenced by outside factors, for example, variety of activities or amount of input, the material used and the input given were the same. The di fference between the groups did not lie in the amount of time spent on the target vocabulary, but rather on how these target items were dealt with in class. For example, both groups in Grade 3 played Dominoes in order to consolidate their knowledge of fruit vocabulary. The control group used the more usual rule of matching like with like. The intervention group, however, was allowed to match seemingly unrelated fruit, provided that they gave an explanation for doing this. In order to ensure that both groups used the vocabulary items productively, the control group had to name the fruit on the cards as they put them down. In contrast, the intervention group was asked to provide the connection between the two fruit depicted on the cards. The reason for this is that the intervention group did not only experience more meaningful communication, but, in addition, they activated their prior knowledge of English in order to give explanations for a match, thereby activating more vocabulary items and making stronger memory connections than their control group counterparts. In the Dominoes example, it becomes apparent that, while the amount of time spent consolidating their vocabulary knowledge was the same, the task di ffered in terms of need and elaboration, to use the terms suggested by Laufer & Hulstijn (2001, see also Section 6.3.2). The rules of this game for the intervention group were explained by modelling. The teacher used a large Dominoes set to illustrate the rules. The learners received five cards each. The rest of the cards were left in a pile. One card was taken from the pile and placed, face up, in the middle of the table. On this card there was, for example, an apple and a pear. One learner started and looked at the card on the table and then at her cards. She might place an orange next to the apple, but has to justify why she has done so, for example: T - L: What’s the same? L - T: They are both round. They taste sweet. The teacher accepted any combination if the learner could say why they placed these specific cards together. The learners in the control group,however, matched a pear with a pear, and orange with an orange, and so forth. Furthermore, learners in the intervention group were encouraged to categorise the vocabulary learned, for example in terms of the material of the packaging (Grade 4). In Grade 3, this was done, for example, by 108 <?page no="125"?> providing them with a worksheet for pair work in which the learners asked questions about the fruit’s taste. The control group, however, did a pair activity in which they only expressed likes and dislikes. The rationale behind this di fference is that more elaborate statements and the increased possibilities for communication and decision-making again encourage learners to activate old and new knowledge and to acquire general conceptual knowledge about vocabulary items. This in turn leads to more connections within the mental lexicon and facilitating durable memory traces. The following lesson plan was given to the substitute teachers and they were instructed to follow this plan as closely as possible to ensure that all groups received the same treatment. 109 <?page no="126"?> 11.1.1 Example Lesson Plan - Grade 3, 2nd Lesson Vocabulary apple, banana, grapefruit, lemon, orange, plum, pear, strawberry, bitter, sour, sweet Materials flashcards depicting new vocabulary, worksheet What does the fruit taste like? Sub-aims consolidating vocabulary knowledge, developing basic listening skills, practising making statements about flavours Warming up - Reviewing Vocabulary Intervention Group Using flashcards, with pictures of the fruit taught in the previous lesson 16 , the teacher elicits the name of the pieces of fruit. Then the appropriate flashcard is put on the board. The teacher then shows the learners flashcards showing three di fferent facial expressions and elicits the words sweet, sour and bitter. Control Group As above. However, The teacher does not elicit the di fferent tastes: sweet, sour and bitter. Connecting Fruit with Flavours Intervention Group Through a series of questions and answers, the teacher elicits what the various fruit taste like, in the opinion of the learners. For example: T - L: What does an orange taste like? (Pulls the three faces.) L - T: Sweet. The learner then comes to the board and puts the ‘sweet’ face next to 16 The flashcards were mostly modified versions of depictions used in Ginger 1 (Hollbrügge & Kraaz 2003a) and Ginger 2 (Hollbrügge & Kraaz 2004a). Additional flashcards were designed by the author. 110 <?page no="127"?> the orange. The teacher then asks the rest of the class if they agree. If somebody disagrees, the teacher explains that taste is a matter of opinion. This is repeated for every piece of fruit. Control Group The teacher shows various pictures of fruit and elicits what fruit the learners like and do not like. ActionStory Intervention Group The learners are asked to stand behind their chairs. The teacher explains that they are about to hear a story of Lucy and her grandmother. While reading the story, the teacher uses suitable gestures, facial expressions and mimes to help the learners understand the story. Di fferent voices are used for Lucy and for the grandmother. The following story is told to the learners who are encouraged to join in. It is repeated eight times, with minor variations. 1) Oh, Grandma, I’m so hungry! (T holds her / his tummy and looks pitiful) I’ve got a hole in my tummy. (T draws a circle in front of his / her tummy) Lucy, look! There is everything here! (T makes a large gesture as if o ffering an variety of food) What do you want, my dear? (T draws a large question mark in the air) Grandma, can I have some fruit? (begging gesture) Sure, have some (grapefruit), it is really good. (T gives the thumbs-up) Brrr, grandma, brr, the (grapefruit) is so (bitter)! (T makes the ´bitter ’ face and shudders) 2-7) Oh, Grandma, I’m still so hungry! (T holds her / his tummy and looks pitiful) T repeats the story inserting a di fferent fruit and a disagreeable flavour each time. 111 <?page no="128"?> 8) T tells the story one last time, this time using strawberry and sweet using a di fferent last line: Mmm, grandma, mmm, the strawberry is so sweet! Thank you, Grandma! (T mimes embracing someone) Control Group The teacher does neither use the words bitter and sour nor pull the appropriate faces. Instead s / he says that all the fruit is yukky. In the last line of the story the word yummy is substituted for the word sweet. Consolidating New Knowledge Intervention Group T distributes the worksheet, What does the fruit taste like? (see Figure 11.1). The teacher elicits what the three faces mean. A learner is asked, “What does an orange taste like? " The learner answers, for example, “It’s bitter." The teacher then draws an orange next to the bitter face. This is modelled until the learners understand what they have to do. They are then put into pairs, ask their partner what they think the various fruit tastes like and fill in the worksheets. At the report back stage, the teacher asks the learners to say what their partners think the various pieces of fruit taste like. Control Group T distributes the worksheet, What does the fruit taste like? (see Figure 11.1). S / he explains to the learners that the smiley face means ‘I like it’ and the face with the scowl means ‘I don’t like it’. S / he models what the learners should do, that is, draw the appropriate fruit next to the appropriate face. The learners are then put in pairs to find out from their partner what fruit they do and do not like and draw the fruit next to the faces. At the report back stage the learners are asked what fruit their partners do and do not like. 112 <?page no="129"?> Figure 11.1: Worksheets for Intervention Group (left) and Control Group (right) 11.1.2 Example Lesson Plan - Grade 3, 5th Lesson Vocabulary cheese, corn, ham, mushroom, onion, pepper, pineapple, salami, tomato, tuna Materials flashcards depicting pizza ingredients (whole), flashcards pizza toppings, worksheet Pizza, pizza memory Sub-aims consolidating new vocabulary, developing speaking skills, learning to make statements about food and ask questions about it Warming up Activity - Reviewing last lesson’s vocabulary Intervention Group The teacher elicits the names of the various pizza toppings covered in the previous lesson. When the learners name a topping, the teacher puts the appropriate picture on the board. 113 <?page no="130"?> Control Group As above Vocabulary Expansion Intervention Group The teacher distributes the pizza worksheet (see Figure 11.2). The teacher models, drills and elicits until the learners understand what they have to do with the sheet. On the worksheet are pictures of a plant and animals. There is a ‘sweet’ face and a salt cellar. The children tick whether a topping is sweet or salty and from a plant or from an animal. When the learners have completed the sheet, they compare answers with their partner. The teacher then discusses some of the examples and stresses that everyone’s sense of taste is di fferent. Figure 11.2: Worksheets for Intervention Group (left) and Control Group (right) 114 <?page no="131"?> Control Group The teacher distributes the pizza worksheet (see Figure 11.2). The teacher models, drills and elicits until the learners understand what they have to do with the sheet. One the worksheet are pictures of various toppings that can be put on a pizza. There is a smiling face and a scowling face. The learners ask their partners what they do and do not like on their pizza and draw the appropriate topping under the appropriate face. After the learners have finished, the teacher discusses some of the answers with the class. Game - Pizza Memory Intervention Group and Control Group The teacher elicits the rules of memory by modelling the game. Once the learners understand what they have to do they start to play the game. The learners are put into groups (of five, if possible) and given two sets of cards. One set is square, the other, rectangular. The rectangular cards are larger. The square cards are placed, face down in rows and columns, on the desk. The larger cards are left in a pile. One learner takes a card from the pile and looks at the picture. The learner then tells the other players what is on the card, for example: L: On my pizza, I like pepper, salami and onion. The learner then takes one of the square cards. On this card there is a pizza. The learner tells the other players what is on the pizza; for example: L: On this pizza there is salami, onion and pineapple. The two pictures do not match so the learner says, “This is not my pizza.", puts the large card on the bottom of the pile and the square card back in its original place. When a child has two cards that match, they keep both cards. The learner who collects the most cards wins the game. Pizza Rhyme Intervention Group and Control Group The learners are asked to stand up. The teacher then chants the following rhyme, accompanying it with appropriate actions, mimes and gestures. Pizza is yummy, pizza is fine, I like pizza, I eat it all the time. 115 <?page no="132"?> I eat it for breakfast, I eat it for lunch, I love eating pizza, munch, munch, munch. On the pizza may be tomato, or cheese, or even green pepper. They all taste so good to me, it doesn’t even matter. Onion, ham, or mushroom, anything will do. I like pizza, how about you? The rhyme is repeated several times and the learners join in, repeating the actions and, if possible, the words of the chant. 11.1.3 Summary of the Di fferences between the Groups The lesson plans for the other lessons followed a similar line to the ones described above. Tables 11.1 and 11.2 give an outline of the di fferences between groups, a more detailed description of the lessons can be found in Appendix A. Task / Activity Intervention Group Control Group Fruit Tasting Fruit are connected with the flavours bitter, sweet, sour Fruit are only talked about in terms of likes and dislikes Revising Vocabulary as above as above Action Story as above The fruit are only referred as yukky or yummy Worksheet Fruit Learners ask their partner which fruit they think taste bitter, sour or sweet Learners ask their partner about likes and dislikes Dominoes Learners have to give reasons for which ways certain fruit are similar Learners match the types of fruit and repeat the English words for them, saying ‘This is a ...’ Song Learners sing the song ‘Here We Go Round the Lemon Tree’ and discuss which fruit grow on trees Learners perform the rap ‘One Banana, Two Bananas’ Worksheet Pizza Learners decide whether a topping is from a plant or an animal and whether it is sweet or salty Learners decide whether they like or dislike the given pizza toppings Table 11.1: Overview of Di fferences between Groups in Grade 3 116 <?page no="133"?> Task / Activity Intervention Group Control Group Recycling Vocabulary Learners categorise shopping items according to their containers Learners repeat vocabulary drill from the last lesson Introducing New Vocabulary Learners are provided with category clues No clues are provided Recycling Vocabulary Learners discuss the di fferent meanings of chips and crisps The issue of di fference of use is not discussed Shopping Dominoes Learners have to give reasons why certain items are similar Learners match the items and repeat the English words from them, saying ‘This is a ...’ Table 11.2: Overview of Di fferences between the Groups in Grade 4 11.2 Participants The research was carried out in four 3rd grade classes (age: 8-10) and two 4th grade classrooms (age: 10-12) 17 . The study was conducted in April, 2007, in School M and in June, 2007, in School P. These dates were chosen to take various factors into consideration, among them the internal organisational constraints of the schools and the fact that pupils had not yet covered the relevant vocabulary. School Grade Treatment Code M 3 Intervention M3I M 3 Control M3C P 3 Intervention P3I P 3 Control P3C P 4 Intervention P4I P 4 Control P4C Table 11.3: How the Groups were Coded All interventions were preceded by an observation of the groups to become acquainted with the learners and to become familiar with the teach- 17 In order to comply with education legislation, an outline of this study was submitted to and approved of by the local education authority (Landesschulbehörde Hannover). 117 <?page no="134"?