Dictionary Look-up Strategies and the Bilingualised Learner's Dictionary: A Think-aloud Study [Reprint 2016 ed.] 9783110918809, 9783484391178

Citation preview

Series Maior

LEXICOGRAPHICA Series Maior Supplementary Volumes to the International Annual for Lexicography Suppléments à la Revue Internationale de Lexicographie Supplementbände zum Internationalen Jahrbuch für Lexikographie

Edited by Sture Allén, Pierre Corbin, Reinhard R. K. Hartmann, Franz Josef Hausmann, Ulrich Heid, Oskar Reichmann, Ladislav Zgusta 117

Published in cooperation with the Dictionary Society of North America (DSNA) and the European Association for Lexicography (EURALEX)

Jenny Thumb

Dictionary Look-up Strategies and the Bilingualised Learner's Dictionary A Think-aloud Study

Max Niemeyer Verlag Tübingen 2004

To my loving parents

Bibliografische Information der Deutschen Bibliothek Die Deutsche Bibliothek verzeichnet diese Publikation in der Deutschen Nationalbibliografie; detaillierte bibliografische Daten sind im Internet über http://dnb.ddb.de abrufbar. ISBN 3-484-39117-0

ISSN 0175-9264

© Max Niemeyer Verlag G m b H , Tübingen 2004 http://w\vw.niemeyer. de 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. Printed in Germany. Gedruckt auf alterungsbeständigem Papier. Druck und Einband: Digital PS Druck AG, Birkach

Acknowledgements

I am very much indebted to Dr. R. K. K. Hartmann, Dr. Phil Benson and Dr. Tse Shek Kam who helped, encouraged and made valuable suggestions. I hope they will find here a proof of my thanks and gratitude.

Table of Contents

List of Tables

XI

List of Appendices

XI

Notes on Transcription and Translation 1. Introduction 1.1. Use of dictionaries in Hong Kong schools - popularity of bilingualised learner's dictionaries 1.2. Aim of this study 2. Literature Review 2.1. Introduction 2.2. User perspective research 2.2.1. Reference needs 2.2.2. Reference skills 2.3. The bilingualised learner's dictionary 2.3.1. Overview 2.3.2. Definitions in bilingualised learner's dictionary 2.3.2.1. Defining policies and comprehensibility 2.3.2.2. Lexicographic definition 2.3.2.3. Defining with genus and differentia 2.3.2.4. Substitutable definition 2.3.2.5. Condensed definition 2.3.2.6. Morphosemantic definition 2.3.2.7. Taxonomic definition 2.3.2.8. Polysemous entries 2.3.2.9. Lexicographic definition deficiencies 2.3.3. Empirical studies on bilingualised learner's dictionary. 2.4. Present state of dictionary use research 2.4.1. What we know about skills and strategies 2.4.2. What we do not know about skills and strategies 2.5. Think-aloud and stimulated recall interview. 2.5.1. Think-aloud 2.5.1.1. Historical background 2.5.1.2. Theoretical framework 2.5.1.3. Think-aloud and instructions 2.5.1.4. Use of think-aloud in language studies 2.5.1.5. Think-aloud protocol analysis 2.5.1.6. Reliability and validity 2.5.2. Stimulated recall interview 2.5.2.1. Theoretical framework

XII 1 1 1 3 3 3 4 8 18 18 21 22 23 24 24 25 26 26 27 28 28 31 31 33 33 33 34 34 35 36 37 38 40 40

VIII 2.5.2.2. 2.5.2.3. 2.5.2.4. 2.5.2.5. 2.6. Conclusion

Stimulated recall interview and instructions Use of stimulated recall interview in research Stimulated recall interview analysis Reliability and validity.

41 41 42 43 43

3. Research Methodology 3.1. Introduction 3.2. The preliminary study 3.2.1. Subjects 3.2.2. Instruments 3.2.2.1. Reading texts 3.2.2.2. Bilingualised dictionaries 3.2.3. Procedures 3.2.3.1. Think-aloud 3.2.3.2. Observation 3.2.3.3. Follow-up questionnaire 3.2.3.4. Stimulated recall interview. 3.2.4. Data analysis 3.3. The main study. 3.3.1. Subjects 3.3.2. Instruments and procedures 3.3.3. Data analysis 3.4. Summary

45 45 45 45 46 46 49 49 50 51 51 51 51 53 53 54 54 55

4. Coding Scheme 4.1. Introduction 4.2. Coding scheme 4.2.1. Transcription 4.2.2. Segmentation and mental operations 4.2.3. Constructing a preliminary coding scheme 4.2.4. Finalizing coding scheme 4.2.4.1. Revising preliminary coding scheme 4.2.4.2. Interrater reliability 4.2.4.3. Intra-rater reliability 4.3. Discussion 4.4. Summary

56 56 56 57 57 58 59 60 61 62 65 65

5. Dictionary Look-up Strategies 5.1. Introduction 5.2. Process of identifying and categorizing strategies 5.3. Description and illustration of look-up strategies 5.3.1. Ignoring strategy. 5.3.2. Assuming strategy. 5.3.3. Minimizing strategy. 5.3.4. Checking strategy

67 67 67 69 69 70 72 75

IX 5.3.5. Paraphrasing strategy 5.3.6. Stretching strategy. 5.3.7. Maximizing strategy 5.4. Discussion 5.4.1. Validity of the method 5.4.2. Strategies 5.5. Summary 6. Using Bilingualised Entries 6.1. Introduction 6.2. Look-up strategies 6.3. Language preference 6.4. Language proficiency. 6.5 Target words 6.6 L2 definitions 6.7. Discussion 6.8. Summary

77 79 82 85 85 86 88 89 89 89 90 92 93 94 96 98

7. Strategy Users 7.1. Introduction 7.2. Overview 7.3. Case studies of two dictionary users 7.3.1. Learner 1: May, female, 19 7.3.2. Learner 2: Pang, male, 20 7.3.3. Comparing the two learners 7.4. Summary

99 99 99 101 101 103 104 106

8. Conclusion 8.1. Introduction 8.2. Contributions of this research 8.2.1. Effectiveness of think aloud 8.2.2. Description of strategies 8.2.3. Strengths of bilingualised learner's dictionaries 8.2.4. Complexity of strategy use 8.3. Limitations of the research 8.3.1. Limitations in the research methodology 8.3.2. Limitations in the scope of the research 8.4. Recommendations for further research 8.5. Summary

107 107 107 107 108 108 108 109 109 109 110 Ill

References

113

Appendices

121

List of Tables

Table 3.1: Number of types for "Cleaver the King of the Kitchen" (Text One) Table 3.2: Number of types for "Why Illegal Hawking Continues on Streets"(Text Two) Table 3.3: Number of types for "Traditional Medicine Going the Way of Its Endangered Ingredients" (Text Three) Table 3.4: Average number of words ENC and RE students underlined for Texts One, Two & Three Table 3.5: Comments on "Cleaver the King of the Kitchen" (Text One) Table 3.6: Comments on "Why Illegal Hawking Continues on Streets" (Text Two) Table 3.7: Comments on "Traditional Medicine Going the Way of Its Endangered Ingredients"(Text Three) Table 3.8: Criteria frequencies Table 3.9: Distribution of bilingualised learner's dictionaries Table 3.10:Distribution of data Table 5.1: Frequency of strategies among the look-ups Table 5.2: The use of ignoring strategy to look up crucial by Wing Table 5.3: The use of assuming strategy to look up garnishes by Yin Table 5.4: The use of minimizing strategy to look up scoring by Bori Table 5.5: The use of checking strategy to look up sumptuous by Samantha Table 5.6: The use of paraphrasing strategy to look up shredding by Hang Table 5.7: The use of stretching strategy to look up stroked by Zoe Table 5.8: The use of maximizing strategy to look up savoured by Wing Table 6.1: Strategies'use of bilingualised entries Table 6.2: Learners' language orientations Table 6.3: Learners' language preference patterns Table 6.4: Learners' language proficiency Table 6.5: Learners' patterns by target words Table 7.1: Look-up strategies used by 18 students

47 47 47 48 48 48 49 53 54 55 68 69 71 73 75 78 80 83 89 91 91 92 93 99

List of Appendices Appendix 1 : Appendix 2a: Appendix 2b: Appendix 3 : Appendix 4:

Questionnaire on Dictionary Use Text one Follow up Questionnaire Foliow up Questionnaire Target Words

119 120 122 124 125

Notes on Transcription and Translation

Transcription adapted from Van Someren and Sandberg (1994).