> Code N Grade 3 Intervention Groups M3I 16 P3I 20 Control Groups M3C 19 P3C 18 Grade 4 Intervention Group P4I 15 Control Group P4C 13 Table 11.4: N for Each Class - Test 1 ing methodology during normal classes to ensure that they were suited for the research. In addition, the intervention was preceded by one lesson in each class. This had the sole purpose of acquainting the substitute teacher with the names of the pupils, and vice versa, during the intervention. These six classes were in two di fferent schools 18 in Hildesheim, Lower Saxony, Germany. In each school one class was randomly selected to form the intervention group and the other class served as a control group. The classes in School M had been taught by the same teacher prior to the intervention. In School P two di fferent teachers were responsible for teaching English in Grade 3, but the topics covered by the classes were comparable. In School P’s 4th grade, one teacher had taught both groups prior to the intervention. Code N Grade 3 Intervention Groups M3I 16 P3I 19 Control Groups M3C 19 P3C 18 Grade 4 Intervention Group P4I 14 Control Group P4C 14 Table 11.5: N for Each Class - Test 2 The tests and questionnaires were coded according to the following coding system: Firstly, the groups were divided according to school, the first letter of the code therefore being M or P. The second part of the code refers to the grade the subjects were in. In third position the kind of treatment 18 Additional information that might identify these schools is withheld to preserve the anonymity of the pupils. 118 <?page no="135"?> was coded, the letter I denoting the intervention (or treatment) group and C, the control group. The individual learners were assigned consecutive numbers. These were entered after the code to preserve anonymity while at the same time ensuring that individuals could be compared. Table 11.3 shows the groups with the labels they were assigned to ensure ease of identification in the statistical analyses. N for Test 1 N for Test 2 Grade 3 Intervention Group 36 35 Control Group 37 37 Grade 4 Intervention Group 15 14 Control Group 13 14 Table 11.6: Pooled N for Each Group Tables 11.4 and 11.5 show the number of participants in each class that were present during the whole period of the intervention. All learners who missed more than two classes during the study were excluded. Some learners were not present for one of the tests, leading to slightly di fferent numbers for Test 1 and Test 2. Furthermore, if parents had refused permission for their children’s data to be included in this study, the children took part in classes, but no data were collected. Table 11.6 states the sample size (N) for both tests after the groups were pooled. The N for grade 4 is very low, because the study in grade 4 was only done in two classes in School P and these were particularly small. Because of the low number of subjects, the statistical tests available for this group were severely limited. 11.3 Test Instruments Used The tests (see Figures 11.3 and 11.4) used were developed specifically for this study. They are based on material available for measuring vocabulary knowledge that form either part of a course book such as Making the Grade with Ginger (Bebermeier et al. 2004) and Show what you know (Becker et al. 2003) or published separately, for example, See what I can do in English (Ehlers et al. 2006). Furthermore, when developing the tests, many of the test items were based on similar ones outlined in the literature on teaching English to young learners (e.g. Cameron 2001; Brewster et al. 2002; Iannou-Georgiou & Pavlou 2003; McKay 2006). 119 <?page no="136"?> The first test was administered two days after the intervention lessons and control group lessons had been completed. This was two days after the last lesson had been taught for all classes. The second test was administered two weeks later. This second test was intended to measure long-term retention of vocabulary knowledge. The test content and overall layout was the same in both tests, but the internal order of the items was changed in the second test to prevent the interference of a memory variable. The tests were graded according to the correctness of the first try. Corrections were excluded, as the main interest lay in the ease and speed of recall. This could be easily decided because any corrections could easily be identified. Figure 11.3: Test used in Grade 3 The test first given to both groups in School M contained an additional item that was excluded from analyses as it was found to be impossible to grade. This item asked learners to listen to a description of what a given person likes on his or her pizza. However, this activity, sometimes referred to as ‘Listen and draw lines’ (Bebermeier et al. 2004), led to learners drawing lines that were so entangled that it was impossible to reliably decide where they were supposed to lead. The possibility of a high error rate in grading and false interpretation could not be excluded, therefore this item was not included in the statistical analyses. The tests (see Figures 11.3 and 11.4) were administered almost immediately after the intervention and then administered again after two weeks. The possibility of measuring delayed long-term achievement was not 120 <?page no="137"?> Figure 11.4: Test used in Grade 4 explored, as it would have been impossible to ensure that all groups had had similar language instruction after the intervention ended. As variables such as the repeated exposure to test items over time could not be controlled, a long-term achievement test after, for example, three months, would not have yielded comparable results. The tests were printed in black and white on A4 paper. The test instructions and listening comprehension tasks were read aloud by the teacher, who was instructed as to how to administer the tests. Each listening comprehension text was read twice only. The learners were informed of the procedure in both English and German in order to ensure that every learner was fully aware of what was expected of them. The teacher checked this by questioning them about the tasks. The maximum score for the tests in Grade 3 was 38, whereas the tests used in Grade 4 had a maximum score of 37. 11.4 Questionnaires The self-assessment questionnaires (Figures 11.5, 11.6 and 11.7) were handed out after the tests. The first version, completed directly after the intervention, included questions asking learners to rate the games played during the intervention. As it would have been redundant to ask this again two weeks after the intervention, these questions were omitted in the second questionnaire. 121 <?page no="138"?> Figure 11.5: Questionnaires 1 (bottom) and 2 (top) for Grade 3 122 <?page no="139"?> Figure 11.6: Questionnaire 1 for Grade 4 123 <?page no="140"?> Figure 11.7: Questionnaire 2 for Grade 4 124 <?page no="141"?> 11.5 Exploring the Data for Grade 3 All statistics were calculated using the Statistics Package for the Social Science (SPSS) 15.0. All statistical values are rounded. 11.5.1 Test 1 The Kolmogorov-Smirnov tests (K-S tests) administered to see whether the distribution of the data is close to normality (Field 2005) indicate for the intervention group, D(36) = 0.25, p < .001, and for the control group, D(37) = 0.22, p < .001, that both distributions are not normal. The means and standard deviation for both groups in Test 1 are summarized in Table 11.7. Intervention Group N 36 mean 36.08 standard deviation 2.822 Control Group N 37 mean 35.95 standard deviation 2.223 Table 11.7: Mean Results for Test 1 in Grade 3 Although the K-S test suggests that the distributions are not normal, it may be assumed that the underlying population is distributed normally (Rost 2007). Therefore, revised tests and a larger sample size are very likely to lead to normally distributed results. Furthermore, it is possible that an existing e ffect is not statistically significant due to the small sample size (Rost 2007), therefore the e ffect size is also reported. Based on this assumption, a t-test was carried out to compare the means of the two groups. On average, subjects in the intervention group had slightly higher scores (M = 36.08, SE = 0.47) than subjects in the control group (M = 35.93, SE = 0.36). This di fference is not significant (t(71) = 0.23, p = .82) and had no e ffect 19 with d = 0.05. 19 E ffect size interpretations are based on Bortz & Döring (2006, 606). 125 <?page no="142"?> Figure 11.8: Score Frequencies for Test 1 in Grade 3 Figure 11.9: Percentages for Test 1 in Grade 3 126 <?page no="143"?> 11.5.2 Test 2 The second test was administered two weeks after the first test in order to measure long-term retention. Intervention Group N 35 mean 35.51 standard deviation 2.853 Control Group N 37 mean 34.38 standard deviation 3.810 Table 11.8: Mean Results for Test 2 in Grade 3 The K-S test scores in Test 2 for the intervention group, D(35) = 0.23, p < .001, and for the control group, D(37) = 0.22, p < .001, both significantly indicate that the distributions are not normal. The means and standard deviations for both groups in Test 2 are summarized in Table 11.8. Here, the t-test showed that again participants in the intervention group achieved slightly higher scores (M = 35.51, SE = 0.48) than subjects in the control group (M = 34.38, SE = 0.63). This di fference is not significant (t(70) = 1.42, p = .16), however, it does represent a moderate e ffect with d = 0.34. 127 <?page no="144"?> Figure 11.10: Score Frequencies for Test 2 in Grade 3 Figure 11.11: Percentages for Test 2 in Grade 3 128 <?page no="145"?> 11.5.3 Comparing Test 1 and Test 2 in Grade 3 Table 11.9 shows an overview of the means and standard deviations for both tests. Test 1 Test 2 Intervention Group N 36 35 mean 36.08 35.51 standard deviation 2.822 2.853 Control Group N 37 37 mean 35.95 34.38 standard deviation 2.223 3.810 Table 11.9: Comparison of Mean Results in Grade 3 In both Grade 3 tests, it becomes apparent that there is a strong ceiling e ffect. By looking at the percentages of test scores, we can see that 50% of all pupils in the intervention group achieved the highest possible score in Test 1 (Figure 11.9), whereas only 24% of the control group did so. In Test 2, the picture is similar. Although overall fewer learners achieved a perfect score (26% of the intervention group, 19% of the control group, see Figure 11.11), the di fference between the two groups is apparent. 11.5.4 Chi-Square Tests Chi-square tests are employed to analyse the relationship between categorical variables, that is, variables that describe categories of entities (Field 2005). The chi-square test can be used on the present data, as they fulfill the assumptions of the test, namely each item fills only one cell in the contingency table (see below) and the expected frequency in each cell is greater than 5 in all cases reported here (ibid., 686). For each chi-square test, the value of χ 2 , the degrees of freedom, and the significance value p will be reported. Where applicable, the value of Cramer ’s V will also be reported, as this is the preferred measure for the strength of association when variables have more than two categories (Field 2005, 693). 129 <?page no="146"?> Self-Evaluation of Vocabulary Knowledge There was no significant association between the type of lessons attended and self-evaluation of the vocabulary knowledge in Test 1, χ 2 (2) = 3.835, p = .14. Group*Self-Evaluation Crosstabulation Vocabulary Knowledge I know I’m not I don’t this sure know this word word Total Group Intervention Count 568 41 21 630 Group Expected Count 556.6 48.9 24.5 630.0 % of Total 43.2% 3.1% 1.6% 47.9% Control Count 593 61 30 684 Group Expected Count 604.4 53.1 26.5 684.0 % of Total 45.1% 4.6% 2.3% 52.1% Total Count 1161 102 51 1314 Expected Count 1161.0 102.0 51.0 1314.0 % of Total 88.4% 7.8% 3.9% 100.0% Table 11.10: Vocabulary Knowledge Self-Evaluation for Test 1 In Test 2, however, there is a significant association between the type of lessons and the self-evaluation of the vocabulary knowledge, χ 2 (2) = 7.72, p = .02, with Cramer ’s V giving an e ffect size of 0.08. Group*Self-Evaluation Crosstabulation Vocabulary Knowledge I know I’m not I don’t this sure know this word word Total Group Intervention Count 559 48 21 628 Group Expected Count 548.1 47.7 32.1 628.0 % of Total 42.1% 3.6% 1.6% 47.3% Control Count 601 53 47 701 Group Expected Count 611.9 53.3 35.9 701.0 % of Total 45.2% 4.0% 3.5% 52.7% Total Count 1160 101 68 1329. Expected Count 1160.0 101.0 68.0 1329.0 % of Total 87.3% 7.6% 5.1% 100.0% Table 11.11: Vocabulary Knowledge Self-Evaluation for Test 2 By looking at the contingency tables, it can be seen that more subjects of the intervention group than expected stated that they know the given vocabulary items, whereas the control group shows the opposite pattern. 130 <?page no="147"?> Fewer participants of the intervention group than expected stated that they are either unsure about an item or do not know it. Again, the control group shows the opposite pattern, with more subjects than expected stating that they did not know an item. Apart from looking at the e ffect size provided by Cramer ’s V, which unfortunately does not indicate even a moderate e ffect for Test 2, it would have been interesting to calculate the odds ratio for these contingency tables. However, this is best done in a 2x2 table (Field 2005, 593). The tables were converted to fit this, grouping the two categories ‘I’m not sure’ and ‘I don’t know this word’ together, but unfortunately the chisquare tests were not statistically significant, giving a value of χ 2 (1) = 3.82 p = .06 for Test 1 and χ 2 (1) = 3.21 p = .08 for Test 2. However, as the lack of statistical significance may be due to the small sample sizes, the odds ratios were calculated and yielded 1.40 for Test 1 and 1.35 for Test 2 . This points to the fact that learners were 1.4 times (Test 1) or 1.35 times (Test 2) more likely to state that they remember a given vocabulary item if they attended the intervention classes (Field 2005, 693). Rating the Games There was a significant association between the type of lessons attended and rating of the games played χ 2 (2) = 7.12, p = 0.03, with Cramer ’s V being 0.12, representing a weak, but nevertheless detectable, e ffect. Group*Games Rating Crosstabulation Games Rating good okay not so good Total Group Intervention Group Count 159 64 8 231 Expected Count 146.5 71.6 12.9 231.0 % of Total 32.9% 13.2% 1.7% 47.7% Control Group Count 148 86 19 253 Expected Count 160.5 78.4 14.1 253.0 % of Total 30.6% 17.8% 3.9% 52.3% Total Count 307 150 27 484 Expected Count 307.0 150.0 27.0 484.0 % of Total 63.4% 31.0% 5.6% 100.0% Table 11.12: Games Rating Crosstabulation More intervention group subjects than expected gave the games a good rating, while fewer than expected stated that the games were merely okay or not so good. 131 <?page no="148"?> Again, the control group shows the opposite pattern. Fewer participants than expected gave the games a good rating, while more than expected rated the games as okay or not so good. 11.6 Exploring the Data for Grade 4 11.6.1 Test 1 The K-S test showed that the distributions for both the intervention and the control group are not normal, the percentages being D(15) = 0.22, p < .05, and D(13) = 0.26, p < .05, respectively. Intervention Group N 15 mean 33.53 standard deviation 2.722 Control Group N 13 mean 32.69 standard deviation 3.966 Table 11.13: Mean Results for Test 1 in Grade 4 A t-test showed that subjects of the intervention group (M = 33.53, SE = 0.70) score slightly higher than those in the control group (M = 3.69, SE = 1.10) 20 . This di fference, however, is again not significant t(20.82) = 0.64, p = .53). Nevertheless, there is a small e ffect (d = 0.25). 20 Levene’s test is significant (p = .04), therefore we can assume that the null hypothesis is incorrect and that the variances are significantly di fferent (Field 2005, 301). 132 <?page no="149"?> Figure 11.12: Score Frequencies for Test 1 in Grade 4 Figure 11.13: Percentages for Test 1 in Grade 4 133 <?page no="150"?> 11.6.2 Test 2 The K-S test administered to see whether the distribution of the data is close to normality (Field 2005) showed the scores in Test 2 for the intervention group, D(14) = 0.15, p < .05, and for the control group, D(14) = 0.14, p < .05. Both significantly indicate that the distributions are not normal. This di fference, however, is significant t(26) = 2.57, p = .02) and represents a strong e ffect of d = 0.97. Intervention Group N 14 mean 33.50 standard deviation 1.990 Control Group N 14 mean 30.79 standard deviation 3.423 Table 11.14: Mean Results for Test 2 in Grade 4 134 <?page no="151"?> Figure 11.14: Score Frequencies for Test 2 in Grade 4 Figure 11.15: Percentages for Test 2 in Grade 4 135 <?page no="152"?> 11.6.3 Comparing Test 1 and Test 2 Test 1 Test 2 Intervention Group N 15 14 mean 33.53 33.50 standard deviation 2.722 1.990 Control Group N 13 14 mean 32.69 30.79 standard deviation 3.966 3.423 Table 11.15: Comparison of Mean Results in Grade 4 A ceiling e ffect also influences the results of the vocabulary test scores in Grade 4. As Table 11.15 illustrates, the means di ffer only slightly, but the spread in the control group is more pronounced. In Test 1, the great majority (80%) of learners in the intervention group, but only 62% percent of learners in the control group, have a score of 33 or more. In Test 2, 70% of learners in the intervention group scored 33 points or more, whereas only 35% of the control group achieved 33 points or more. 11.6.4 Chi-Square Tests It is not feasible to do a chi-square test on the self-evaluation of vocabulary knowledge or the ratings of the games, as the N for each group is too small and therefore the expected frequencies for some cells are below 5. 136 <?page no="153"?> 11.7 Discussion of Results Although some of the statistical analyses led to inconclusive results, this exploratory study has nevertheless yielded interesting insights into the research question as to whether teaching vocabulary according to the principles outlined in Part 1 and 2 leads to better long-term retention of the vocabulary taught. The comparison of all test score means was e ffected by the ceiling e ffect, which was strongest in Grade 3. This e ffect is due to the fact that the vocabulary tests used were too easy and therefore too many learners achieved top scores, leading to non-normal distributions. The tests used for young learners in this study contained familiar elements and repetitive routines, as is recommended in the literature (see e.g. McKay 2006), but the adherence to this view might have led to the ceiling e ffects described above. However, the mean scores of the intervention group were in all cases higher than those of the control group in all tests. This di fference, however, was too small to yield significant results. Therefore, the possibility that the di fferences between the two groups can be attributed to chance cannot be excluded. The chi-square tests of the self-evaluated vocabulary knowledge in the 3rd grade all show a similar pattern, namely that more learners than expected stated that they know the vocabulary items taught in the study, while in the control group more learners than expected were either unsure about items or stated that they did not know them at all. Although only Test 2 yielded a significant result, this pattern emerges in both Test 1 and Test 2. This indicates that vocabulary taught with meaningful connections, facilitating deep processing and encouraging the learners to make connection in their mental lexicon, may indeed lead to better long-term retention, at least in the eyes of the learners themselves. Furthermore, although there was no significant association between self-perceived vocabulary knowledge and group membership in Test 1, the second self-evaluation showed that, after the knowledge had time to be consolidated in the minds of the learners, an association between group membership and self-perceived vocabulary knowledge emerged. The e ffect size is only small, but further research may be able to show that encouraging deep processing indeed leads to long-term retention of language, especially over a longer period of time. To verify this, a follow-up study with more tests spread over a longer period of time is needed. The chi-square test of the games rating for the 3rd grade provided 137 <?page no="154"?> additional information that demonstrate that instruction adhering to the principles outlined seems to be more interesting, and thus more motivating, for learners because the games played in the intervention group were generally rated higher that those in the control group. One of the reasons for this e ffect might be that the games played by the intervention group encouraged genuine communication and exchange of information. They were therefore perceived as being more meaningful and valuable by the learners. The influence of this teaching methodology on learners’ motivation needs to be investigated further. The results for Grade 4 pose more problems when it comes to their interpretation. Due to the small sample sizes, most statistical tests could not be carried out, therefore all results have to be interpreted with caution, as the di fferences in mean scores were not significant and could be chance findings. Some of these problems could be due to the lesser extent of di fference between the two groups, therefore the instructions for a follow-up study need to be revised. For example, this could be done by laying an even stronger emphasis on collocational information provided for the intervention group than for the control group. However, the overall picture allows the conclusion that a larger sample size, a longer exposure to the two di fferent ways of teaching vocabulary and an improved, that is, more complex, vocabulary test would lead to significant di fferences in long-term retention of vocabulary. 11.8 Discussion of Classroom Observation The teaching during the time of the study was done by teachers specifically instructed for this research, therefore the classroom interaction was observed. For this purpose, the following guiding questions were considered: • What role does the L1 play during vocabulary instruction? • Do beginners show a tendency to associate words according to clang? 138 <?page no="155"?> • Do learners in the intervention group take part in more meaningful communication? • Do learners communicate enough while playing the games? • Do learners make unexpected connections between words? Although the observations were not quantified and often do not go beyond impressions, they are nevertheless interesting in view of the theoretical background discussed in Part 1 and 2. Role of the L1 The L1, predominantly German, although some children speak additional languages at home, seemed to be an important resource for young learners. Apart from persistently asking ‘What’s that in German? ’, learners often based their guesses of unknown words on the German language, suggesting, for example, pfloom for plum or citrus and citron for lemon. Additionally, they also used their L1 to fill gaps in formulaic sequences, by stating, for example ‘This is a Birne (pear).’ Word Association Patterns It was extremely noticeable in each class, regardless of whether it was the intervention or control group, that some children seemed to rely heavily on clang associations. It was common for these learners to not only repeat the target vocabulary, but also to supply a rhyming association, sometimes in the L1, sometimes the L2 and even using nonsense words. For example, one learner repeatedly said pear, hair, another remarked ‘tuna, wie tuning beim Auto’ (‘tuna, like tuning a car ’). Surprisingly, another learner at the other school came up with the exact same association. Others repeatedly said ‘Matschraum’ instead of mushroom. In grade 4, many learners seemed to like the alliteration patterns of ‘a jar of jam’ and ‘a tin of tuna’, stating that this was easy to remember. Another learner explained that he found corn easy to remember as it consists of Körner (seeds), which sounds similar. Creative Connections Another point that emerged, during lesson observations, was that learners in the intervention group were very quick to form their own categories 139 <?page no="156"?> and associations in addition to those provided by the input. While carrying out the activity that categorised vocabulary items in terms of their containers, chocolate was hotly debated over, as it is wrapped in paper and tin foil or sometimes even plastic. The learners finally settled the dispute by agreeing that chocolate is ‘the odd one out’ and therefore a special case. In Grade 3, many children referred to grapefruit as the ‘big fruit’ or ‘big orange’ as the illustrations used were indeed very similar, although the grapefruit was larger and lighter in colour. While filling in the worksheet ‘What does it taste like? ’ (see Figure 11.1), one learner put in an additional column for those fruit that he thought were disgusting. In Grade 4, learners came up with their own categories for the party items, stating that there are things that you can bite into, for example crisps, and other things you can slurp, for example coke and juice. Interaction during Games Unfortunately, many parents refused to give permission to record their children while they played the games. Therefore, no recordings were done, as it would have been very di ffi cult to ensure that only data from children whose parents gave permission to use their data were analysed. While observing the di fferent groups, however, it became apparent that the interaction in the intervention groups was more animated and much more varied. This was partly due to the modified rules that di fferentiated between the control and intervention group and partly because the children felt the need to genuinely communicate information. For example, while playing Dominoes in the intervention group, an explanation as to why two items matched had to be given and the competitors had to allow the combination. Therefore, authentic dialogues ensued, with learners arguing their respective points, using an astonishing amount of English. Learners in the control group, however, repeated the required chunks of language, for example ‘This is an orange and this is also an orange’, but there was no need to justify the matchings. Therefore, almost no variation of output occurred and learners seemed to revert to German more quickly when problems arose. 140 <?page no="157"?> Chapter 12 Conclusion After decades of vocabulary being neglected in the fields of foreign and second language acquisition, research into vocabulary acquisition, learning and teaching now covers a wide range of areas related to these disciplines. One main reason for this is that the relevance of vocabulary acquisition and teaching in FLA and SLA is now undisputed (see e.g. Lewis 1993; Read 2004b; Nation 2001). Research on the acquisition of vocabulary in a second language primarily focuses on di fferent learning conditions, for example, on incidental vocabulary acquisition through reading (see e.g. Laufer & Hulstijn 2001). As outlined in Chapters 7 and 8, beginners, and young learners in particular, require a di fferent research focus, because they need to be taught explicitly in order to acquire enough word families to read at the level required for incidental vocabulary acquisition through extensive reading (Laufer 1997). Additionally, the main focus of many other empirical studies on vocabulary acquisition in a foreign language still lies on the analysis of adult learners of intermediate or high proficiency. This is because research in this area often equates second language learning with adult language learning (Nicholas & Lightbown 2008, 28). A very similar picture emerges when looking at the research on vocabulary acquisition within a Cognitive Linguistics framework. Studies are predominantly conducted using adolescent or adult learners (see Verspoor & Lowie 2003 or the contributions by Beréndi et al. 2008, Skoufaki 2008, Condon 2008 and MacArthur & Littlemore 2008 in Boers & Lindstromberg 2008a). There seems to be no reason, however, why only older or adult learn- 141 <?page no="158"?> ers should be given vocabulary instruction based on Cognitive Linguistic principles, because “the gradual development of the L2 lexicon is a fundamental part of the process of learning a foreign language" (Piquer Píriz 2008, 219). Therefore young learners may benefit in the same way as adult learners (Boers & Lindstromberg 2008c). Studies (see e.g. Piquer Píriz 2008) have shown that young EFL learners benefit from enriching their understanding of vocabulary items “by extending their meaning in motivated ways" (ibid., 236). Concrete methodology, based on these implications, needs to be introduced into foreign language classrooms, because it might then have a positive impact on learning processes and knowledge expansion. Unfortunately, the practical implementation of theoretical frameworks and hypotheses in formal language instruction is still rare, although many course books (e.g. Discovery (Behrendt et al. 2004b) and Sally (Bredenbröcker et al. 2005a)) claim that they base their methodology and curriculum on findings of applied linguistics and Cognitive Linguistics. The actual content of the course books, however, often falls short of making full use of research findings in this area. This is deplorable for many reasons, one of them being that teaching vocabulary in German school is still very much based on pair associations, a practice that does not promote successful language acquisition. Instead of introducing and recycling vocabulary items in a rich linguistic environment and meaningful communication, that is, in usage events, vocabulary teaching is still given very little priority in many German schools. This study aimed at investigating how lack of emphasis on usage events could be remedied. It focuses on young learners who are beginners with a very limited proficiency in the target language. It has been argued that, given the right mode of instruction, vocabulary teaching does not only promote the acquisition of lexical items, but, ideally, also aids the development of the learner ’s mental lexicon. The intervention study showed that, although it is possible to devise activities, tasks and instructions that implement the theoretical implications of research carried out in the area of the mental lexicon and SLA, what has not been done is to see if the implementations are really e ffective. Although the results of the present study o ffer only limited proof that the intervention was successful and does result in long-term retention, encouraging learners to elaborate on words and to use them in meaningful communication does indeed seem to help them acquire a language with more ease. This leads to a more profound understanding of language, 142 <?page no="159"?