(

)

' (

Short pause Long pause Unusual silence Action protocols are inserted in brackets next to a cognitive operation code, e.g. (turning dictionary pages) Utterances originally in English are placed within single inverted commas, e.g. 'stroke down1 Romanized Chinese equivalents are placed within brackets next to a communication unit, e.g. (sing2)

' )

Cantonese sounds and tones are represented according to Lau's (1977) system of romanization. Six basic Cantonese tones are used: Tone Tone No. Example

High falling 1 Wan1

Middle rising 2 Wan2

Middle level 3 Wan3

Low falling 4 Wan4

Low rising 5 Wan5

Low level 6 Wan6

1. Introduction

1.1. Use of dictionaries in Hong Kong schools - popularity of bilingualised learner's dictionaries

Bilingualised learner's dictionaries are "the result of an adaptation of unilingual English learners' dictionaries which have all or part of their entries translated into the mother tongue of the learner" (Hartmann 1994a: 243). High sales figures show that two bilingualised dictionaries have been especially well received by the Chinese learners of English in Hong Kong: Oxford Advanced Learner's English-Chinese Dictionary (OALECD) and Longman English-Chinese Dictionary of Contemporary English (LECDCE). However, the commercial success of bilingualised learner's dictionaries has not acted as a catalyst for serious research on how Chinese learners actually use their bilingualised dictionaries (Hartmann 1994a: 244). In fact, published research on the use of bilingualised learner's dictionaries elsewhere in the world has been scarce. To the knowledge of the researcher, only five such studies have been published (Hartmann 1994b, Laufer & Melamed 1994, Laufer & Hadar 1997, Laufer & Kimmel 1997, and Fan 2000). There has also been little research on look-up behaviour in dictionaries generally, in part because it is a difficult field drawing on the adjacent fields of sociolinguistics and psycholinguistics (Bejoint 1994: 154). The complex lexicographical text, the variety of dictionaries, the individual parameters of users, and the variety of situations in which dictionaries are consulted also interact to keep the number of studies on dictionary look-up at a low level (ibid.: 154). One of the prevalent issues discussed in some of the published dictionary use studies is the question of whether failure to use dictionaries effectively is a result of poor reference skills or deficiencies in the dictionaries themselves (Cowie 1999: 188). Until quite recently the dictionary researchers and commentators tended to 'blame' dictionary-makers rather than dictionary users (Tickoo 1989). Their blame, however, was not based on any empirical evidence. Tomaszczyk (1979) was the first researcher to produce empirical evidence showing that elementary and intermediate learners of foreign languages used their dictionaries ineffectively because of their limited reference skills. Though the prevailing view now is that successful use of a dictionary calls for a special 'competence' (Herbst and Stein 1987: 115, Cowie 1999: 88), there is still very little published research, particularly in look-up strategies, to support this. This investigative study aims to add to our knowledge in this area in the context of bilingualised dictionary use.

1.2. Aim of this study

This study aims to add to our knowledge in the area of look-up behaviour by focusing on look-up strategies using the think-aloud protocol as the research method. The main aim of

2 the study is to examine qualitatively how Hong Kong tertiary students look up word meanings in bilingualised English-Chinese learner's dictionaries while reading. Specifically, the researcher aims to identify and describe the look-up strategies used by these students. The study focuses on recovery of word meanings because most dictionary use studies have reported that this is the most frequent user activity (e.g. Bejoint 1981: 215, Snell-Homby 1987: 167, Summers 1988: 114, Cowie 1999: 181, Diab and Hamdan 1999: 298). The students were given a reading task because previous research indicates that one of the most common language contexts in which dictionaries are used is reading (e.g. Bejoint 1981: 216, Cowie 1999:185). The researcher's unpublished dictionary use study also suggested that reading is the most frequent linguistic activity and that looking up word meaning is the most common purpose among university students in Hong Kong1. Since think-aloud protocol was chosen as the main methodology for data collection in the present study, another important aim is to investigate if this is an effective instrument for studying look-up behaviour. The four research questions addressed in this study are: 1) How can we analyze and describe look-up strategies in the bilingualised learner's dictionary? 2) What look-up strategies do students use when they are reading and need to find the meaning of a target word in a bilingualised English-Chinese learner's dictionary? 3) To what extent do different look-up strategies make use of the bilingualised entries in the English-Chinese learner's dictionary? 4) To what extent are look-up strategy patterns specific to individual students or are there common patterns among die students? Investigation of the first question aims to devise a methodological framework for analyzing and describing look-up strategies. Investigation of the second question aims to identify and describe types of look-up strategies by using the newly formed coding scheme. The answer to the third question may deepen our understanding of the students' use of the bilingualised entries in the dictionary. The answer to the fourth question may enhance our understanding of the strategy use of the learners.

1

In 1994, the researcher conducted a questionnaire survey on dictionary use at the Hong Kong Polytechnic University. Nearly all of the 137 first-year students enrolled on the 'English for Academic Purposes' course indicated they used their bilingualised English-Chinese learner's dictionaries for finding word meanings while reading. Over half of them said they used their dictionaries several times per week while reading.

2. Literature Review

2.1. Introduction

In this chapter, the literature in three areas that are particularly relevant to this research is reviewed:- user perspective dictionary research, the bilingualised learner's dictionary, and think-aloud method. Section 2.2 gives a critical account of two aspects of user perspective dictionary research - reference needs and reference skills - with particular focus on look-up strategies. Section 2.3 focuses on the bilingualised learner's dictionary. An overview of its design features and the principles of composing definitions is given in Section 2.3.1 and 2.3.2. Section 2.3.3 presents the current state of research about the use of the bilingualised learner's dictionary, with special emphasis on look-up strategies. Sections 2.4 summarizes what we know or what we do not know about the present state of research into reference skills and strategies. Finally, Section 2.5 reviews the strengths and weaknesses of thinkaloud and stimulated recall interview as methods for research into look-up behaviour.

2.2. User perspective research

Hartmann and James (1998) defined the 'user perspective' as: "an approach that considers lexicography from the point of view of the user. Issues to be considered include the status of users, their previous background and experience, their familiarity with available reference works and look-up strategies required to consult them successfully and whether deliberate instruction can develop and improve reference skills" (p. 25). Since the 1960 conference on lexicography at Indiana University, there has been a prevalent feeling that dictionaries are not serving their users well. "The combined forces of lexicography, commerce and linguistics have not produced a satisfactory account of the reference needs of the dictionary user" (Hartmann 1987: 20). Tomaszczyk (1979:103) argues that though commercial dictionaries are very popular, very little research is on the needs of dictionary users. The only study known to him concerns the use of dictionaries of English by native speakers of British English. Stein (1984, cited in Winkler 2001: 1) concurs and notes that since it is obvious that dictionaries are written for their users, "we therefore need much more research on the dictionary user, his needs, his expectations, and his prejudices". Householder and Saporta (1962) even recommend that "dictionaries should be designed with a special set of users in mind and for their specific needs" (p. 279). However, research on the 'user perspective' is still relatively young in lexicography.

4 2.2.1. Reference needs The last two decades or so have witnessed a growing awareness of the user perspective and the number of published studies on reference needs has increased at such a fast pace that Bejoint (1994) remarks that "it is difficult to keep track of all of them" (p. 141). Many of the studies on dictionary use elicit the habits and preferences of native English speakers, students learning a foreign language, secondary teachers, tertiary teachers or school children. Most of the studies are carried out with English dictionaries, native speaker's or learner's dictionaries, and monolingual or bilingual dictionaries. Most of the research is based on questionnaires and tests. In this study, 'reference needs' are defined as "the circumstances that drive individuals to seek information in reference works such as dictionaries" (Hartmann and James 1998: 116). The study of reference needs includes the types of dictionaries preferred by users e.g. monolingual, bilingual; the types of activities e.g. reading, writing, listening, speaking or translating; and the types of information sought e.g. meaning, spelling, pronunciation, grammar (Hartmann 1987: 20). Barnhart (1962), who emphasized that the role of the dictionary was to answer the questions which the dictionary users asked, conducted a pioneering questionnaire survey in 1955 in 27 American states. The survey aimed at finding out the relative importance of different types of information in popular American college dictionaries. 108 English teachers were asked to rate six types of information according to their importance to college freshmen. It was found that the most frequently looked-up information was meaning. Though the numerical results were not reported and the questionnaire was not reproduced, Barnhart's contribution to the field of dictionary research was valuable. The study was the earliest attempt to supply quantitative data on the reference needs of dictionary users. Quirk (1973) studied the use of monolingual general-purpose English dictionaries by 220 English-speaking British male and female undergraduates majoring in the humanities or the sciences at the University of London. A 30-item questionnaire was used to elicit information on dictionary use. The results showed that the students predominantly used their dictionaries to look up word meaning. Only a minority of them had the habit of retrieving information on pronunciation and form class. However, a majority of students indicated strong dissatisfaction over the comprehensibility and adequacy of definitions. Apparently, they had difficulty in understanding the metalanguage of the definitions which encapsulated word meanings. Quirk found that this difficulty was associated with unfamiliarity which resulted from infrequent dictionary use. Quirk's work was also pioneering in that it sought to investigate the reference needs by directly questioning the users. It gave us a more realistic picture of the dictionary user and helped lay the foundations for later empirical research on the user perspective. Quirk's study was replicated by Greenbaum et al. (1984) at the University of Wisconsin in the United States of America. A population of 240 humanities or sciences undergraduates was surveyed with the help of a 32-item questionnaire. The findings were similar to Quirk's in two main respects: the dictionary was seen as an instrument for finding word meanings and the students were dissatisfied with the definitions because of their use of difficult metalanguage. The above studies were about the reference needs of native speakers. The first study of any scope and complexity into foreign users of dictionaries was conducted by Tomaszczyk (1979) who studied how university foreign languages students (284 persons), teachers and