> which in turn aids the committing of vocabulary to long-term memory. Learners appeared to gain more confidence when they were shown connections between lexical items, an outcome that is predicted by a Cognitive Linguistics approach to teaching vocabulary (Boers & Lindstromberg 2008c). This is because the deeper level of processing in combination with meaningful input that goes beyond the purely linguistic level helps learners to develop connections within their mental lexicon. It also has an impact on the activation threshold of vocabulary items, therefore making these items more readily available to the learner. The intervention outlined in Chapter 11 attempted to provide young learners with the “pedagogical potential of creating meaningful, and therefore potentionally memorable, assocations" (Boers & Lindstromberg 2008c, 20) between vocabulary items. This should be seen as an ongoing process and ought to be fostered throughout the learners’ learning process, because repeated encounters with items also help develop a deeper understanding of all the facets of meaning of them, including figurative meaning extensions. The study, on the other hand, also revealed that there is still a great deal of research to be done to fully understand and evaluate the benefits of a Cognitive Linguistic approach to second language vocabulary learning. Apart from the need to develop better vocabulary tests that accurately measure vocabulary knowledge and vocabulary gain in both breadth and depth, it is also necessary to conduct further studies in which the di fferences between intervention and control groups are more pronounced. The sample sizes need to be large enough to make statistically sound statements about the e ffectiveness of this approach. With regard to the vocabulary knowledge, a good starting point would be to look into measuring vocabulary depth in EFLT, using, for example, an adaptation of the WAT test (Schoonen & Verhallen 2008). In addition, the methodological side of language testing in primary school needs to be investigated further, ideally leading to a standardized test instrument that can be used to compare learners’ achievements and to implement a proper pre-test / post-test design. This should enable an accurate measure of vocabulary gain over a given period of time. Although it is often advocated that the tests used for young learners should contain familiar elements and repetitive routines (see e.g. McKay 2006), an adherence to this view might have led to the ceiling e ffects described in Chapter 11.7. By including too many familiar routines or tasks already practiced during lessons, the tests may have been too easy to complete, leading to skewered results. More complex tasks and a larger 143 <?page no="160"?> variation of test items might be a way to address this problem. Another possibility is to include other test types, for example production tasks, and not only the selected-response tasks. These types of tasks measure productive vocabulary knowledge and would allow for the evaluating, in more depth, to what extent learners benefit from usage-events in language learning. The operationalisation of test variables, therefore, needs to be addressed because a sharper focus on only one aspect might lead to more conclusive results. On the other hand, as DST suggests, language learning is a highly complex process and isolating certain variables in a test design might lead to severely limited results that have only little relevance for language teaching methodology. Language learning is a process that has long been treated as if it were a linear, additive process, but in reality it is an iterative one (De Bot et al. 2007; Larsen-Freeman & Cameron 2008). It is fundamentally based on experiences, the “constant adaptation and enactment of language-using patterns" (Larsen-Freeman & Cameron 2008, 158) in communication. Further studies, therefore, should aim at addressing this by implementing vocabulary knowledge tests that are based on communicative events rather than standardized tests. 144 <?page no="161"?> Bibliography Achard, M. & Niemeier, S., (Eds.) (2004). Cognitive Linguistics, Second Language Acquisition, and Foreign Language Teaching. Berlin: Mouton de Gruyter. Ahlsén, E. (2006). Introduction to Neurolinguistics. Amsterdam: John Benjamins. Aitchison, J. (1992). Introducing Language and Mind. Harmondsworth: Penguin. Aitchison, J. (2003). Words in the Mind. An Introduction to the Mental Lexicon. Third Edition. Oxford: Blackwell. Aßbeck, J. (1996). Schaut Euch dann bis zur nächsten Stunde bitte die Wörter an! Der fremdsprachliche Unterricht Englisch, 2, 25-30. Aßbeck, J. (2002). Thesen zur Wortschatzüberprüfung - ‘An der Tafel habe ich immer Mattscheibe’. Der fremdsprachliche Unterricht Englisch, 36, 28-32. Baddeley, A. D. & Hitch, G. J. (1974). Working memory. In G. H. Bower (Ed.), The Psychology of Learning and Motivation (pp. 47-89). London: Academic Press. Barlow, M. & Kemmer, S., (Eds.) (2000). Usage-Based Models of Language. Stanford, CA: CSLI Publications. Bebermeier, H., Fredrichs, M., Hartmann-Kleinschmidt, E., & Stoll, U. (2004). Making the Grade with Ginger 1. Berlin: Cornelsen. Becker, C., Junakoavic, S., Gerngross, G., & Puchta, H. (2003). Show what You Know. Stuttgart: Klett. Behrendt, M., Christmann, B., & Mayer, N. (2004a). Discovery 4 - Teacher’s Book. Braunschweig: Westermann. Behrendt, M., Christmann, B., Mayer, N., & Stelter, A. (2004b). Discovery 3 - Teacher’s Guide. Braunschweig: Westermann. Beréndi, M., Csábi, S., & Kövecses, Z. (2008). Using conceptual metaphors and metonymies in vocabulary teaching. In F. Boers & S. Lindstromberg 145 <?page no="162"?> (Eds.), Cognitive Linguistic Approaches to Teaching Vocabulary and Phraseology (pp. 65-99). Berlin: Mouton de Gruyter. Bickes, H. (2004). Bilingualismus, Mehrsprachigkeit und mentales Lexikon evolutionsbiologische, soziokulturelle und kognitionswissenschaftliche Perspektiven. Fremdsprachen Lehren und Lernen, 33, 27- 51. Bock, K. & Levelt, W. J. M. (1994). Language production, grammatical encoding. In M. A. Gernsbacher (Ed.), Handbook of Psycholinguistics. San Diego, CA: Academic Press. Boers, F. (2000). Metaphor awareness and vocabulary retention. Applied Linguistics, 21, 553-571. Boers, F., Eyckmans, J., & Stengers, H. (2007). Presenting figurative idioms with a touch of etymology: more than mere mnemonics? Language Teaching Research, 11(1), 43-62. Boers, F. & Lindstromberg, S., (Eds.) (2008a). Cognitive Linguistic Approaches to Teaching Vocabulary and Phraseology. Berlin: Mouton de Gruyter. Boers, F. & Lindstromberg, S. (2008b). From empirical findings to pedagogical practice. In F. Boers & S. Lindstromberg (Eds.), Cognitive Linguistic Approaches to Teaching Vocabulary and Phraseology (pp. 375-393). Amsterdam: John Benjamins. Boers, F. & Lindstromberg, S. (2008c). How cognitive linguistics can foster e ffective vocabulary teaching. In F. Boers & S. Lindstromberg (Eds.), Cognitive Linguistic Approaches to Teaching Vocabulary and Phraseology (pp. 1-61). Amsterdam: John Benjamins. Börner, W. & Vogel, K. (1997). Mentales Lexikon und Lernersprache. In W. Börner & K. Vogel (Eds.), Kognitive Linguistik und Fremdsprachenerwerb: Das Mentale Lexikon (pp. 1-17). Tübingen: Narr. Bortz, J. & Döring, N. (2006). Forschungsmethoden und Evaluation für Human- und Sozialwissenschafter. 4. Auflage. Heidelberg: Springer. Bredenbröcker, M., Elsner, D., Gleixner, S., Gutwerk, S., Lugauer, M., & Spangenberg, A. (2005a). Sally Lehrwerk für den Englischunterricht ab Klasse 3 - Lehrermaterialien 3. München: Oldenbourg. Bredenbröcker, M., Elsner, D., Gleixner, S., Gutwerk, S., Lugauer, M., & Spangenberg, A. (2005b). Sally Lehrwerk für den Englischunterricht ab Klasse 3 - Schülerbuch 3. München: Oldenbourg. Brewster, J., Ellis, G., & Girard, D. (2002). The Primary English Teacher’s Guide - New Edition. Harlow: Penguin. Cameron, L. (2001). Teaching Languages to Young Learners. Cambridge: CUP. 146 <?page no="163"?> Carter, R. (1998). Vocabulary. Applied Linguistic Perspective. Second Edition. London: Routledge. Carter, R. (2001). Vocabulary. In R. Carter & D. Nunan (Eds.), The Cambridge Guide to Teaching English to Speakers of Other Languages (pp. 42-47). Cambridge: CUP. Cenoz, J., Hufeisen, B., & Jessner, U., (Eds.) (2003). The Multilingual Lexicon. Dordrecht: Kluwer. Chomsky, N. (1980). Rules and Representations. Oxford: Blackwell. Chomsky, N. (1988). Language and Problems of Knowledge: The Nicaraguan Lectures. Cambridge, MA: MIT Press. Coady, J. (1997a). L2 vocabulary acquisition: A synthesis of the research. In J. Coady & T. Huckin (Eds.), Second Language Vocabulary Acquisition: A Rationale for Pedagogy (pp. 273-290). Cambridge: CUP. Coady, J. (1997b). L2 vocabulary acquisition through extensive reading. In J. Coady & T. Huckin (Eds.), Second Language Vocabulary Acquisition: A Rationale for Pedagogy (pp. 225-237). Cambridge: CUP. Condon, N. (2008). How cognitive linguistic motivations influence the learning of phrasal verbs. In F. Boers & S. Lindstromberg (Eds.), Cognitive Linguistic Approaches to Teaching Vocabulary and Phraseology (pp. 133-158). Berlin: Mouton de Gruyter. Conklin, K. & Schmitt, N. (2008). Formulaic sequences: Are they processed more quickly than nonformulaic language by native and nonnative speakers? Applied Linguistics, 29, 72-89. Corson, D. J. (1995). Using English Words. Dordrecht: Kluwer. Council of Europe (2001). Common European Framework of Reference for Languages: Learning, teaching, assessment. Cambridge: CUP. Craik, F. I. M. & Lockhart, R. S. (1972). Levels of processing: A framework for memory research. Journal of Verbal Learning and Verbal Behavior, 11, 671-684. Craik, F. I. M. & Tulving, E. (1975). Depth of processing and the retention of words in episodic memory. Journal of Experimental Psychology: General, 104, 268-294. Croft, W. (2001). Radical Construction Grammar: Syntactic Theory from a Typological Perspective. Oxford: OUP. De Bot, K. (1992). A bilingual production model: Levelt’s ’speaking’ model adapted. Applied Linguistics, 13, 1-24. De Bot, K. (2003). Bilingual speech: from concepts to articulation. Fremdsprachen Lehren und Lernen, 32, 92-103. De Bot, K. (2004). The multilingual lexicon: Modeling selection and control. International Journal of Multilingualism, 1, 17-32. 147 <?page no="164"?> De Bot, K., Lowie, W., & Verspoor, M., (Eds.) (2005). Second Language Acquisition - An Advanced Resource Book. London: Routledge. De Bot, K., Lowie, W., & Verspoor, M. (2007). A Dynamic Systems Theory approach to second language acquisition. Bilingualism: Language and Cognition, 10, 7-21. De Bot, K., Paribakht, T. S., & Wesche, M. B. (1997). Toward a lexical processing model for the study of second language vocabulary acquisition. Studies in Second Language Acquisition, 19, 309-329. De Groot, A. M. B. (1993). Word-type e ffects in bilingual processing tasks: Support for a mixed-representational system. In R. Schreuder & B. Weltens (Eds.), The Bilingual Lexicon (pp. 27-51). Amsterdam: John Benjamins. De Groot, A. M. B. (1995). Determinants of bilingual lexicosemantic organisation. Computer Assisted Language Learning, 8, 151-180. De Groot, A. M. B. & Poot, R. (1997). Word translation at three levels of proficiency in a second language: the ubiquitous involvement of conceptual memory. Language Learning, 47, 215-264. Dijkstra, T. (2003). Lexical processing in bilinguals and multilinguals: The word selection problem. In J. Cenoz, B. Hufeisen, & U. Jessner (Eds.), The Multilingual Lexicon (pp. 11-26). Dordrecht: Kluwer. Do ff , S. & Klippel, F. (2007). Englischdidaktik - Praxishandbuch für die Sekundarstufe I und II. Berlin: Cornelsen Scriptor. Dong, Y., Gui, S., & MacWhinney, B. (2005). Shared and separate meaning in the bilingual mental lexicon. Bilingualism: Language and Cognition, 8, 221-238. Eckerth, J. (2008). Investigating consciousness-raising tasks: Pedagogically targeted and non-targeted learning gains. International Journal of Applied Linguistics, 18, 119-145. Ehlers, G., Michailow-Drews, U., Rohde, A., & Zeich-Pelsis, A.-K. (2006). See What I can do in English. Lernstandserhebung 3 / 4. Braunschweig: Schroedel. Ellis, N. (2002). Frequency e ffects in language processing: A review with implications for theories of implicit and explicit language acquisition. Studies in Second Language Acquisition, 24, 143-188. Ellis, N. & Beaton, A. (1993). Psycholinguistic determinants of foreign language vocabulary learning. Language Learning, 43, 559-617. Ellis, N. C. (1994). Vocabulary acquisition: the implicit ins and outs of explicit cognitive mediation. In N. C. Ellis (Ed.), Implicit and Explicit Learning of Languages (pp. 211-282). London: Academic Press. Ellis, N. C. (1995). Vocabulary acquisition: Psychological perspectives 148 <?page no="165"?> and pedagogical implications. The Language Teacher, 19, 12-16. Ellis, N. C. (1996). Sequencing in SLA: phonological memory, chunking, and points of order. Studies in Second Language Acquisition, 18, 91-126. Ellis, N. C. (1997a). Vocabulary acquisition: word structure, collocations, word-class, and meaning. In N. Schmitt & M. McCarthy (Eds.), Vocabulary: Description, Acquisition, Pedagogy (pp. 122-139). Cambridge: CUP. Ellis, N. C. (2001). Memory for language: Phonological memory, chunking, and points of order. In P. Robinson (Ed.), Cognition and Second Language Instruction (pp. 33-68). Cambridge: CUP. Ellis, R. (1997b). SLA Reseach and Language Teaching. Oxford: OUP. Ellis, R. (1999). Item versus system learning: Explaining free variation. Applied Linguistics, 20, 460-480. Ellis, R. (2003). Task-based Language Teaching and Learning. Oxford: OUP. Field, A. (2005). Discovering Statistics Using SPSS - Second Edition. London: Sage. Fodor, J. A. (1983). The Modularity of Mind. Cambridge, MA: MIT Press. Folse, K. S. (2004). Vocabulary Myths. Applying Second Language Research to Classroom Teaching. Ann Arbor, MI: The University of Michigan Press. Geeraerts, D. (2006). Introduction: A rough guide to Cognitive Linguistics. In D. Geeraerts (Ed.), Cognitive Linguistics: Basic Readings (pp. 1-28). Berlin: Mouton de Gruyter. Gleason, J. (1958). The child’s learning of English morphology. Word, 14, 150-177. Gullberg, M. & Indefrey, P., (Eds.) (2006). The Cognitive Neuroscience of Second Language Acquisition. Oxford: Blackwell. Hagoort, P. (2006). What we cannot learn from neuroanatomy about language learning and language processing. Comment on Uylings. In M. Gullberg & P. Indefrey (Eds.), The Cognitive Neuroscience of Second Language Acquisition (pp. 91-97). Oxford: Blackwell. Handke, J. (1997). Zugri ffsmechanismen im mentalen und maschinellen Lexikon. In W. Börner & K. Vogel (Eds.), Kognitive Linguistik und Fremdsprachenerwerb: Das Mentale Lexikon (pp. 89-106). Tübingen: Narr. Hatch, E. & Brown, C. (1995). Vocabulary, Semantics, and Language Education. Cambridge: CUP. Hollbrügge, B. & Kraaz, U. (2003a). Ginger 1 Lehrwerk für den früh beginnenden Englischunterricht - Activity Book Klasse 3. Berlin: Cornelsen. Hollbrügge, B. & Kraaz, U. (2003b). Ginger 1 Lehrwerk für den früh beginnenden Englischunterricht - Handreichung für den Unterricht Klasse 3. Berlin: Cornelsen. 