5 translators in Poland and America (165 persons) used dictionaries. These two groups of subjects reported learning or speaking 16 different languages. They completed a 57-item questionnaire aimed at investigating the use of monolingual and bilingual dictionaries in the study of a foreign language. The results showed that the extent of dictionary use depended on three factors: the nature of look-up skills, the subjects' level of language proficiency, and the extent to which a given skill was practised. There was more extensive use of dictionaries for writing than for reading, for L1-L2 translation than for L2-L1 translation, and for speaking than for listening. Subjects in both groups were more satisfied with their monolingual than with their bilingual dictionaries and those with more advanced levels of competence in the foreign language appeared to get the most out of the monolingual dictionaries. Similar to the findings of Quirk and Greenbaum et al., the subjects used their dictionaries for meanings most frequently and often found the definitions of monolingual dictionaries difficult to comprehend. It also appeared that the advanced learners and speakers were getting more out of their dictionaries than the beginning and intermediate learners. Another major finding was that a vast majority of the foreign language learners and speakers used dictionaries, but their dependence on dictionaries decreased as their language proficiency increased. Although the use of monolingual dictionaries became more extensive and frequent as the subjects' language proficiency increased, nearly all of them continued to use the bilingual ones. Though Tomaszczyk conducted one of the most comprehensive questionnaire surveys on reference needs at the time, the reliability and validity of the results are undermined by unclear presentation of numerical evidence and incomplete statistical analysis. The questionnaire was also not reproduced for close scrutiny. Prompted by Tomaszczk's research, Béjoint (1981) launched his own questionnaire survey to study the reference needs of foreign students. He studied the use of monolingual English dictionaries by 122 French students of English at the University of Lyon with the aid of a 21-item questionnaire, which was partly based on Tomaszczyk's. The results show a striking similarity with those of Tomaszczyk: the vast majority of foreign language learners and speakers used dictionaries to look up word meanings (87%), placing this well ahead of syntax (53%) and synonyms (52%). The dictionary was moreover, used for decoding more often than for encoding. The subjects indicated that bilingual dictionaries were in general more satisfactory and useful than monolingual dictionaries. Unlike Tomaszczyk, who found that the advanced learners and speakers were getting the most out of their monolinguals, Béjoint concluded that many students were not using their monolinguals as fully as they should be. Many of his students were not even aware of the riches of information contained in their dictionaries. Like Béjoint and Tomaszczyk, Baxter (1980) conducted a user-oriented study on nonnative speakers of English. He surveyed a population of 342 Japanese university students who were either English majors or non-English majors with the aid of a 6-item questionnaire. The conclusions were similar to that of previous research: preference of bilingual to monolingual dictionaries and the difficulty of definitions. Non-English majors rarely used a monolingual English dictionary while English majors did so more frequently. Both groups of students said they used a bilingual English-Japanese dictionary more often because it was easier to use. English majors also indicated their preference for a learner's dictionary especially one in which a controlled defining vocabulary has been employed. On the basis of the questionnaire data, Baxter strongly argued for the use of a monolingual learner's dictionary which had controlled defining vocabulary. He thought long-term use of such dictionary

6 would allow students to acquire a lexical element in the context of other words of the same language (definitions and examples). On the contrary, long-term use of the bilingual dictionary developed the tendency of learning single lexical items with their equivalents in other language. Baxter concluded by stating that the monolingual dictionary was more useful than the bilingual dictionary in the teaching of English as a foreign language. This conclusion, however, was not directly derived from the questionnaire data. Though Baxter's survey was quite limited in scope and methodology, it brought into sharper focus the relationship between acquisition of dictionary and frequency of use. If university students of the research owned bilingual dictionaries, it was expected that they would use them more often. The converse was also expected to be true. The results also suggest that there is correlation between level of language proficiency and frequency of use of a monolingual dictionary. It is possible that English majors use monolingual dictionaries more frequently because they have higher level of English proficiency. Hartmann's (1983) study aimed at investigating the use of dictionaries among learners of German in South West England. The subjects were a group of 16 teachers and 118 students at more than 200 educational institutions including secondary, further and higher education. A 23-item questionnaire with questions such as the types of dictionaries used, the contexts of use and the purposes of use was used. The main findings were that the most popular activity was translating, followed by reading and writing; that the use of bilingual dictionaries was predominant within and outside formal language classes; that semantic and syntactic information were retrieved more often than phonetic and etymological information; and that users received little dictionary instruction. Hartmann's findings partly confirmed those of Tomaszczyk (1979) and Béjoint (1981) but contradicted that of Baxter (1980). The subjects' preference for the bilingual dictionary suggests that its role in language learning should not be underestimated. Any ban or discrediting of its use in favour of the monolingual dictionary would therefore appear to be unrealistic. Though the findings were only valid within the confines of the research area and issues, they indicate that students may find a 'translating' or a "bilingualised1 dictionary beneficial and helpful. Hatherall's (1984) research results appear to support this assertion. His subjects were given one hour to translate a difficult English text into German. During the translation hour, the subjects were each given a recording form on which they entered the title and the edition of the dictionary used; recorded the type of information retrieved in each dictionary consultation; indicated if they used the retrieved information in their translation; and indicated whether the retrieved information of a word was useful or otherwise. After the task, the subjects were asked to complete a short questionnaire on their translation strategies, attitudes and expectations. Results indicated that groups of students tended to excessively translate word-for-word and bilingual dictionary users did not usually look up commonly occurring items such as prepositions. The finding that more advanced students were found to use the dictionary more often than less advanced ones was unexpected because it was generally assumed that the need of the latter to use the dictionary was greater (Tomaszcyk 1979). Less advanced students may have been more reluctant to try to consult their dictionaries because they were less confident of retrieving the necessary information. Taylor (1988) conducted a small-scale dictionary use survey on 122 students at the City Polytechnic of Hong Kong. It was assumed that the students needed an English dictionary, that they knew how to use it, and that they used a dictionary. It was also assumed that they used a monolingual dictionary since two monolinguals, the Oxford Advanced Learners' Die-

7

tionary of Current English (OALDCE) and the Longman Dictionary of Contemporary English (LDOCE) had been recommended. His findings did not match all his assumptions because only 30% of the students said they usually used a monolingual English learner's dictionary. 55% of them said they usually used the bilingual (bilingualised, according to this study) Oxford Advanced Learner's English-Chinese dictionary and their dictionary choice was influenced mainly by their school. However, the assumption that the students did use their dictionaries was supported by the findings. It was also discovered that the most frequent use of dictionaries was finding out word meanings and the least frequent use was looking up grammatical information. The major problems of dictionary use were understanding pronunciation symbols, identifying the right meaning of words, and the amount of time used to consult a dictionary. Taylor's findings, however, cannot be generalized because it was only a small-scale survey involving only a small number of tertiary students at one tertiary institution in Hong Kong. Li's (1997) survey involved 691 students and 110 teachers at Wuxi University of Light Industry in China. Interviews, translation tests and protocol techniques were also used, in addition to questionnaires, to elicit data. The investigation was "based on the hypothesis that different users have different needs for dictionaries, and that the use is influenced by the level of language proficiency and specialty" (ibid.: 63). The survey results showed that the subjects had a high level of dictionary awareness and positive attitudes towards dictionary use. Every student and teacher owned at least one English-Chinese dictionary, including bilingualised dictionaries, but the ownership of other types of dictionaries, especially monolingual ones, was low. It was apparent that the types of dictionary used were closely related to the users' education level and professional needs. English teachers who had higher education levels used monolingual (English-English) dictionaries much more often than science undergraduates did. Dictionary use also appeared to be much affected by the subjects' specialty. The humanities students majoring in international business used dictionaries most frequently and the science teachers consulted dictionaries least frequently. Looking up the meaning of words was the main reason for dictionary use but more than half of the users (64.1%) also consulted dictionaries for pronunciation purposes. These main findings of Li's questionnaire study were similar to some previous studies except one. Her discovery that a significant number of users sought pronunciation information does not support other studies (e.g. Tomaszczyk 1979, Bejoint 1981) which had tended to report that pronunciation was infrequently sought by their subjects. A more recent undertaking has been that of Diab and Hamdan (1999) who reported a case study that investigated how 50 Jordanian Arab university students of English used dictionaries while reading an introductory text in linguistics entitled "Phonology: The Sound Patterns of Language' - a 25-page chapter with approximately 9000 words in total. The subjects were given three weeks to complete the reading task with an understanding that a 75minute session would be devoted to explaining and discussing the content of the chapter at the end of that period. Diab and Hamdan collected dictionary use data by 'dictionary use record sheets' which were specially designed by the researchers to elicit the following data: the problematic word or phrase which motivated dictionary use, types of information sought, pre-dictionary use strategies; types of dictionary consulted and the students' satisfaction with dictionaries. A structured interview was conducted for each student immediately after the student's submission of the completed dictionary use records. Besides explaining and discussing the content of the chapter, the students answered questions covering the follow-

8 ing: pre-dictionary use strategies, the value of dictionary consultation, the importance of vocabulary mastery in understanding and answering questions in linguistics tests, the value of specialized bilingual dictionaries, and awareness of the existence of glossaries in course textbooks. The results revealed that the subjects looked up meaning in 85% of the look-ups. Information on pronunciation was also extracted in 15% of the look-ups. These findings were expected since it was essential to know the meaning and/or pronunciation of L2 words in order to prepare for the discussion. Other types of information such as morphology and grammar were not used. Perhaps these were not perceived as relevant since they were not usually needed for discussion.