149 <?page no="166"?> Hollbrügge, B. & Kraaz, U. (2004a). Ginger 2 Lehrwerk für den früh beginnenden Englischunterricht - Activity Book Klasse 4. Berlin: Cornelsen. Hollbrügge, B. & Kraaz, U. (2004b). Ginger 2 Lehrwerk für den früh beginnenden Englischunterricht - Handreichung für den Unterricht Klasse 4. Berlin: Cornelsen. Hulstijn, J. H. (1997). Mnemonic methods in foreign language vocabulary learning: Theoretical considerations and pedagogical implications. In J. Coady & T. Huckin (Eds.), Second Language Vocabulary Acquisition: A Rationale for Pedagogy (pp. 203-224). Cambridge: CUP. Hulstijn, J. H. (2001). Intentional and incidental second language vocabulary learning: a reappraisal of elaboration, rehearsal and automaticity. In P. Robinson (Ed.), Cognition and Second Language Instruction (pp. 258-286). Cambridge: CUP. Iannou-Georgiou, S. & Pavlou, P. (2003). Assessing Young Learners. Oxford: OUP. Joe, A. (1995). Text-based tasks and incidental vocabulary learning. Second Language Research, 11, 149-158. Johnson, K. (2008). An Introduction to Foreign Language Learning and Teaching - 2nd edition. Learning about Language. London: Longman. Johnstone, R. (2000). Context-sensitive assessment of modern languages in primary (elementary) and early secondary education: Scotland and the European experience. Language Testing, 17, 123-143. Kemmer, S. & Barlow, M. (2000). Introduction: A usage-based conception of language. In M. Barlow & S. Kemmer (Eds.), Usage-Based Models of Language (pp. vii-xxviii). Stanford, CA: CSLI Publications. Klippel, F. (1995). Wörternetze. In K.-R. Bausch, H. Christ, F. G. Königs, & H.-J. Krumm (Eds.), Erwerb und Vermittlung von Wortschatz im Fremdsprachenunterricht: Arbeitspapiere der 15. Frühjahrskonferenz zur Erforschung des Fremdsprachenunterrichts (pp. 101-107). Tübingen: Narr. Ordner Literatur IV. Knapp-Pottho ff , A. (2000). Vokabeln im Kopf. In C. Riemer (Ed.), Kognitive Aspekte des Lehrens und Lernens fremder Sprachen. Cognitive Aspects of Foreign Language Learning and Teaching. Festschrift für Willis J. Edmondson zum 60. Geburtstag (pp. 293-307). Tübingen: Narr. Krashen, S. (1981). Second Language Acquisition and Learning. Oxford: Pergamon. Kroll, J. F. (1993). Accessing conceptual representations for words in a second language. In R. Schreuder & B. Weltens (Eds.), The Bilingual Lexicon (pp. 53-81). Amsterdam: John Benjamins. Kroll, J. F. & Stewart, E. (1994). Category interference in translation 150 <?page no="167"?> and picture naming evidence for asymmetric connections between bilingual memory representations. Journal of Memory and Language, 33, 149-174. Krumm, H.-J. (1995). Die Bedeutung von Wortschatz und Wortschatzerwerb in einem integrativen Fremdsprachenunterricht. In K.-R. Bausch, H. Christ, F. G. Königs, & H.-J. Krumm (Eds.), Erwerb und Vermittlung von Wortschatz im Fremdsprachenunterricht: Arbeitspapiere der 15. Frühjahrskonferenz zur Erforschung des Fremdsprachenunterrichts (pp. 119- 125). Tübingen: Narr. Ordner Literatur IV. Kuiper, K. (2004). Formulaic preformance in conventionalised varieties of speech. In N. Schmitt (Ed.), Formulaic Sequences. Acquisition, processing and use. (pp. 37-54). Amsterdam: Benjamins. Lako ff , G. (1987). Women, Fire, and Dangerous Things - What Categories Reveal about the Mind. Chicago, IL: University of Chicago Press. Langacker, R. W. (2000). A dynamic usage-based model. In M. Barlow & S. Kemmer (Eds.), Usage-Based Models of Language (pp. 1-63). Stanford, CA: CSLI Publications. Langacker, R. W. (2008). Cognitive Grammar: A basic introduction. Oxford: OUP. Larsen-Freeman, D. & Cameron, L. (2008). Complex Systems and Applied Linguistics. Oxford: OUP. Laufer, B. (1997). The lexical plight in second language reading: Words you don’t know, words you think you know, and words you can’t guess. In J. Coady & T. Huckin (Eds.), Second Language Vocabulary Acquisition: A Rationale for Pedagogy (pp. 20-34). Cambridge: CUP. Laufer, B. & Hulstijn, J. (2001). Incidental vocabulary acquisition in a second language: the construct of task-induced involvement. Applied Linguistics, 22, 1-26. Lazar, G. (2003). Meaning and Metaphor: Activities to Practise Figurative Language. Cambridge: CUP. Lefrançois, G. R. (2006). Theories of Human Learning. What the old woman said (5th edition). Belmont, CA: Thomson Wadsworth. Levelt, W. J. M. (1989). Speaking: From intention to articulation. Cambridge, MA: MIT Press. Levelt, W. J. M. (1993). The architecture of normal spoken language use. In G. Blanken, E. Dittmann, H. Grimm, J. Marshall, & C. Wallesch (Eds.), Linguistic Disorders and Pathologies. An International Handbook (pp. 1-15). Berlin: Mouton de Gruyter. Lewis, M. (1993). The Lexical Approach. Hove: Language Teaching Publications. 151 <?page no="168"?> Lewis, M. (1997a). Implementing the Lexical Approach. Hove: Language Teaching Publications. Lewis, M. (1997b). Pedagogical implications of the lexical approach. In J. Coady & T. Huckin (Eds.), Second Language Vocabulary Acquisition: A Rationale for Pedagogy (pp. 255-270). Cambridge: CUP. Lightbown, P. M. & Spada, N. (2006). How Languages are Learned. 3rd Edition. Oxford: OUP. Lowie, W. (2000). Cross-linguistic influence on morphology in the bilingual mental lexicon. Studia Linguistica, 54, 175-185. Lowie, W. & Verspoor, M. (2004). Input versus transfer? The role of frequency and similarity in the acquisition of L2 prepositions. In M. Achard & S. Niemeier (Eds.), Cognitive Linguistics, Second Language Acquisition, and Foreign Language Learning (pp. 77-93). Berlin: Mouton de Gruyter. Lutjeharms, M. (2004). Der Zugri ffauf das mentale Lexikon und der Wortschatzerwerb in der Fremdsprache. Fremdsprachen Lehren und Lernen, 33, 10-26. MacArthur, F. & Littlemore, J. (2008). A discovery approach to figurative language learning with the use of corpora. In F. Boers & S. Lindstromberg (Eds.), Cognitive Linguistic Approaches to Teaching Vocabulary and Phraseology (pp. 159-188). Berlin: Mouton de Gruyter. Marslen-Wilson, W. D. & Tyler, L. (1980). The temporal structure of spoken language understanding. Cognition, 8, 1-71. McCarthy, M. (1984). A new look on vocabulary in EFL. Applied Linguistics, 5, 12-22. McClelland, J. L. & Rumelhart, D. (1981). An interactive activation model of context e ffects in letter perception, Part I: An account of basic findings. Psychological Review, 88, 375-405. McKay, P. (2006). Assessing Young Language Learners. Cambridge: CUP. McNeill, D. (1979). The Conceptual Basis of Language. Hillsdale, N.J: Erlbaum. Meara, P. (1980). Vocabulary acquisition: A neglected aspect of language learning. Language Teaching and Linguistics Abstracts, 13, 221-246. Meara, P. (1993). The bilingual lexicon and the teaching of vocabulary. In R. Schreuder & B. Weltens (Eds.), The Bilingual Lexicon (pp. 279-297). Amsterdam: John Benjamins. Meara, P. (1997). Towards a new approach to modelling vocabulary acquisition. In N. Schmitt & M. McCarthy (Eds.), Vocabulary: Description, Acquisition, Pedagogy (pp. 109-121). Cambridge: CUP. Meara, P. (2002). The rediscovery of vocabulary. Second Language Research, 152 <?page no="169"?> 18, 393-407. Meara, P. (2004). Modelling vocabulary loss. Applied Linguistics, 25, 137- 155. Meara, P. (2006). Emergent properties of multilingual lexicons. Applied Linguistics, 27, 620-644. Meara, P. & Wolter, B. (2004). V-Links: Beyond vocabulary depth. In D. Albrechtsen, K. Haastrup, & B. Henriksen (Eds.), Angles on the English speaking world Vol. IV: Writing & Vocabulary in Foreign Language Acquisition (pp. 88-97). Copenhagen: Museum Tusculanum Press / University of Copenhagen. Müller, H. M. (2003). Neurobiologische Aspekte des Fremdsprachenlernens. Fremdsprachen Lehren und Lernen, 32, 167-177. Möhle, D. (1997). Deklaratives und prozedurales Wissen in der Repräsentation des mentalen Lexikons. In W. Börner & K. Vogel (Eds.), Kognitive Linguistik und Fremdsprachenerwerb: Das Mentale Lexikon (pp. 39-49). Tübingen: Narr. Moir, J. & Nation, P. (2008). Vocabulary and good language learners. In C. Gri ffi ths (Ed.), Lessons from Good Language Learners (pp. 159-173). Cambridge: CUP. Morton, J. (1969). Interaction of information in word recognition. Psychological Review, 76, 165-178. Nagy, W. & Herman, P. (1985). Incidental vs. instructional approaches to increasing reading vocabulary acquisition. Contemporary Educational Psychology, 17, 286-292. Nation, I. S. P. (1990). Teaching and Learning Vocabulary. Boston, MA: Heinle. Nation, I. S. P. (2001). Learning Vocabulary in Another Language. Cambridge: CUP. Nation, P. (1993). Vocabulary size, growth, and use. In R. Schreuder & B. Weltens (Eds.), The Bilingual Lexicon (pp. 115-156). Amsterdam: John Benjamins. Nation, P. & Gu, P. Y. (2007). Focus on Vocabulary. Sydney: NCELTR Publishing. Nation, P. & Newton, J. (1997). Teaching vocabulary. In J. Coady & T. Huckin (Eds.), Second Language Vocabulary Acquisition: A Rationale for Pedagogy (pp. 238-254). Cambridge: CUP. Nattinger, J. R. & DeCarrico, J. S. (1992). Lexical Phrases and Language Teaching. Oxford: OUP. Neveling, C. (2004). Wörternetze als Abbilder mentale Lexikonstrukturen - Untersuchungen zur Güte eines neuen Forschungsvorhabens. 153 <?page no="170"?> Fremdsprachen Lehren und Lernen, 33, 128-146. Neveling, C. (2005). Die Konstruktion von Wörternetzen. Eine Speicherstrategie nach dem mentalen Lexikon. PRAXIS Fremdsprachenunterricht, 2, 28-32. Näf, A. (2004). Assoziationsgesteuerte Lexemsuche in der L2. In B. Hufeisen & N. Marx (Eds.), Beim Schwedischlernen sind Englisch und Deutsch ganz hilfsvoll - Untersuchungen zum multiplen Sprachenlernen (pp. 137-162). Frankfurt am Main: Peter Lang. Nicholas, H. & Lightbown, P. M. (2008). Defining child second language acquisition, defining roles for L2 instruction. In J. Philp, R. Oliver, & A. Mackey (Eds.), Second Language Acquisition and the Younger Learner (pp. 27-51). Amsterdam: John Benjamins. Niedersächsisches Kultusministerium (1995). Didatisch-methodische Empfehlungen für das Fremdsprachenlernen in der Grundschule. Hannover: Schroedel. Niedersächsisches Kultusministerium (2006). Kerncurriculum für die Grundschule Schuljahrgänge 3-4. http: / / db2.nibis.de / 1db / cuvo / ausgabe / . Niemeier, S. (2005). Applied Cognitive Linguistics and newer trends in foreign language teaching methodology. In A. E. Tyler, M. Takada, Y. Kim, & D. Marinova (Eds.), Language in Use. Cognitive and Discourse Perspectives on Language and Language Learning (pp. 100-111). Washington: Georgetown University Press. Niemeier, S. (2008). The notion of boundedness / unboundedness in the foreign language classroom. In F. Boers & S. Lindstromberg (Eds.), Cognitive Linguistic Approaches to Teaching Vocabulary and Phraseology (pp. 307-327). Amsterdam: John Benjamins. Obler, L. K. & Gjerlow, K. (1999). Language and the Brain. Cambridge: CUP. O’Gorman, E. (1996). An investigation of the mental lexicon of second language learners. TEANGA, 16, 15-31. Oliver, R., Philp, J., & Mackey, A. (2008). The impact of teacher input, guidance and feedback on ESL children’s task-based interactions. In J. Philp, R. Oliver, & A. Mackey (Eds.), Second Language Acquisition and the Younger Learner (pp. 131-147). Amsterdam: John Benjamins. Oppenheim, N. (2000). The importance of recurrent sequences for nonnative speaker fluency and cognition. In H. Riggenbach (Ed.), Perspectives on Fluency. Ann Arbor, MI: University of Michigan Press. Paradis, J. (2007). Second language acquisition in childhood. In E. Ho ff & M. Shatz (Eds.), Blackwell Handbook of Language Development (pp. 154 <?page no="171"?> 387-406). Malden, MA: Blackwell. Paradis, M. (2004). A Neurolinguistic Theory of Bilingualism. Amsterdam: John Benjamins. Penke, M. (2006). Flexion im Mentalen Lexikon. Tübingen: Niemeyer. Philp, J., Mackey, A., & Oliver, R. (2008). Child’s play? Second language acquisition and the younger learner in context. In J. Philp, R. Oliver, & A. Mackey (Eds.), Second Language Acquisition and the Younger Learner (pp. 3-23). Amsterdam: John Benjamins. Piquer Píriz, A. M. (2008). Reasoning figuratively in early EFL: Some implications for the development of vocabulary. In F. Boers & S. Lindstromberg (Eds.), Cognitive Approaches to Teaching Vocabulary and Phraseology (pp. 219-240). Berlin: Mouton de Gruyter. Randall, M. (2007). Memory, Psychology and Second Language Learning. Amsterdam: John Benjamins. Raupach, M. (1997). Das mehrsprachige mentale Lexikon. In W. Börner & K. Vogel (Eds.), Kognitive Linguistik und Fremdsprachenerwerb: Das Mentale Lexikon (pp. 19-37). Tübingen: Narr. Rea-Dickins, P. (2000). Assessment in early years language learning contexts. Language Testing, 17, 115-122. Read, J. (2004a). Plumbing the depths: How should the construct of vocabulary knowledge be defined? In P. Bogaards & B. Laufer (Eds.), Vocabulary in a Second Language: Selection, Acquisition, and Testing (pp. 209-227). Amsterdam: John Benjamins. Read, J. (2004b). Research in teaching vocabulary. Annual Review of Applied Linguistics, 24, 146-161. Robinson, P. (2007). Criteria for classifying and sequencing pedagogic tasks. In M. d. P. García Mayo (Ed.), Investigating Tasks in Formal Language Learning (pp. 7-26). Clevedon: Multilingual Matters. Rosch, E. (1975). Cognitive representations of semantic categories. Journal of Experimental Psychology: General, 104, 192-233. Rost, D. H. (2007). Interpretation und Bewertung pädagogisch-psychologischer Studien. Weinheim: UTB. Rubin, J. & McCoy, P. (2008). Tasks and good language learners. In C. Gri ffi ths (Ed.), Lessons from Good Language Learners (pp. 294-305). Cambridge: CUP. Rudzka-Ostyn, B. (2003). Word Power: Phrasal Verbs and Compounds. A Cognitive Approach. Berlin: Mouton de Gruyter. Scherfer, P. (1997). Überlegungen zu einer Theorie des Vokabellernens und -lehrens. In W. Börner & K. Vogel (Eds.), Kognitive Linguistik und Fremdsprachenerwerb: Das Mentale Lexikon (pp. 185-215). Tübingen: 155 <?page no="172"?