2.2.2. Reference skills To meet their reference needs in the context of various tasks, such as reading, dictionary users need to apply reference skills. 'Reference skills' are defined as "the abilities required on the part of the dictionary user to find the information being sought" (Hartmann and James 1998: 116). A look-up strategy is defined as a systematic application of certain skills to retrieve the meaning of a word during the process of dictionary consultation (Scholfield 1982: 185). More precisely, a look-up strategy includes "those mental operations and decisions made by users in the process of consultation, from the selection of the relevant reference work, through the appropriate search and retrieval acts, to integration of the information obtained with the original reference needs" (Hartmann & James 1998: 152). Several lexicograhers and metalexicographers have pinpointed the complexity of the psycholinguistic processes involved in dictionary consultation while performing receptive tasks. Miller and Gildea (1985) briefly describe a systematic application of a sequence of reference skills (i.e. a strategy) used in a situation when the user is reading and needs to find out the meaning of a word in a monolingual dictionary. Miller and Gildea (ibid.: 13) describe the processes as involving a high-level of complex skills: search through the dictionary alphabetically for the target word; choose among two or three different entries on the basis of part of speech; decide among several alternative senses within the chosen entry; compare the context of original passage with a succession of dictionary contexts until a best guess can be made about the intended sense. Context-matching is referred to as a high-level cognitive skill because the user must keep in mind the source context while searching for a target word in a dictionary and must compare it with the definitions in the dictionary context until a fitting sense is decided. Scholfield (1982: 186-93 / 1999: 13-14) analyzed the procedure of consulting monolingual dictionaries in a hypothetical situation into a sequence of seven steps involving several skills and strategies: 1) Locate the word(s) or phrase which the learner does not understand. 2) Recover the canonical form of inflected unknown word. 3) Search for the unknown word in the alphabetic list. 4) Try the following procedures if at least one main entry for the unknown cannot be found: - If the unknown seems to be set phrase, idiom, or compound word, look up each main element. - If the unknown seems to have a suffix, look up the entry for the stem.

9 - If the unknown appears to be an irregularly inflected form or a spelling variant, scan nearby entries. - If there is an addendum, search there. 5) Reduce multiple senses or homographic entries by elimination but scan all of the definitions in the entry before making any decision. 6) Understand the definition and integrate it into the context where the unknown was met. The sub-steps: a) looking up unknown words in the definition itself, b) adjusting for complementation and collocation, c) adjusting for part of speech, and d) adjusting for breadth of meaning. If any of the above sub-steps is not used, the central definitional in formation about the unknown has still to be combined with the source context to see if it fits. 7) tnfeT one appropriate sense that fits the context from the senses entered, if none of these senses seems to fit. If more than one sense fits, seek further contextual clues in the source text to disambiguate. The above analysis hypothesizes that dictionary consultation process is unitary. Steps 1 to 4 deal with the macrostructure of the dictionary and therefore are straightforward. Steps 5 to 7 concern the microstructure of the dictionary and users are required to consider context when making various decisions. Users may not have to go through all seven steps in order to look up the meaning of a word. They are, however, required to apply a considerable amount of prior knowledge to the information at hand. They may also have to set up alternative hypotheses at a number of stages and to make inferences based on the nature of the target word and its context. Scholfield's steps also indicate that a successful dictionary user is expected to have quite a high level of language proficiency. Schelbert (1988: 63) argues convincingly that dictionary users should be equipped with quite a high level of linguistic skills such as: knowing when a sign acts as a signal or a symbol; reading and handling meanings dynamically; juggling with variations and prototypes; knowing how meanings adjust and bow to context; understanding modalities; using the whole/part relationship syntactically and rhetorically; sensing what is the same, similar and dissimilar, and paraphrasing. Not only do the users need to be reasonably linguistically competent, they should also be able to apply the skills successfully to have a successful look-up. If the learner is unable to apply these linguistic skills, the result may be an unsuccessful look-up and s/he may start the consultation process again, or use another dictionary, or ask a friend or teacher or give up. Hartmann (1989: 105) constructed a look-up strategies model based loosely on Scholfield's dictionary consultation procedure:

10 The model specifies seven main constituent strategies of dictionary reference, four related to the macrostructure of a dictionary and three related to its microstructure. Unlike Scholfield, Hartmann arranged the strategies in a recursive fashion. Hartmann (1991: 9) demonstrates the application of this model by using a sentence 'You look fagged to death', said Kate\ 1) select a monolingual learner's dictionary (ALD, LDOCE or COBUILD) to consult 2) choose fagged from the sentence as the target word 3) determine that the canonical or dictionary form offagged is fag 4) search for appropriate headword fag or fagged 5) decide where in the entry fag or fagged is to be found (ALD: 2nd paragraph of the entry fag either subentry 1 or 3; LDOCE: fagged as separate entry: COBUILD: fagged as sepa rate entry) 6) extract the information relevant for the sentence 7) relate retrieved information to the original context i.e. the sentence According to the hypothetical model, after skilful application of these seven strategies, the appropriate meaning of the target word fag or fagged is either retrieved (success) or not retrieved (failure). If meaning retrieval is unsuccessful, the user may have to start from strategy 1 again or give up. The application of the model clearly demonstrates that dictionary consultation involves complex psycho-linguistic processes and that the model could serve as a theoretical framework for empirical studies on look-up skills and strategies. These lexicographers and metalexicographers have shown the difficulties foreign learners had with reference skills. More importantly, they have also indicated that ordinary people also had the same difficulties when using dictionaries. However, these observations have yet to be supported by strong empirical evidence. The study of reference skills is "the most recent and least explored aspect" of dictionary use (Hartmann 1987: 23) and thus it is less advanced than the study of reference needs. Bejoint (1994) argues that "the complexity of the lexicographical text, the variety of dictionaries, the innumerable individual parameters of the users, and the variety of situations in which dictionaries are consulted" (p. 154) are the contributory factors. Bensoussan et al. (1984) studied how university students used monolingual and bilingual dictionaries when taking a reading comprehension test of 20 multiple choice questions. They conducted three separate studies at two different universities in Israel. The subjects of the first and second empirical study were 761 EFL students of similar English proficiency, attending reading comprehension courses at Haifa and Ben Gurion Universities. A third study was carried out by the researchers at Haifa University because they were skeptical of the results of the first and second studies. 740 students in the required EFL reading comprehension course, received eight different tests with each of them taking only one of the eight. The findings of all three studies showed that there was a general preference for bilingual dictionaries and, unexpectedly, that there was no significant correlation between the use of a dictionary and test results. Students who did not use a dictionary performed just as well as those who used a dictionary. It seemed that neither test scores nor time limitations were significantly influenced by dictionary use. The third study revealed that advanced English learners were more likely to use dictionaries than elementary learners. The latter also preferred bilingual dictionaries while those more proficient frequently consulted monolingual

11 works. Concerning the lack of improvement with reading test scores, Bensoussan et al. suggested: "the status quo stands: less proficient students lack the language skills to benefit from the dictionary, whereas more proficient students know enough to do without it. The dictionary can be used in a test situation only to fill in places where the context is already clear, not to create context"(p. 271). In an attempt to clarify the test results and to understand underlying attitudes and expectations of dictionary users, a follow-up questionnaire on dictionary use and preferences was administered to students of English, their teachers and a group of 13 third-year psychology students (with near-native proficiency). The survey results indicated that the most linguistically proficient learners (third year students) were most critical about dictionaries and had fewer expectations. They used dictionaries less but more selectively than less proficient learners. Almost half of them did not expect the dictionary to affect their test scores. Many less proficient learners (first-year students), on the other hand, were more frustrated and confused with monolingual dictionaries and often consulted bilingual works. They thought the definitions contained too many difficult, unknown words which in turn had to be looked up as well. Some said that even after referring to the monolingual dictionary, they often had to refer to the bilingual dictionary if they really wanted to understand the word. There was also frustration with the great number of meanings given in a polysemous entry, and the inability to find the exact meaning required by the context. Their use of bilingual dictionaries, on the other hand, did not appear to be any more satisfactory. They indicated that bilingual dictionaries often failed to give complete definitions or to include enough idioms. Moreover, in many instances, they still had to choose from among a wide range of meanings, to understand the context. Using the information from the questionnaire, Bensoussan et al. (1984: 270) inferred that the skills and strategies involved in consulting dictionaries were: establishing the lexical item which posed a comprehension problem; mastering the alphabetical ordering of headwords; and reading the first definition in an entry. Herbst and Stein's (1987) study aimed at determining the skills of German learners of English language when using dictionaries. Unlike Bensoussan et al. (1984) who used a test as the research instrument, they gathered data by means of a questionnaire sent to 160 first year English language university students and 60 English language teachers, as well as on conversations held with groups of learners and teachers. Questionnaire responses revealed that monolingual dictionaries were introduced in only two or three lessons, and little guidance was provided afterwards. Students were unable to use the dictionary to their full advantage. In addition, many teachers reported that problems with dictionary use seemed to be caused more by students' lack of interest and motivation than to deficiencies within dictionaries. The researchers speculated that dictionary training had been discouraged within second and foreign language classrooms because of the adoption of the 'communicative' teaching approach. They recommended that dictionary skills be taught and practised in language classes because users needed to be shown how to apply them. In one of the few published 'think-aloud' studies concerning dictionary use, Ard (1982) examined the use of bilingual dictionaries by two high-intermediate English as a second language students as a part of their composing process while writing. One student (Japanese female) who frequently used a bilingual dictionary and one (Arabic-speaking male) who

12

never did were asked to write a short composition and simultaneously orally describe what they were doing. The video-recorded protocols of the female student indicated that bilingual dictionary use often led to mistakes in lexical choice (paronym, word-combinations etc.) in her composition. Since the Arab student who did not use a dictionary also made similar kinds of lexical errors, Ard concluded that these types of lexical errors could arise even without the use of a bilingual dictionary but that certain errors might be motivated by the use of a bilingual dictionary, depending on the differences between the user's first and second language. However, he refrained from concluding that the use of a bilingual dictionary should be discouraged in order to minimize writing errors. Though the results indicated that the use of a bilingual dictionary could cause certain lexical errors, interlingual lexical interference may have caused some other errors. These results should not be generalized because of the limited scope of the research. The technique used, however, was valuable as it provided the researcher with a comparatively reliable way of observing dictionary use. Studies like Ard's also motivated Neubach and Cohen (1988: 4) to conduct research on six students selected from high, intermediate, and low-level EFL classes at the PreAcademic Center of the Hebrew University of Jerusalem using think-aloud and interview. The students were asked to perform two tasks, which required the use of a dictionary. They could either use a dictionary of their preference (monolingual or bilingual), or choose not to use a dictionary. The lower-level students used the Longman Active Study Dictionary of English and advanced students used Collins English Learner's Dictionary. All students used the same English-Hebrew dictionary: Megido Modern Dictionary. The research data revealed strategies and problems associated with the use of monolingual and bilingual dictionaries to look up polysemous words. Before the search, the students could use the strategy of'formulating an expectation of the possible meaning of a word'. This strategy created problems when an incorrect expectation was formulated. During or at the end of the search, the students could use the strategy of 'reading only the first definition in the monolingual dictionary' when dealing with a polysemous entry. Another important finding was concerned with the relationship between student's language proficiency and the use of pre-consultation strategies. It was found that the students with high proficiency tended to start the look-up with correct expectations of the word meaning both at the sentence and word level and they were able to draw suitable conclusions for the dictionary entries. The two intermediateproficiency students did not necessarily determine which part of speech a word was in and had incorrect expectations of word meaning before starting the look-up. During the look-up, they had problems with other words in the definition. The low-proficiency students used ineffective skills and encountered problems such as problems with the format and the dictionary entry itself. They felt frustrated during the search because they often could not understand the definition and they had to spend more time looking up words. A third important finding concerned the effect of certain word searches on skills. Words with more "easy-tounderstand" contextual clues seemed to produce the largest number of correct hypotheses prior to a word search and eventually led to a successful search. In contrast, words with fewer contextual clues produced more incorrect expectations and eventually led to faulty search. Neubach and Cohen's data also revealed the dictionary preference of students at different proficiency levels. Both high proficiency students preferred the monolingual dictionary because it gave them more precise word meanings. The intermediate and low students, however, preferred the bilingual dictionary partly because they disliked having to read a lot of

13 text in a monolingual dictionary in order to understand a word. The intermediate learners also used the monolingual dictionary when they did not understand or were not satisfied with the information in the bilingual dictionary, unlike the low proficiency students who used the bilingual dictionary only. The researchers suggested that the users needed to have proficiency in both the foreign language (English) and the native language (Hebrew) in order to find the best native-language equivalent. The last major finding dealt with the problems inherent in the dictionaries themselves. The bilingual dictionary tended to have problems like including equivalents in a strange and uncommon register and excluding the essential meaning of some common words. The monolingual dictionaries also presented problems such as using difficult metalanguage in some definitions and incorporating difficult or unknown examples. Neubach and Cohen's findings supported Schelbert's (1988) argument that a successful dictionary user needed to have a reasonable level of linguistic proficiency. In the study, when high-proficiency students did refer to the dictionary, they were generally successful with it. The low-proficiency learners, on the other hand, did not have sufficient language proficiency to make efficient use of the dictionary. The results reinforced Bensoussan et al.'s (1984) conclusion that advanced students did not need the dictionary so much, while the reverse was true for the weak ones. The findings also pointed to the need of providing dictionary skills training for learners. Problems such as looking in the wrong place, giving up the search before reaching the appropriate meaning, continuing the search without realizing that they had reached the correct meaning, and looking for the inflected form in the dictionary could be avoided if ample dictionary skills training was given. Unfortunately, Neubach and Cohen did not report the look-up strategies used by the subjects in a systematic manner and failed to show conclusively that the errors were caused by inadequate skills and strategies rather than deficiencies in the dictionaries used. Tono (1984), however, was able to produce evidence to show that performance errors were due to inadequate skills rather than deficiencies in the dictionaries used. He conducted an experiment to collect research data. The experiment involved the control of two key variables: dictionary entries and look-up process. By using invented words, he ensured that he could make sure the subjects did not have prior knowledge of the entries involved. He also chose a look-up process of accessing meanings that he was certain the subjects had experience in. Four hundred and two Japanese university students were divided into groups and given the task of translating a number of English texts into Japanese. The texts contained a number of nonsense words for which he had compiled special bilingual minidictionaries. Not all the groups received the same dictionary entries. Some, for example, had to use syntactic information in order to retrieve the appropriate information. Tono was able to demonstrate that of the different Japanese equivalents available for each nonsense word, the students had the tendency to choose the first sense as the word meaning. They would only continue to the next sense if the first sense was patently unsuitable. They were also not guided by syntactic and other clues which were only used when the equivalent could not be fitted in the context. Tono attributed poor standards of information retrieval to inadequate reference skills and strategies. Tono (1991) later conducted another study to find out the characteristics of the good dictionary user by using a user profile questionnaire, an English proficiency test, a task on the understanding of dictionary conventions, and a translation task. Tono suggested that good dictionary users could: a) retrieve necessary information from the entry faster; b) understand

14 the dictionary conventions better; c) have sufficient language proficiency to understand the content of the entry; and d) have the ability to choose the right word to look up. In other words, the research showed that those who are good language learners could also be good dictionary users. This finding lends some support to Schelbert's (1988) argument that dictionary users need to have quite a high level of linguistic skills. Using the direct observational method, Bareggi (1989, cited in Cowie 1999:189) also found that first-year university students were unable to use contextual clues to locate the appropriate sense in polysemous entries of their monolingual dictionaries. Students had the tendency to adopt the strategy of retrieving the first sense. Nuccorini (1994) examined students' errors in written translation tests (English to Italian) against the microstructure of the dictionaries used. She observed that though the majority of her sample used dictionaries which had user-unfriendly organization of microstructure, most errors were caused by students' low levels of look-up skills. When dealing with polysemous entries, her students adopted the strategy of not reading beyond the first listed sense. Blachowicz et al. (1990) likewise found their fourth grade students used this strategy of choosing the first-listed sense. The students were instructed to use the dictionary to locate an entry for a targeted word presented in a sentence. The data was collected by means of observation, interview and test. The results indicated that the students adopted different skills for looking-up the target words. Before starting the look-up, some students appeared to adopt the strategy of formulating a possible word meaning (see also Neubach and Cohen 1988: 6-8). When locating the target headword, the students either adopted the skill of plodding through each entry letter page-by-page or the skill of using the guidewords. Upon locating the target entry, they either referred to or ignored the pronunciation key. They were then observed to adopt one of the following strategies (Blachowicz et al. (1990: 12): 1) they searched for a meaning that seemed to fit the target sentence and stopped as soon as their criteria were met without looking past the definition they chose (in some cases, they stopped after reading the first definition); 2) they read through each and all definitions and then went back and chose the one thought to have the best fit; 3) they read all the definitions but proceeded to formulate their own definitions which were more encompassing of all the definitions, looking for something in common among the definitions. Though the research concerned the skills and strategies of unskilled dictionary users using a native speaker's dictionary, its results were comparable to some of the previously reviewed studies which involved the use of learner's dictionaries. Like the adult learners of previous studies, the children in this study concentrated mainly on definitional or central information rather than ancillary information in an entry. In other words, they failed to take full advantage of the wealth of information included in an entry. The preceding studies were valuable in revealing weaknesses of skill, but they were all small in scale. The grand-scale project undertaken by Atkins et al. (Atkins et al. 1987, Atkins and Knowles 1990) aimed at investigating the efficacy of dictionaries in helping foreign learners of English to perform various reading and translation tasks; effectiveness of bilingual and monolingual dictionaries in meeting specific needs; and training of reference skills. One thousand six hundred sets of papers (400 in French, German, Italian and Spanish, respectively) were distributed and over 1,100 students responded. A questionnaire and two

15

tests (a language proficiency test and a research test) were used as research instruments. The results relevant to this study included the discovery that a majority of respondents had never received any dictionary use instruction, while only a minority (12.9%) had been given systematic instruction in reference skills. The respondents showed differences in their levels of skills. These appeared to be related to the type of task rather than the respondents' language proficiency level. Nearly all students were able to interpret grammatical labels. However, when they were asked to locate multi-word items such as lame duck, 55.7% of the most proficient students expected to find it under entry duck in spite of the fact that all monolingual learner's dictionaries entered compounds like this according to their initial element (in this case, under entry lame). Though the researchers did not explain why this might have been the case, Béjoint (1981) has argued that more advanced students assume that 'structural factors' rather than 'linear order factor' determine the location of a dictionary entry because they perceive that in adjective-noun compounds the noun is the head (and thus the headword they should look up). In an earlier study, Mitchell (1983) studied how some secondary school children in Scotland used their dictionaries while reading. The children were also required to complete a number of writing tasks, one of which involved completing sentences in which target words appeared. Mitchell reported that the children tended to make mistakes when adopting either one of the following two strategies: when they focused on the easily-understood parts of the definition or when they misread an unfamiliar word as a familiar word. To illustrate, the looking up of the target word aspen was used. Poplar was found in the definition: aspen noun, a kind of poplar whose leaves quiver even in a light breeze Some children wrote sentences such as: An aspen is a kind of leaves, or An aspen is a kind of quiver. because they focused on the parts they could understand easily such as leaves and quiver. Mitchell also discovered that some children extracted the elaboration instead of the meaning of a word in a definition. For example a child looked up: vestment noun. A ceremonial robe, especially one worn by the clergy during religious services. He wrote a sentence: A vestment is a kind of religious service. Other dictionary use skills identified were searching for the appropriate headword, understanding the structure of the entry, identifying the relevant part of the definition, relating the appropriate sense to the context provided, and merging the word with the context of the source text. Miller and Gildea (1985: 13 - 26) asked ten 11-year-old native speakers of English to look up given words in a dictionary and then write sentences illustrating their use. They found that the single most common cause of error was the use of a substitution strategy they called 'kidrule', which is made up of a sequence of four steps: 1) Read the definition 2) Select a short familiar segment 3) Compose a sentence containing that segment

16

4) Substitute the target word for the selected segment An example sentence produced by the 'kidrule' strategy is as follows: I was meticulous about falling off the cliff.