> Narr. Schmidt, R. (2000). Attention. In P. Robinson (Ed.), Cognition and Second Language Instruction. Cambridge: CUP. Schmitt, N. (1997). Vocabulary learning strategies. In N. Schmitt & M. McCarthy (Eds.), Vocabulary: Description, Acquisition, Pedagogy (pp. 199-227). Cambridge: CUP. Schmitt, N. (2000). Vocabulary in Language Teaching. Cambridge: CUP. Schmitt, N., Schmitt, D., & Clapham, C. (2001). Developing and exploring the behaviour of two new versions of the Vocabulary Levels Test. Language Testing, 18, 55-88. Schoonen, R. & Verhallen, M. (2008). The assessment of deep work knowledge in young first and second language learners. Language Testing, 25, 211-236. Selinker, L. (1972). Interlanguage. International Review of Applied Linguistics, 10, 209-231. Singleton, D. (1999). Exploring the Second Language Mental Lexicon. Cambridge: CUP. Singleton, D. (2000). Language and the Lexicon. An Introduction. London: Arnold. Singleton, D. (2003). Perspectives on the multilingual lexicon: A critical synthesis. In J. Cenoz, B. Hufeisen, & U. Jessner (Eds.), The Multilingual Lexicon (pp. 167-176). Dordrecht: Kluwer. Singleton, D. (2007). How integrated is the integrated mental lexicon? In Z. Lengyel & J. Navracsics (Eds.), Second Language Lexical Processes. Applied Linguistic and Psycholinguistic Perspectives (pp. 3-16). Clevedon: Multilingual Matters. Skoufaki, S. (2008). Conceptual metaphoric meaning clues in two idiom presentation methods. In F. Boers & S. Lindstromberg (Eds.), Cognitive Linguistic Approaches to Teaching Vocabulary and Phraseology (pp. 102- 132). Berlin: Mouton de Gruyter. Snodgrass, J. G. (1993). Translating versus picture naming: Similarities and di fferences. In R. Schreuder & B. Weltens (Eds.), The Bilingual Lexicon (pp. 83-114). Amsterdam: John Benjamins. Söderman, T. (1993). Word associations of foreign language learners and native speakers - A shift in response type and its relevance for the theory of lexical development. In H. Ringbom (Ed.), Near-native proficiency in English. Åbo: Åbo Akademi. Sökmen, A. J. (1997). Current trends in teaching second language vocabulary. In N. Schmitt & M. McCarthy (Eds.), Vocabulary: Description, Acquisition, Pedagogy (pp. 237-257). Cambridge: CUP. 156 <?page no="173"?> Sparks, R., Ganshow, L., & Javorsky, J. (2000). Déja vu all over again: A response to Saitu, Horrowitz and Garza. Modern Language Journal, 84, 251-255. Spöttl, C. & McCarthy, M. (2003). Formulaic utterances in the multilingual context. In J. Cenoz, B. Hufeisen, & U. Jessner (Eds.), The Multilingual Lexicon (pp. 133-151). Dordrecht: Kluwer. Stahl, S. A. & Fairbanks, M. M. (1986). The e ffects of vocabulary instruction: A model-based meta-analysis. Review of Educational Research, 56, 72-110. Stork, A. (2003). Vokabellernen. Eine Untersuchung zur E ffi zienz von Vokabellernstrategien. Tübingen: Narr. Thornbury, S. (2002). How to Teach Vocabulary. Harlow: Longman. Tomasello, M. (2000). First steps towards a usage-based theory of language acquisition. Cognitive Linguistics, 11, 61-82. Tomasello, M. (2003). Constructing a Language - A Usage-Based Theory of Language Acquisition. Cambridge, MA: Harvard University Press. Tyler, A. E., Takada, M., Kim, Y., & Marinova, D., (Eds.) (2005). Language in Use. Cognitive and Discourse Perspectives on Language and Language Learning. Washington: Georgetown University Press. Van de Guchte, C. & Vermeer, A. (2003). Een passende woordkeus. Het kiezen van woorden voor woordenschatlessen. Toegepaste Taalwetenschap in Artikelen, 69, 9-23. Van der Rhoer, F. & Vermeer, A. (2005). Hiërarchische relaties en woordenschatonderwijs. Toegepaste Taalwetenschap in Artikelen, 73, 99-110. Van Geert, P. (1995). Dimensions of change: A semantic and mathematical analysis of learning and development. Behavioral and Brain Sciences, 21, 615-656. Verhallen, M. & Schoonen, R. (1998). Lexical knowledge in L1 and L2 of third and fifth graders. Applied Linguistics, 19(4), 452-470. Verspoor, M. (2008). What bilingual word associations can tell us. In F. Boers & S. Lindstromberg (Eds.), Cognitive Linguistic Approaches to Teaching Vocabulary and Phraseology (pp. 261-289). Amsterdam: John Benjamins. Verspoor, M. & Lowie, W. (2003). Making sense of polysemous words. Language Learning, 53, 547-586. Waring, R. (1997). The negative e ffects of learning words in semantic sets. System, 25, 261-274. Weinreich, U. (1953). Languages in Contact. Findings and Problems. New York: Humanities Press. Willis, D. & Willis, J. (2001). Task-based language learning. In R. Carter 157 <?page no="174"?> & D. Nunan (Eds.), The Cambridge Guide to Teaching English to Speakers of Other Languages (pp. 172-179). Cambridge: CUP. Willis, J. (1996). A Framework for Task-Based Learning. London: Longman. Wol ff , D. (2002). Das mentale Lexikon - Grundlage der Sprachkompetenz in der Muttersprache und der Fremdsprache. Der fremdsprachliche Unterricht Englisch, 36, 11-16. Wolter, B. (2001). Comparing the L1 and L2 mental lexicon. A depth of individual word knowledge model. Studies in Second Language Acquisition, 23, 41-69. Wolter, B. (2006). Lexical network structures and L2 vocabulary acquisition: The role of the L1 lexical / conceptual knowledge. Applied Linguistics, 27(4), 741-747. Wray, A. (1999). Formulaic language in learners and native speakers. Language Teaching, 32, 213-231. Wray, A. (2000). Formulaic sequences in second language teaching: Principle and practice. Applied Linguistics, 21, 463-489. Wray, A. (2002). Formulaic Language and the Lexicon. Cambridge: CUP. Wray, A. & Perkins, M. R. (2000). The functions of formulaic language: an integrated model. Language & Communication, 20, 1-28. 158 <?page no="175"?> Sources of Illustrations Illustrations used for Lessons in Grade 3 All pictures adapted from Ginger 1 (Hollbrügge & Kraaz 2003a). www.vskrems-lerchenfeld.ac.at / arbeitsmaterialien / anlautbilder / birne.jpg www.vivat-lingua.de / wortderwoche_archiv.htm 159 <?page no="176"?> All pictures Saskia Kersten All pictures adapted from Ginger 1 (Hollbrügge & Kraaz 2003a). adapted from www.clipartguide.com / _search_terms / snack.html fish adapted from www.prepolino.ch / bildersammlung / tiere / fisch / fisch.gif Saskia Kersten 160 <?page no="177"?> www.landmaschinen.tu-dresden.de / Lehrstuhl / Forschung / Forschung_alt / Kuh / kuh.html www.schafplanet.de / klugschaf / Wissen / schafe / schafe1.html www.kilde.com / de / portfolio / projektarchiv / 9465_salzwerk / 9465_1.html All pictures Saskia Kersten All pictures adapted from vilt creaties by HEMA. 161 <?page no="178"?> All pictures adapted from vilt creaties by HEMA. 162 <?page no="179"?> Illustrations used for Lessons in Grade 4 adapted from www.cadburyschweppes.com / NR / rdonlyres / 2E56A6EF- 7A17-4DBB-B738-1A01AD78767F / 0 / cdm200.jpg adapted from www.spielhandlung.de / pd1212772794.htm? categoryId = 33 adapted from www.haba.de / Knabbereien_Sue.knabber.0.html? &tx _habaproducts_pi1[showUid] = 394&tx_habaproducts _pi1[toysView] = 1&tx_habaproducts _pi1[pointer] = 1&cHash = dae7548db5 #hp_toy_1381 adapted from www.sainsburys.com / groceries / index.jsp? bmUID = 1209118529542 adapted from www.worldofstock.com / closeups / PFO3529.php adapted from images.mytoys.com / intershoproot / eCS / Store / de / images / 100 / 63 / 1006380-l.jpg adapted from www.indfish.co.nz / products / retail / batteredfishfingers.php en.wikipedia.org / wiki / Fish_finger 163 <?page no="180"?> All pictures adapted from vilt creaties by HEMA. 164 <?page no="181"?> adapted from www.tidespoint.com / food / tetley.shtml www.istockphoto.com / file _closeup / style_and_design / illustrations / vector_objects / 343840 _pot_of_jam.php? id = 343840 www.united-states-map.org / americanflag.htm www.postershop.co.uk / Anonymous / Anonymous- Union-Jack-5000572.html www.fundraw.com / clipart / categories / Everyday_Objects / Toys_-_Balloons / 00000047 www.psfk.com / 2007 / 11 / uwe-lubbermanntalks-to-psfk-about-premium-cola.html www.thesage.com / images / prod / SmallTub.jpg www.abcteach.com / free / i / icecreamcone2rgb.jpg 165 <?page no="182"?> adapted from www.sainsburys.com / media / images / products / AWSF001 / 5000328750255_AWSF001_2_Spec2 _v1_m56577569833618481.jpg www.wdr5.de / service / service_kleine_anfrage / 816135.phtml adapted from wwww.cadbury.com.au / sites / cadbury / index.php? pageId = 92 www.kado.de / index.php / cat / c10_Pfe fferminzund-Bonbons.html www.einzinger.de / catalog / product_info.php? cPath = 5 _36&products_id = 1725 &osCsid = 29555c6c7610e5873cd4ac4fb14158d7 www.stomp.com.sg / stfoodiesclub / poshnosh / 88 / lemon.html 166 <?page no="183"?> All pictures adapted from Ginger (Hollbrügge & Kraaz 2004a). 167 <?page no="185"?> Appendix <?page no="187"?> Appendix A Lesson plans Preliminary remarks: • All lessons are for elementary level learners and have the aim of introducing and / or practising new vocabulary. • The exact location and time of the lessons are omitted to preserve anonymity. • All lessons are planned to fit into a 45 minute slot. • Class composition is described in detail in Chapter 11.2. • The intervention was preceded by one lesson in order for the new teacher, responsible for teaching the lessons during the intervention study, and the learners to become familiar with one another. • All the lessons are designed for primary school level and conform to the appropriate style of teaching deemed necessary for this level according to the Kerncurriculum (Niedersächsisches Kultusministerium [Lower Saxony Ministry of Education] 2006. • In the following lessons plans, for the sake of brevity, only a brief description of what actually happened in the lessons will be given. It has been done in this way because the main thrust of the work is not methodology. 171 <?page no="188"?> Lessons for 3rd Grade 1st Lesson - Introducing New Vocabulary Vocabulary apple, banana, grapefruit, lemon, orange, plum, pear, strawberry, bitter, sour, sweet Materials basket containing fruit, prepared pieces of fruit for tasting, worksheet for listening comprehension task, large version of the worksheet for display on the chalkboard Sub-aims introducing new vocabulary, practising these vocabulary items, developing basic listening skills, enabling students to make statements about fruit Warming up Activity Intervention Group The teacher distributes copies of a fruit shop picture (see illustration below). The learners look at the pictures and describe what they can. This activates the necessary vocabulary for the lesson and encourages the learners to recycle vocabulary learned before the intervention lessons. Control Group As above Introducing New Vocabulary Intervention Group The learners (LL) form a circle. The teacher (T) places a basket of fruit in the middle of the circle. The basket is covered to hide its contents. The teacher asks a learner to take something out of the basket without looking inside. When the learner has done so the teacher asks the learner what it is. The learner can say the name of the fruit in German or in English. The teacher repeats the name in English, and then asks the class to repeat the name, for example: 172 <?page no="189"?> T - L : What is it? L - T: It’s an orange. T - LL: Yes, it’s an orange. T- LL: What is it? LL - T: (chorally) It’s an orange. The teacher then asks another learner to take something out of the basket. The above is repeated. The teacher names two pieces of the fruit, points to them, and the class repeat. This continues until all eight pieces of fruit have been removed. Control Group As above Game ‘What is missing? ’ Intervention Group The teacher asks the class to close their eyes. S / he then names some of the fruit and asks the class to repeat the names after her / him. During this choral repetition, the teacher quietly removes one or two pieces of fruit and hides them. The learners then open their eyes and are asked to name the missing pieces of fruit. When the learners answer, the teacher corrects the pronunciation, if necessary and puts the correctly named fruit back. This is repeated several times until every piece of fruit has been hidden at least once. The learners then return to their seats. Control Group As above Listening Comprehension Intervention Group T distributes a worksheet that shows eight boxes, each contain three fruit (see illustration below). The teacher models, with the learners, what they should do with the worksheets, until the task is understood by the learners. The teacher also ensures, through question and answer, that the learners know all the fruit 173 <?page no="190"?> depicted. Task: The learners have to listen while the teacher says, ‘Box number one, orange, apple, plum’. They then number the appropriate box, 1. This is repeated until all the boxes are numbered. In the report back stage, a large version of the worksheet is pinned to the board. The learners can check if they have the correct answers. Control Group As above Fruit Tasting, Learning about Flavours Intervention Group The teacher walks around the class, o ffering the learners a piece of fruit. After the L has tasted the fruit the teacher asks what it tastes like. Using appropriate gestures, sounds and facial expressions, the teacher explains the meaning of sweet, sour and bitter. This is repeated until all the learners have a piece of fruit. Each time the teacher elicits what the fruit tastes like. Control Group As above. However, the teacher does not teach the words sweet, sour and bitter, s / he only asks the learners if they like the fruit they have. 174 <?page no="191"?> 2nd Lesson - Practising New Vocabulary, Action Story Vocabulary apple, banana, grapefruit, lemon, orange, plum, pear, strawberry, bitter, sour, sweet Materials flashcards depicting new vocabulary, worksheet What does the fruit taste like? Sub-aims consolidating vocabulary knowledge, developing basic listening skills, practising making statements about flavours Warming up - Reviewing Vocabulary Intervention Group Using flashcards, with pictures of the fruit taught in the previous lesson, the teacher elicits the name of the pieces of fruit. Then the appropriate flashcard is put on the board. The teacher then shows the learners flashcards showing three di fferent facial expressions and elicits the words sweet, sour and bitter. Control Group As above. However, The teacher does not elicit the di fferent tastes: sweet, sour and bitter. Connecting Fruit with Flavours Intervention Group Through a series of questions and answers, the teacher elicits what the various fruit taste like, in the opinion of the learners, for example: T - LL: What does an orange taste like? (Pulls the three faces.) L - T: Sweet. The learner then comes to the board and puts the ‘sweet’ face next to the orange. The teacher then asks the rest of the class if they agree. If somebody disagrees, the teacher explains that taste is a matter of opinion. This is repeated for every piece of fruit. Control Group The teacher shows various pictures of fruit and elicits what fruit the learners like and do not like. 175 <?page no="192"?