Here the subject selected a familiar piece (i.e. very careful) of the definition: 'meticulous...very careful or too particular about small details'. He then composed a sentence, and substituted the entry word for the familiar piece in his sentence. Step (2) of the Tddrule' procedure involves the look-up strategy of reading/selecting only familiar segments from the dictionary definition. Dictionary defining style was also suspected to cause the production of inappropriate sentences. Difficult defining language and inadequate one-word definitions may have caused some Tddrule' sentences. The findings of the study are in many ways similar to that of Mitchell (1983) though they designed their experiments independently. The 'kidrule' sentence is similar in kind to those produced by Mitchell's subjects. The evidence from the studies conducted by Mitchell (1983) and Miller & Gildea (1985) shows that native-speaker children can misread long and explicit entries which may create their own particular problems. It is also possible that the children will deal with only part of a long entry, and this part may not even be the kernel definition, but may be an example phrase which simply provides context in the dictionary. When it comes to dictionary use, it is possible that adult non-native speakers adopt the Tddrule' strategy. Like Miller and Gildea (1985), Nesi and Meara (1994), collected data by asking their subjects to compose sentences after consulting dictionary entries. However, Nesi and Meara's subjects were asked to use a pair of words instead of a single word in a single sentence. A computer program was also used to present the task and store the data. Results showed that 'kidrule' was responsible for about a quarter of all errors in sentences the subjects composed. Four other strategies led to errors: "failing to apply grammatical information supplied by the dictionary, failing to choose appropriate lexical collocations, confusing words which looked or sounded similar to each other, or rejecting information from the entry which did not match the subjects' preconceived notions of what words meant" (ibid.: 11). All except the last strategy were well supported by the empirical evidence. The last strategy apparently was based on speculation because the subjects were unable to give any explanation and nothing in the entry offered a clue as to what the subjects intended to express in the sentences that they composed after looking up. Thus, the conclusion that the error was caused by the subjects' preconceptions may have been too hastily made. This points to the value of process-oriented research methods, such as think-aloud, which may be able to reveal cognition and metacognition in the look-up process. By devising a 'paper equivalent' of the think-aloud protocol, Atkins and Varantola (1997) recorded what a group of skilled dictionary users (translation students and lexicographers) really did when they consulted dictionaries to solve a linguistic problem in translation and what the group's level of satisfaction was after each search. The subjects worked in pairs, one partner using dictionaries, and the other monitoring and recording every step of his partner's activity on forms designed by the researchers for this purpose. The researchers did not attempt to discover in these experiments how 'successful' the look-ups were, by rating the choices made as correct or incorrect because they argued that normally dictionary users

17 translating into a foreign language often knew if they felt satisfied with the information they retrieved rather than knowing immediately whether they had the right solution. Thus, the researchers invited an account of "the level of user satisfaction after each search" (p. 12). The subjects were asked if they were convinced that they had got the right information, doubtful, or sure they had not been successful (p. 7). It was reported that the subjects consulted dictionaries mostly in order to find a translation. A majority of the users consulted bilingual rather than the monolingual dictionaries and their look-ups in the former tended to be more successful. A monolingual dictionary was often used after a series of unsuccessful bilingual look-ups. Only in 59% of 1000 look-ups were the users satisfied with the dictionaries they used. 91% of the look-ups in their study were for meaning. Therefore, the main results of the research are similar to those of the questionnaire studies reviewed in Section 2.2.1.1. Three case studies were cited by Atkins and Varantola to illustrate the strategies used in 'unsuccessful' look-ups. In the first case when the single-word target was not found in the L2 monolingual dictionaries as well as in a bilingual dictionary, the users usually adopted the strategy of abandoning the search. In the second case when a multi-word target was partially found in the dictionary (e.g. finding either 'low', 'wage' or 'wage earner' but not the phrase 'low wage earner'), the users either gave up and composed some kind of paraphrase or chose to continue the search. The researchers speculated that this behaviour could be the result of the abnormal experimental circumstances of the dictionary use. In the third case when an idiom target was partially found in the dictionary (e.g. finding only 'put' of the idiom 'to put all one's eggs in one basket'), the strategy of abandoning the search was again used. The researchers concluded that: ..."case studies throw some light on dictionary users' strategies for handling the frustrating situation of finding what they want in a dictionary entry. Sometimes the tactics employed are sophisticated and sensible; sometimes with the benefit of hindsight the user might have wished to give up earlier; sometimes a more intelligent search strategy would have found the item in the dictionary" (p. 11).

Wingate (2000) asked two main research questions: whether a bilingual or monolingual dictionary is most effective for Hong Kong Chinese intermediate learners of German for reading comprehension and what features make monolingual dictionary definitions effective. The investigation was conducted by means of a questionnaire, experiments and think-aloud. The results indicated that the German-English bilingual dictionary was not significantly more effective for intermediate language learners than the monolingual dictionary which was linguistically too difficult for this proficiency level. The first think-aloud study identified the reasons for successful and unsuccessful look-ups in the bilingual and monolingual dictionary. It appeared that if the target meaning was found in the initial position of an entry, the success rate was high. 'Kidrule' was also found to be the cause of a majority of look-up errors. The main reason for successful look-ups in the monolingual dictionary was related to definition. The definition must contain a segment which was familiar to the subjects and could be substituted as an equivalent for the target word. The second main think-aloud study identified features which made monolingual dictionary definitions effective for intermediate learners. The effective features were: redundancies and examples as part of the definition; if available, equivalents; and in initial position. The ineffective features were: unrestricted defining vocabulary; derivational definitions; and definitions that contain unknown superordi-

18

nates (ibid.: 315). The choice of look-up strategies also was shown to affect the outcome, although the study only dealt with this aspect superficially. There were also empirical studies of bilingualised dictionary use, but these will be reviewed after considering the structure of the dictionary in more detail.

2.3. The bilingualised learner's dictionary

This section examines three aspects of the bilingualised learner's dictionary. First, it provides an overview of this genre of learner's dictionary. Second, it deals with the principles of defining. Finally, it reviews the empirical studies into the use of bilingualised learner's dictionaries.

2.3.1. Overview The studies reviewed so far are of monolingual or bilingual dictionaries, but there is a third kind of dictionary called here the bilingualised dictionary. 'Bilingualised', 'glossed', 'translated', 'hybrid' or 'semibilingual' learner's dictionaries for foreign learners are the result of partial or full adaptations of existing monolingual learner's dictionaries (Hartmann 1994a, Cowie 1999). As revealed by Hartmann (1994a), monolingual learner's dictionaries are bilingualised either by the original publisher e.g. Oxford University Press, or by a publishing house (such as Kernerman) which has acquired the bilingualisation rights from another publisher (e.g. Harrap). They have been in circulation for more than thirty years2 and, since the early 1980s, more than fifty such dictionaries have been published for learners of English who speak languages such as Chinese, German, Hebrew and Japanese. Bilingualised versions of English learner's dictionaries attempt to "bridge the present gulf between the bilingual and the monolingual" (Atkins 1985: 22). The gulf has been formed because of the on-going debate as to which dictionary type is a better educational tool for foreign students. For many years, most language educationists have accepted the orthodoxy that learners should be encouraged to use monolingual dictionaries rather than bilingual dictionaries (Underhill 1985, Atkins 1985). Foreign language learners, on the other hand, tend to favour the traditional general-purpose concise or pocket bilingual dictionaries (SnellHornby 1987: 159, Thompson 1987: 283) and do not support the use of the monolinguals. The arguments against the use of monolingual dictionaries (Thompson 1987: 283 - 284) are: 1) For production, when learners want to look up a word they know in their own native but not in the foreign language, they do not know where to find it in the dictionary. 2) For comprehension, learners may not directly use much of the information provided, e.g. grammatical behaviour. 1

Hartmann (1994a) has pointed out that translated dictionaries have, in fact, been compiled in Europe since the Renaissance and such dictionaries deserve a mention in dictionary history. These pioneers and predecessors are, however, beyond the scope of this study.