> Action Story Intervention Group The learners are asked to stand behind their chairs. The teacher explains that they are about to hear a story of Lucy and her grandmother. While reading the story, the teacher uses suitable gestures, facial expressions and mimes to help the learners understand the story. Di fferent voices are used for Lucy and for the grandmother. The following story is told to the learners who are encouraged to join in. It is repeated eight times, with minor variations. 1) Oh, Grandma, I’m so hungry! (T holds her / his tummy and looks pitiful) I’ve got a hole in my tummy. (T draws a circle in front of his / her tummy) Lucy, look! There is everything here! (T makes a large gesture as if o ffering an variety of food) What do you want, my dear? (T draws a large question mark in the air) Grandma, can I have some fruit? (begging gesture) Sure, have some (grapefruit), it is really good. (T gives the thumbs-up) Brrr, grandma, brr, the (grapefruit) is so (bitter)! (T makes the ´bitter ’ face and shudders) 2-7) Oh, Grandma, I’m still so hungry! (T holds her / his tummy and looks pitiful) T repeats the story inserting a di fferent fruit and a disagreeable flavour each time. 8) T tells the story one last time, this time using strawberry and sweet using a di fferent last line: Mmm, grandma, mmm, the strawberry is so sweet! Thank you, Grandma! (T mimes embracing someone) Control Group The teacher does neither use the words bitter and sour nor pulls the 176 <?page no="193"?> appropriate faces. Instead s / he says that all the fruit is yukky. In the last line of the story the word yummy is substituted for the word sweet. Consolidating New Knowledge Intervention Group T distributes the worksheet ’What does the fruit taste like? ’ ( illustration below). The teacher elicits what the three faces mean. A learner is asked, “What does an orange taste like? " The learner answers, for example, “It’s bitter." The teacher then draws an orange next to the bitter face. This is modelled until the learners understand what they have to do. They are then put into pairs, ask their partner what they think the various fruit tastes like and fill in the worksheets. At the report back stage, the teacher asks the learners to say what their partners think the various pieces of fruit taste like. Control Group T distributes the worksheet ‘What does the fruit taste like? ’ (see illustration below). S / he explains to the learners that the smiley face means ‘I like it’ and the face with the scowl means ‘I don’t like it’. S / he models what the learners should do, that is, draw the appropriate fruit next to the appropriate face. The learners are then put in pairs and find out from their partners what fruit they do and do not like and draw the fruit next to the faces. At the report back stage the learners are asked what fruit their partners do and do not like. 177 <?page no="194"?> 3rd Lesson - Dominoes and ‘Here we go round the lemon tree’ Vocabulary apple, banana, grapefruit, lemon, orange, plum, pear, strawberry, bitter, sour, sweet Materials worksheet What does the fruit taste like? , domino game, large version of domino game for display on the chalkboard, CD-player, flashcards ‘Grows on a tree’ and ‘Does not grow on a tree’ Sub-aims consolidating vocabulary knowledge, developing speaking skills, consolidating recently taught chunks and vocabulary, listening for gist Warming up - Repeating new vocabulary from the previous lesson Intervention Group T - LL: Please take out the worksheet from last week. Which fruit is sour / bitter / sweet? The teacher elicits what fruit is sweet, sour or bitter. If the learners’ opinions di ffer, the teacher encourages each learner to state their own opinion. Control Group The teacher asks the learners what fruit they do and do not like. Dominoes Intervention Group The teacher checks that the learners know how to play dominoes. The rules of this domino game are explained by modelling until the teacher is sure that everyone understands them. The teacher uses a large domino set to illustrate the rules. The learners receive five cards each. The rest of the cards are left in a pile. One card is taken from the pile and placed, face up, in the middle of the table. On this card there is, for example, an apple and a pear. One learner starts (perhaps the youngest). She looks at the card on the 178 <?page no="195"?> table and then at her cards. She might place an orange next to the apple, but has to justify why she has done so, for example: T - L1 : What’s the same? L1 - T: They are both round. T: (looks at L2) L2: They taste sweet. The teacher accepts any combination if the learner can say why they have placed their card next to another card. The learners form five groups and play the game. Control Group As above except the learners match a pear with a pear, and orange with an orange, and so forth. Song ‘Here we go round the lemon tree’ Intervention Group Here we go round the lemon tree, 21 The lemon tree, the lemon tree. Here we go round the lemon tree, Come and dance with me! The learners are asked to form a circle. The teacher explains that they are going to sing a song about a tree. The song is played once. The teacher asks the learners what the song is about and elicits as many answers as possible. A volunteer is asked for and is given a picture of a lemon to hold over his / her head. The rest of the class form a circle around the ‘lemon tree’. The learners dance around the lemon tree and join in with the song. This is repeated for plum and apple tree. Using two flashcards, one with a tree and one with a crossed out tree, the teacher explains that lemons grow on trees. It is then elicited, from the learners, what other fruit grows on trees. The appropriate pictures are placed beside the tree or the crossed out tree. The song is sung one final time. 21 Adapted from Sally 3 (Bredenbröcker et al. 2005b, 24). 179 <?page no="196"?> Control Group The learners are asked to form a circle. The teacher tells the learners they are going to do a rap. S / he then performs the rap once: One banana, two bananas, three bananas, four! Five bananas, six bananas, seven bananas, more! The rap is repeated several times, using the names of di fferent fruit. The learners are encouraged to move in time to the rap. The learners then repeat the rap on their own. 180 <?page no="197"?> 4th Lesson - Introducting pizza vocabulary Vocabulary cheese, corn, ham, mushroom, onion, pepper, pineapple, salami, tomato, tuna Materials flashcards depicting pizza ingredients (whole), flashcards pizza toppings, Bingo cards, paper chips, worksheet Find the right pizza Sub-aims introducing and practising new vocabulary, developing listening skills, developing speaking skills Warming up - Talking about pizza Intervention Group Teacher draws an outline of a pizza on the board. The learners are asked to guess what it is. Any answers in German are translated, by the teacher, into English. The teacher then gives some clues, such as It’s something to eat / It’s hot / Many children really like this. After one of the learners has guessed that it is a pizza, the teachers asks the class what their favourite pizza toppings are: T - LL: What do you like on a pizza. L - T: Tomato If any child uses German, the teacher translates into English: L - T: Käse. T - L: Yes, I like cheese on my pizza, too. While the learners are naming their favourite pizza toppings, the teacher puts flashcards of these on the board. The pictures show the toppings before they have been cut up. If a child names a topping for which there is no card, the teacher searches for the card and then says: T - L: Yes, you can have (broccoli) on a pizza, but I don’t have this in my restaurant. Sorry. Would you like something else on your pizza? The teacher can then list other toppings until the child chooses one. After 181 <?page no="198"?> every new topping is placed on the board, the teacher lists all of them. T-LL: Let’s see what we’ve got on our pizza: onion / mushroom / pepper... The children are then asked to join in and repeat. Control Group As above ‘What’s Missing? ’ Intervention Group In order to recycle the new vocabulary, the teacher asks the learners to close their eyes and then removes one or more flashcards from the board. The learners then open their eyes and say what is missing. This is repeated until each topping has been removed at least once. Control Group As above Listening Comprehension - Pizza Bingo Intervention Group Teacher elicits whether or not the children are familiar with the game - Bingo. If not, through modelling and eliciting, the rules are explained. Every learner is then given a bingo board (all with di fferent combinations of the unchopped pizza toppings pictured) and some counters. The teacher calls out the names of the toppings and, if the learners have that picture, they cover the picture with a counter. When a learner has covered three lines in a row, horizontally, diagonally or vertically, s / he shouts Bingo! The learner then tells the rest of the class what is in that line. This is repeated several times. Control Group As above Find the Right Pizza Intervention Group The teacher puts the flashcards of the original, whole ingredients on the board. S / he then elicits what topping can be matched with which 182 <?page no="199"?> ingredient, using flashcards with the chopped up toppings on them. The chopped up topping is then put beside the original ingredient on the board. The flashcards with the whole ingredients are removed and the children are asked to name the toppings. Control Group As above 183 <?page no="200"?> 5th lesson - Consolidating Pizza Vocabulary Vocabulary cheese, corn, ham, mushroom, onion, pepper, pineapple, salami, tomato, tuna Materials flashcards depicting pizza ingredients (whole), flashcards pizza toppings, worksheet Pizza, pizza memory Sub-aims consolidating new vocabulary, developing speaking skills, learning to make statements about food and ask questions about it Warming up Activity - Reviewing last lesson’s vocabulary Intervention Group The teacher elicits the names of the various pizza toppings covered in the previous lesson. When the learners name a topping, the teacher puts the appropriate picture on the board. Control Group As above Vocabulary Expansion Intervention Group The teacher distributes the pizza worksheet (see illustration below). The teacher models, drills and elicits until the learners understand what they have to do with the sheet. On the worksheet are pictures of a plant and animals. There is a ‘sweet’ face and a salt cellar. The children tick whether a topping is sweet or salty and from a plant or from an animal. When the learners have completed the sheet, they compare answers with their partner. The teacher then discusses some of the examples and stresses that everyone’s sense of taste is di fferent. 184 <?page no="201"?> Control Group The teacher distributes the pizza worksheet (see the illustration below). The teacher models, drills and elicits until the learners understand what they have to do with the sheet. One the worksheet are pictures of various toppings that can be put on a pizza. There is a smiling face and a scowling face. The learners ask their partners what they do and do not like on their pizza and draw the appropriate topping under the appropriate face. After the learners have finished, the teacher discusses some of the answers with the class. Game - Pizza Memory Intervention Group The teacher elicits the rules of memory by modelling the game. Once the learners understand what they have to do they start to play the game. The learners are put into groups (of five, if possible) given two sets of cards. One set is square, the other, rectangular. The rectangular cards are larger. The square cards are placed, face down in rows and columns, on the desk. The larger cards are left in a pile. One learner takes a card from the pile and looks at the picture. The learner then tells the other players what is on the card, for example: L: On my pizza, I like pepper, salami and onion. The learner then takes one of the square cards. On this card there is a pizza. The learner tells the other players what is on the pizza; for example: L: On this pizza there is salami, onion and pineapple. The two pictures do not match so the learner says, “This is not my pizza.", and puts the large card on the bottom of the pile and the square card back in its original place. When a child has two cards that match, they keep both cards. The learner who collects the most cards wins the game. Control Group As above 185 <?page no="202"?> Pizza Rhyme Intervention Group The learners are asked to stand up. The teacher then chants the rhyme (see below), accompanying it with appropriate actions, mimes and gestures. Pizza is yummy, pizza is fine, I like pizza, I eat it all the time. I eat it for breakfast, I eat it for lunch, I love eating pizza, munch, munch, munch. On the pizza may be tomato, or cheese, or even green pepper. They all taste so good to me, it doesn’t even matter. Onion, ham, or mushroom, anything will do. I like pizza, how about you? The rhyme is repeated several times and the learners join in, repeating the actions and, if possible, the words of the chant. Control Group As above 186 <?page no="203"?> Lessons for 4th Grade 1st Lesson - Introducing New Vocabulary NB: Items in square brackets have been encountered by pupils prior to this study. Vocabulary a bar of [chocolate], a packet of [biscuits], a tin of baked beans, a tin of tuna, a carton of [eggs], a carton of [milk], a box of [cornflakes], a packet of fish fingers, a packet of [tea], a carton of [orange] juice, a jar of [honey], a jar of [jam] Materials shop pictures, basket containing shopping items, worksheet Shopping List Sub-aims introducing new vocabulary, practising these vocabulary items, developing basic speaking skills Warming up - Picture Description Intervention Group The teacher distributes copies of shop pictures (see illustration below). The learners look at the pictures and describe what they can see. This activates the necessary vocabulary for the lesson and encourages the learners to recycle vocabulary learned before the intervention lessons. Control Group As above Introducing new Shopping Vocabulary Intervention Group The LL form a circle. The T places a basket of shopping in the middle of the circle. The basket is covered to hide its contents. The teacher asks a learner to take something out of the basket without looking inside. When the learner has done so the teacher asks the learner what it is. The learner can say the name of the item in 187 <?page no="204"?> German or in English. The teacher repeats the name in English, and then asks the class to repeat the name, for example: T - L : What is it? L - T: It’s chocolate. T - LL: Yes, it’s a bar of chocolate. T- LL: What is it? LL - T: (chorally) It’s a bar of chocolate. The teacher then asks another learner to take something out of the basket. The above is repeated. The teacher names two items, points to them, and the class repeat. This continues until all items have been removed. Control Group As above Game ‘What is missing? ’ Intervention Group The teachers asks the class to close their eyes. S / he then names some of the shopping items and asks the class to repeat the names after her / him. During this choral repetition, the teacher quietly removes one or two items and hides them. The learners then open their eyes and are asked to name the missing items. When the learners answer, the teacher corrects the pronunciation, if necessary, and puts the correctly named items back. This is repeated several times until every item has been hidden at least once. The learners then return to their seats. Control Group As above Practising new Vocabulary Intervention Group The teacher writes Shopping List on the board and explains that you need a shopping list to go shopping. S / he asks the learners what they want to buy and pins the appropriate flashcards on the board. 188 <?page no="205"?> This is repeated until all flashcards are on the board. Control Group As above Information Gap Activity Intervention Group The teacher distributes a worksheet that shows a shopping list. S / he then instructs learners to work in pairs and compare their respective lists. The teacher models, with the learners, what they should do with the worksheets, until the task is understood by the learners. In the report back stage, the teacher asks learners to say what they want to buy. Control Group As above 189 <?page no="206"?> 2nd Lesson - Practising New Vocabulary Vocabulary a bar of [chocolate], a packet of [biscuits], a tin of baked beans, a tin of tuna, a carton of [eggs], a carton of [milk], a box of [cornflakes], a packet of fish fingers, a packet of [tea], a carton of [orange] juice, a jar of [honey], a jar of [jam] Materials flashcards depicting shopping dialogue, worksheet Shopping Dialogue Sub-aims practising new vocabulary, developing listening skills, developing communication skills Warming up - Categorizing Shopping Items Intervention Group The teacher pins flashcards showing di fferent types of containers on the board, asking learners to categorize the new vocabulary items according to the material the containers are made out of. T - LL: I’ll show you a thing you can buy and you come to the board and put it under the right container. For example: honey. Where does this go? L: Jar. T - L: That’s right. Please come to the board and put it under the jar. L comes to the front and puts the flashcard depicting the jar of honey under the appropriate container. T - L: Honey is in a jar. This is repeated until all vocabulary items have been categorized. Control Group The new vocabulary items are repeated without categorizing them. Using new vocabulary - Shopping Dialogue Intervention Group The learners are asked to form a circle in front of the board. The teacher 190 <?page no="207"?> explains that they are about to hear a story of Lucy who is going shopping. While hearing the story, the teachers uses suitable gestures, facial expressions and mimes to help the learners understand the story. Di fferent voices are used for Lucy and for the shop assistant. The following story is told to the learners who are encouraged to join in. T: This is the shop assistant. He works in the supermarket. He is there to help customers. And this is Lucy. She wants to buy some things. Lucy enters the shop and says hello. The shop assistant also says hello. Lucy tells the shop assistant what she wants to buy. She says: ‘I want to buy a carton of milk, a jar of honey, a bar of chocolate, a carton of eggs and a packet of tea, please.’ The shop assistant says: ‘Sorry, no tea.’ After that, the assistant adds up the prices of Lucy’s shopping ‘A carton of milk, a jar of honey, a bar of chocolate, a carton of eggs: That’s three pounds, please.’ Lucy pays and says goodbye. The shop assistant says goodbye. Control Group As above Practising and customizing the dialogue Intervention Group LL return to their seats and T distributes the dialogue worksheet and the cut & paste shopping items. Task: Learners have to paste the shopping items to make their own shopping list. This personalized shopping list is then used as the basis for practising the shopping dialog with their partner. In the report back stage, the teacher asks learners to perform their shopping dialog as a role play in front of the whole class. Control Group As above 191 <?page no="208"?> 3rd Lesson - Shopping Memory Vocabulary a bar of [chocolate], a packet of [biscuits], a tin of baked beans, a tin of tuna, a carton of [eggs], a carton of [milk], a box of [cornflakes], a packet of fish fingers, a packet of [tea], a carton of [orange] juice, a jar of [honey], a jar of [jam] Materials flashcards depicting shopping dialogue, worksheet Shopping Dialogue, Shopping Memory Sub-aims practising new vocabulary, developing listening skills, developing communication skills Warming up - Repeating the Shopping Dialogue Intervention Group The teachers asks those learners who did not perform their role play yet to do so now. Control Group As above Game - Shopping Memory Intervention Group The teacher elicits the rules of memory by modelling the game. Once the learners understand what they have to do they start to play the game. The learners are put into groups (of five, if possible) given a sets of square cards and a shopping trolley and a shopping list each. The square cards are placed, face down in rows and columns, on the desk. One learner takes a card from the pile and looks at the picture. The learner then tells the other players what is on the card, for example: L: This is a jar of honey. The learner compares with his or her shopping list. The learner tells whether the item is on the list; for example: L: Yes, I want to buy a jar of honey. 192 <?page no="209"?> If the two pictures do not match, the learner says, “No, I don’t want to buy a jar of honey.", and puts the card back in its original place. When a child has a match, they put the item in their shopping trolley. The first learner to fill his or her shopping trolley wins the game. Control Group As above 4th Lesson - Planning a Party Vocabulary a bar of [chocolate], a carton of [orange] juice, a packet of [balloons], a bottle of coke, a tub of [ice cream], a bag of crisps, a bag of chips, a bag of sweets, a bag of wine gums, a roll of streamers Materials flashcards depicting party items, Party Bingo, worksheet Party Shopping Sub-aims introducing new vocabulary, developing listening skills Introducing New Vocabulary Intervention Group The teacher tells the learners that they are going to go shopping again, but this time it is for a party they want to have. The teacher then asks the children to name things you need for a party. The teacher gives certain clues, for example, the category an item belongs, for example drink, decoration, food, and so forth. If the learners name an item that there is no flashcard for, the teacher o ffers a similar item instead L: Apple juice! T - L: Yes, juice, but I have only got orange juice. Is that okay, too? The teacher put the appropriate flashcards on the board and elicits the container the item comes in. T: Orange juice is in a ..... L: A carton! This is repeated until all vocabulary items have been named. 193 <?page no="210"?> Control Group As above, but the teacher does not provide category clues. What’s missing Intervention Group The teachers asks the class to close their eyes. S / he then names some of the party items and asks the class to repeat the names after her / him. During this choral repetition, the teacher quietly removes one or two items and hides them. The learners then open their eyes and are asked to name the missing items. When the learners answer, the teacher corrects the pronunciation, if necessary and puts the correctly named items back. This is repeated several times until every item has been hidden at least once. Control Group As above Listening Comprehension - Bingo Game Intervention Group Teacher elicits whether or not the children are familiar with the game - Bingo. If not, through modelling and eliciting, the rules are explained. Every learner is then given a bingo board (all with di fferent combinations of the party items pictured) and some counters. The teacher calls out the names of the items and, if the learners have that picture, they cover the picture with a counter. When a learner has covered three lines in a row, horizontally, diagonally or vertically, s / he shouts Bingo! The learner then tells the rest of the class what is in that line. This is repeated several times. Control Group As above Integrating new vocabulary - Shopping Dialogue Intervention Group The teacher reminds the learners about the shopping dialogue they did in an earlier lesson. S / he then distributes a worksheet and asks learners 194 <?page no="211"?> to work in pairs. Task: The learners ask their partner what they want to buy and how many of these they want to buy. The children then draw a line from the item to the trolley and write the number of items inside the trolley. In the report back stage, the teacher asks learners about what their partner wants to buy for a party. Control Group As above 195 <?page no="212"?> 5th Lesson - Planning a Party continued Vocabulary a bar of [chocolate], a carton of [orange] juice, a packet of [balloons], a bottle of coke, a tub of [ice cream], a bag of crisps, a bag of chips, a bag of sweets, a bag of wine gums, a roll of streamers Materials flashcards depicting party items, flashcards depicting flags, flashcards depicting containers, worksheet Party Shopping, Party Domino, large version of Party Domino Sub-aims introducing new vocabulary, developing listening skills Warming up - Repeating Party Vocabulary Intervention Group The teacher asks the learners to name the items they remember from the last lesson and pins the corresponding flashcards on the board. Then the teacher explains the di fferent uses of chips, fries and crisps in America and Britain. Control Group As above, but without the information about the di fferent meanings of chips and crisps. Speaking Skills - Party Shopping Intervention Group Then the teacher asks those student who did not give feedback on the party shopping worksheet during the last lesson to tell the others what their partner wants to buy for a party. Control Group As above 196 <?page no="213"?> Consolidating Vocabulary - Shopping Dominoes Intervention Group The teacher checks that the learners know how to play dominoes. The rules of this domino game are explained, by modeling, until the teacher is sure that everyone understands them. The teacher uses a large domino set to illustrate the rules. The learners receive five cards each. The rest of the cards are left in a pile. One card is taken from the pile and placed, face up, in the middle of the table. On this card there is, for example, a carton of orange juice and a bar of chocolate. One learner starts (perhaps the youngest). She looks at the card on the table and then at her cards. She might place a bottle of coke next to the orange juice, but has to justify why she has done so, for example: T - L1 : What’s the same? L1 - T: You can drink it. T: (looks at L2) L2: They taste sweet. The teacher accepts any combination if the learner can say why they have placed their card next to another card. The learners form five groups and play the game. Control Group As above, except the learners match a bottle of coke with a bottle of coke, and a bag of sweets with a bag of sweets, and so forth. 197 <?page no="215"?> Narr Francke Attempto Verlag GmbH + Co. KG Postfach 25 60 · D-72015 Tübingen · Fax (0 7071) 97 97-11 Internet: www.narr.de · E-Mail: info@narr.de This book presents the first detailed empirical study of irregular verb morphology in the German-English interlanguage. Starting with a theoretical in-depth account of irregular verb morphology both in English and German, three widely discussed theses about its nature are tested against a wide range of empirical data, both from L1 and L2. Although the findings partly confirm existing models and theories, the data also show the need for further re-examination of some fundamental questions through associative or probabilistic computer models. All in all this book provides a rigorous, profound and thought-provoking discussion of the phenomenon in question, involving an up-to-date revision of existing theories as well as the cross-linguistic examination of L1 and L2 data by an array of sophisticated multivariate statistical models. It is mainly directed towards university students of German and English linguistics, psycholinguistics and language acquisition at an advanced level. Thomas Wagner Interlanguage Morphology Irregular Verbs in the Mental Lexicon of German-English Interlanguage Speakers Language in Performance, Band 42 2010, 194 Seiten, €[D] 58,00/ SFr 98,00 ISBN 978-3-8233-6547-1 <?page no="216"?> Narr Francke Attempto Verlag GmbH + Co. KG Postfach 25 60 · D-72015 Tübingen · Fax (0 7071) 97 97-11 Internet: www.narr.de · E-Mail: info@narr.de Within the United States of America, French is of importance in only two areas, Louisiana and New England, the latter often being referred to as the Québec d’en bas for its high number of French-Canadian immigrants. Among the six states that constitute New England, Massachusetts is the one that attracted most of them, Québécois as well as Acadiens. Despite the high number of citizens of French-Canadian origin and the proximity to Canada, French has been losing ground as a langue du foyer in all of New England but especially in the southern part. This sociolinguistic study concentrates on the process of language decay among the French-Canadian population of Massachusetts. Based on a corpus consisting of 87 qualitative interviews and a quantitative questionnaire survey of 392 questionnaires in 7 areas (covering the centers of French- Canadian immigration throughout Massachusetts), this study approaches the topic in a new, broader angle by encompassing the following aspects: an analysis of U.S. Census data on ancestry and language use, an overview of the history of French- Canadian presence in Massachusetts, various specificities of the varieties of Canadian French spoken there, as well as an analysis of the extralinguistic factors, such as the heterogeneity of the French -speaking population, and the intralinguistic consequences, such as unskilled code-switching, of language decay. Edith Szlezák Franco-Americans in Massachusetts “No French no mo’ ‘round here” Language in Performance, Band 40 2010, 325 Seiten, zahlreiche Abb. und Tab., €[D] 64,00/ SFr 100,00 ISBN 978-3-8233-6449-8