19 3) Learners may find it difficult to understand the foreign language definitions 4) Circularity in the definitions may cause problems for language learners. 5) Even when restricted defining vocabulary is used in the definitions, if learners do not know the vocabulary, they will not be helped. If they know the vocabulary, then the definitions are unnecessary. 6) Learners will not benefit from being exposed to the foreign language via the definitions, which are mostly complicated and of a special register (see also Tomaszczyk 1983, Run dell 1988, Piotrowski 1989, Stein 1990: 404). 7) Learners are quite likely to pass over semantic information in a definition crucial in un derstanding a word or distinguishing it from other words of related meanings because they do not understand the basic content of the word. 8) Learners may be led to a never-ending search for explanations of the words used in the definitions (Tomaszczyk 1983: 46, Stein 1990: 404). 9) There is also no guarantee that learners really know a foreign lexical item until they can refer to its LI equivalent (Tomaszczyk 1983: 46, Snell-Hornby 1987: 159). The essential arguments against the use of bilingual dictionaries are: 1) Learners are not encouraged to think directly in the foreign tongue. 2) Learners' inclination to translate from and into their mother tongue is reinforced. 3) The belief in one-to-one correspondence between LI and L2 may be wrongly reinforced. In turn, the internalization of L2 could be prevented (Atkins 1985: 19, Snell-Hornby 1987: 165, Stein 1990: 403 - 404). 4) Learners may fail to find adequate descriptions of the syntactic behaviour of words (Thompson 1987: 282). 5) Looking up translation equivalents does not develop the learners' skill in paraphrasing or defining (Baxter 1980: 329). 6) It is not unusual to discover an absence of an equivalent to a given L2 lexical item (i.e. nil-equivalence) due to reasons such as the differences in culture and in the surrounding world (Tomaszczyk 1983: 48, Zgusta 1984: 149, Snell-Hornby 1987). 7) An equivalent does not necessarily cover the same semantic area as the L2 lexical item (Kromann et al. 1991: 2718). 8) Stylistic and/or connotational equivalence is not guaranteed. An equivalent may not con vey to its reader the same aesthetic and denotative meaning as conveyed by the original text (Zgusta 1984: 151, Tomaszczyk 1983: 47, Reif 1987: 154). Taking the above arguments into consideration, it seems logical to assume that the userlearner would benefit most from the two dictionary types by adopting the strategy of moving back and forth between them (Thompson 1987: 282, Stein 1990: 402, Zofgen 1991: 2897). The bilingualised learner's dictionary, which combines these two genres of dictionary into one, and consequently combines the best features of each type, would certainly facilitate such a strategy if learners choose to use it. Bilingualisation of a monolingual work is achieved by one of the two approaches: partial and full bilingualisation. The former approach usually only translates the L2 definitions into LI whereas the latter translates everything in each entry. Harrap's English Dictionary for Speakers of Chinese (HEDSOC), published in Beijing in 1991, is a partial adaptation of Easy English Dictionary. As can be seen from the following entry (i) for fag (HEDSOC,

20

cited in Hartmann 1994a: 246), only one-phrase translations of the main senses of each headword are provided and the English definitions and examples are left intact. (i) Entry for fag from HEDSOC (1991) f a g [ f ® j ] a. 1 a tiring/boring work. 1 ( a M j ) cigarette. • ». 1 to work hard. S j b l f l s X fm Jagged out, ( ¿ o f ) tired out. H49&

Longman and Oxford, the two main competing dictionary publishers in Hong Kong, produced fully bilingualised dictionaries e.g. LECDOCE (1988), OALECD (1984). The entries for fag (ii & iii) below (cited in Hartmann 1994a: 246) show that the entries are fully translated in Chinese, from headword and definition to all illustrative examples. (ii) Entries for fag from OALECD (1984) fag1 Itxo; tutl « 1 IC U1 (slug only! (colloq) lirini job: (dm*«) (B)CJMU'.CVi£r the illus. of arm. (FM) Formulating Meaning: formulating own meaning/definition by using own words as far as possible after reading dictionary definition and/or after reading contextual clues. On Both Reading and Dictionary Texts: (FP) Focusing Pronunciation : concentrating on pronunciation of target word or concentrating on different pronunciations/phonetic symbols of target headword. (M) Matching: realizing spelling of target word in reading text partly/wholly matches with that of target headword in dictionary text or realizing spelling of target word partly matches with that of target headword. (FF) Focusing Features: Focusing on features such as pictorial illustrations related to target headword in dictionary text or mechanical features such as italics, bold face related to target word/headword in reading/dictionary text. (RSM) Realizing Shared Meaning: realizing some target words share the same or similar meaning after looking them up. (CD) Choosing Definition: choosing LI translation equivalent and/or L2 definition which is/are considered to fit the reading text best. (RM) Refining Meaning: fine-tuning meaning after formulating word meaning from contextual clues or fine-tuning meaning after formulating own meaning/definition from dictionary definition(s). (IM) Inferring Meaning: Inferring meaning of target headword with the help of contextual clues. Coding 3: Metacognitive Operations During Dictionary Look-up (PL) (UPR) (ISS) (ISP) (MO)

Planning: Making plans at the beginning of look-up task on how to tackle the task or setting look-up task goals or task demands. Using Prior Knowledge: drawing on prior linguistic or semantic knowledge of target word to decide whether to start, continue or discontinue looking it up. Initiating Search Strategy by Spelling: starting the spelling-driven word search by spelling out the whole or part of the target word. Initiating Search Strategy by Pronouncing: starting the spelling-driven word search by pronouncing the whole or part of target word. Monitoring: monitoring the progress or process of the look-up task.

65 (PMS) Postponing Meaning Search: postponing search for meaning of target headword after dictionary user expresses doubts or have queries about the given dictionary definitions. (TMS) Terminating Meaning Search : ending search for meaning of target headword after dictionary user decides that s/he knows the meaning. (ET) Evaluating Task: making general comment(s) on look-up task, think-aloud method, and dictionary by summarizing, drawing on personal experience and/or linking concepts. (TLT) Terminating Look-up Task: ending look-up task after dictionary user decides that s/he has looked up all the words s/he intends to look up. Note: The same look-up word is termed 'target word' in the reading text or 'target headword' in the dictionary text.

4.3. Discussion

As mentioned in Section 4.1, previous models of look-up behaviour were based on the researchers' introspection. Studies based on empirical data have not attempted to systematically tap the cognitive aspect of look-up strategies. No attempt has yet been made to offer a tool to analyze and describe look-up strategies. The present study has attempted to bridge this research gap by offering a coding scheme of look-up operations which was derived from verbal data. The coding scheme could also be used to analyze and describe look-up strategies in either monolingual or bilingual dictionaries because it was generated from data involving the bilingualised learner's dictionary, which had both monolingual and bilingual parts in one. Besides achieving the aim of devising coding categories to analyze the data the quality of the data collection procedures - think-aloud and stimulated recall interview - was also assessed. The results showed that both think-aloud and stimulated recall interview were to a large extent practicable. The administration of the procedures was in general smooth and unproblematic so only minor revisions of the practical aspects of the procedures were necessary. This coding scheme, however, was not comprehensive because it was derived from 20 sets of think-aloud data. By increasing the amount of data, more operations could be identified.

4.4. Summary

This chapter has addressed the first research question. After identifying operations from the look-up data, it is possible to describe look-up strategies in terms of regularly occurring se-

66

quences of operations. The next chapter will investigate the number of possible ways the operations can be sequenced, i.e. how many types of look-up strategies.

5. Dictionary Look-up Strategies

5.1. Introduction

This chapter will address the second research question: What look-up strategies do students use when they are reading and need to find the meaning of a target word in a bilingualised English-Chinese learner's dictionary? Although some codes generated in Chapter 4 refer to English or Chinese (e.g. FE/FC), strategic patterns were first scrutinized independently of this distinction. The question of whether or how strategies are distinguished in terms of the language focus is one that can wait until Chapter 6. Section 5.2 of this chapter will describe the process of identifying and categorizing lookup strategies and Section 5.3 will describe and illustrate each look-up strategy identified in the data. Section 5.4 will discuss the validity of the method and the strategy descriptions.

5.2. Process of identifying and categorizing strategies

In this study, a look-up refers to the looking up of one entry of one target word, once, in one dictionary. The success of a look-up is judged by rating the choice of word meaning as correct or incorrect by the researcher. This choice of meaning could be English explanation or Chinese equivalent. With 55 target words (see Appendix 4) and 18 students participating in the main study, 264 look-ups were identified in the think-aloud data. The success rate was high - 27 wrong solutions in total. Other researchers (Hartmann 1994b, Laufer and Hadar 1997) have also observed high success rate - evidence for the strength of the bilingualised dictionary. As discussed in Chapter 4, the coding procedure assigned to each operation, in a look-up, an executive code with a matching cognitive or metacognitive code. As a result, a coded look-up consisted of a sequence of operations. A look-up strategy is therefore defined as a systematic application of certain operations. More precisely, a look-up strategy is a regular pattern of operations used by one or more than one user. All 264 coded look-up strategies were found to include a sequence of basic or 'core' executive, metacognitive or cognitive operation codes. Each strategy incorporated the following 'core' executive operations: starting search (SS), searching headword (SH), referring headword (RH), referring definition (RD), and writing definition (WD). The matching 'core' metacognitive operations for the first and last 'core' executive operations (i.e. SS and WD) were: initiating search strategy by spelling or pronouncing (ISS/ISP) and terminating meaning search (TMS). The 'core' executive operations in between were paired up with the following cognitive operations: searching alphabetically (SA), locating (L), and focusing English (FE) and/or focusing Chinese (FC). In the coded think-aloud data, the 'core' operations were presented in three columns:

68

Executive operations SS SH RH RD WD

Cognitive operations

Metacognitive operations ISS/ISP

SA L FE/FC TMS

Hereafter, these operations were referred to as 'core operations'. To identify and categorize look-up strategies, a card-sorting method was used. 264 coded strategies were printed on cards and piled. The pile of 264 cards was eventually sorted into separate piles by the researcher who began going through the pile of cards by asking a contrast question, "Is the strategy different from the last one in some way?" Consequently, seven piles of cards corresponding to seven types of look-up strategy emerged. The strategies (see Table 5.1 for frequency among the look-ups) were named according to their criterial features: 1 Ignoring - ignoring the use of ancillary information. 2 Assuming - assuming the only meaning in an entry was appropriate for the target word without showing evidence of context-matching. 3 Minimizing - minimizing the use of central definitional and ancillary information in an entry. 4 Checking - checking inferred meaning against the dictionary meaning to see if they matched. 5 Paraphrasing - paraphrasing a chosen dictionary definition in order to achieve a better contextual fit. 6 Stretching - stretching an idea which was common for all definitions in an entry. 7 Maximizing - maximizing the use of central definitional and ancillary information in an entry. Table 5.1 Frequency of strategies among the look-ups Look-up Strategies Ignoring Assuming Minimizing Checking Paraphrasing Stretching Maximizing Total

Frequency 40 45 36 6 14 11 112 264

5.3. Description and illustration of look-up strategies

The following sub-sections (5.3.1 to 5.3.7) describe and illustrate the seven strategies identified in detail. Relevant segments of the coded think-aloud protocols are used to illustrate the strategies. It should be emphasized that, though the protocols probably provide more information about the cognitive and metacognitive processes of these students than any other single technique, they are only capable of capturing traces of the cognition which takes place during their look-ups. Thus it was necessary for the researcher to use the students' stimulated recall interview data and her interpretative skills to create a coherent picture of the students' look-up behaviour.

5.3.1. Ignoring strategy The ignoring strategy is characterized by its complete neglect of the ancillary information in an entry. The subjects used the ignoring strategy to look up words in both monosemous and polysemous entries. They monitored the start of the consultation of the target word. If the target word was an inflected verb, they recovered its canonical form before starting to search for the target headword in the alphabetic list of the dictionary. After locating the entry of the target headword, they scanned all the senses, chose one sense, ignoring the ancillary information in the entry such as grammar. The chosen sense was subsequently fitted into the context. The ending of the consultation process was closely monitored. The protocol data showed that the students who used this strategy were consciously aware of the operations and decisions they had made during the consultation process. Therefore, all ignoring strategy examples involved the 'core operations' (see 5.2) and the matching operations of "referring text - constructing from context" (RT - CC). Table 5.2, which contains a segment of the coded think-aloud protocol, shows how Wing successfully found the meaning for the target word crucial in his OALECD (1984) by using the ignoring strategy. The top row of the table contains the reading context of crucial. Column 1 shows the utterances of Wing (Cantonese original - English translation; English utterances within single quotation marks) during the dictionary consultation process. Columns 2, 3, and 4 contain executive, cognitive and metacognitive codes respectively (refer to 4.3.4 for definitions of the coding categories). Table 5.2 The use of ignoring strategy to look up crucial by Wing Target word in context Perhaps, but all are crucial to any venue specialising in Chinese cuisine. Think-aloud protocol segment Communication Units (CU) Cognitive Metacognitive Executive operations operations operations 9 CRUCIAL SS ISS 10 CRU... SH (turning SA pages) 11 CRUCIAL RH L

70 12 Means decisive (kuei3 ding6 sing3), very important (chung5 yiu3) 13 Em.doesn't have any problem.it matches here

RD

FC

RT/WD (writing down chung5 yiu3)

CC

TMS

Here is entry crucial from OALECD (1984) cru-cial /'kncjl ; 'krufsl/ ad} decisive; critical: ik tht - ust/vtstim; ifcJE&iW WKBUB): «' th* ~ moment,tesgoyust.—ly /-fab;-W : She knows haw to scare off people who ask sellingraywdL o She ¿¡¡¡oUt questions. tt«#H«JMMiiliiuttAatt *"*ed against htm In quoting hts earlier statementtt-fl-fjg aeon ath Out/through draw a lioe or Ones ffl(|W*»Kleiktl9IT. o score an instant success a through sth M«M£(:jCT atroka , 1 (To f act aa a stroke'p) to fa boat or rowing crew) oarsman who sits nearest the stem of a 1 crew) H •. • M M M f . 2 [Tn-pr, To-p] strike (a ball) V radng boat, and seta the speed of the strokes (MMW4 W.stwked the baU cleverly past her opponent »Pito* 99 < « [ » » « . * « M * ) . Cf • » aowr2. 4 -»«»JT«»*. — Of ath single successful or effective action or atroka' /stnuk; strok/ •> [Tn] paaa the band gently over occurrence (of toe specified kind) atroka n (usu sing act of stroking: misfortune led to his ruin. stroking movement 0: give her hat an (mark made by a) single movement of a pen or brush affectionate stroke KtttJMKSIitttMK. thin/thick strokes ° with a stroke of the pen ffl*—W o put the finishing strokes to a painting ft-bftttAff. 6 sound

¡M*.

B^li. 5 flifttJ

(net«*». 0 5ttMM • XAMtt. 0 IK; Hi

made by a bell or dock striking the hours (WWW on the stroke of three, ie at three o'clock exactly

The think-aloud data showed that Zoe began the consultation process by focusing on the pronunciation of the infinitive form of stroked (CU 170). Although her verbalization of the spelling of stroke during the search process was minimal and fragmented, one could still assume that its complete spelling was in her cognition (CU 171) simply because she required complete spelling to locate the headword. Before starting the word search, Zoe determined that stroked was inflected. Thus she recovered its infinitive form (i.e. stroke) before searching. Upon finding entry shred (CU 172), she referred to the reading text and found that stroked was followed by on. Armed with this piece of information, she decided to look up this verb-particle combination (CU 173). From this point onwards, she shifted from searching for stroke to stroked on. However, instead of finding stroked on, she found the phrasal verb stroke down and its meaning (CU 174, see verb entry stroke). She then fitted the meaning into the context. She did so probably because she thought that stroked on and stroke down shared similar meaning. As a result of context fitting, she found that the meaning of stroke down (i.e. on' foo2) was not the right meaning for stroked on (CU 175). The context was verbalized as follows: blades can be stroked on an inverted earthenware

She moved on to the first (i.e. da2 gik') of the 8 senses listed in the noun entry of stroke (see noun entry stroke), selected da2 gik1 tentatively as the meaning of stroked on, and verbalized that da2 gik1 was the extension of the idea of posture (CU 176). Next, she searched the reading text for more clues and indicated that stroked on did not seem to have a matching meaning in the noun entry (CU 177). She then verbalized the 6th English definition in the noun entry i.e. single movement of pen or brush (CU 178). She rejected it and then guessed that gui2 hei2 (lift up) was the meaning of stroked on (CU 179). According to her gui2 hei2 was the extension of the idea of movement which was common for all the definitions in the entry. She could also infer this idea from the source context. In the interview, she elaborated this inference of meaning more fully. The relevant excerpt reads:

82 It [stroked on] seems to be describing how to hold the cleaver, .the movement or direction. I don't know if this is correct so choose this first. Because I want to find stroked on, that is, the phrasal verb..but it [verb entry of stroke] has none.

She then decided to end the meaning search of the target word stroked (CU 180). Stroked on is not a phrasal verb and the meaning (i.e. gui2 hei2) is wrong, therefore this look-up was unsuccessful. Though the look-up was unsuccessful, it illustrates the stretching strategy well.

5.3.7. Maximizing strategy The maximizing strategy maximizes the central definitional and ancillary information in an entry. Like the ignoring, minimizing, checking, paraphrasing and stretching strategies, it performs context-fitting. The subjects used the maximizing strategy to look up words in monosemous and polysemous entries. If the target word was an inflected verb, they recovered its canonical form before starting to search for the target headword in the alphabetic list of the dictionary. They monitored the initiation of the word search. After locating the entry of the target headword in the dictionary, they scanned all the senses. To help them reduce the number of possibilities to one in a polysemous entry, they retrieved all the available ancillary information in the entry: grammar, illustrative example, usage, or idiom. Next, they fitted the chosen meaning to the context where the target word was found. Finally, they monitored the end of the meaning search. In general, the students indicated that conscious awareness was involved in many of the operations and decisions made in the consultation process. Therefore, like the above six strategies, the maximizing strategy had the 'core operations'. Unlike the assuming strategy, it included the matching operations of "referring text - constructing from context" (RT - CC). Unlike the ignoring strategy, it involved the "ancillaryinformation-related" operations. It differed from the minimizing strategy in that it involved 'all' the available "ancillary-information-related" operations. Table 5.8 shows how Ha successfully found the meaning for the target word savoured in her OALECD (1984) by using the maximizing strategy.

83 Table 5.8 The use of maximizing strategy to look up savoured by Ha Target word in context Before the tastes of kung po chicken, moo shu pork and spring rolls can be savoured, you can be certain that all the fillings have come under the trusted blade of the multi-purpose cleaver. Think-aloud protocol segment Communication Units (CU) MetaExecutive Cognitive cognitive operations operations operations 2 SAVOURED SS ISS 3 so this 'ed' is..eh..this is 'passive voiceV.so RG UG if I look up the dictionary I'll just look up SAVOUR..E 4 so I'll try to see if I can look this word SH (turnSA up..first I'll open this dictionary..to S..the ing dicpart..then turn to this page..left tionary page..comer..top left corner is..SPACE.. pages) now have to look up is SAVOURED..eh.. SAVOUR..RE..so this is SP and I've to look up SA..so that means it's a few pages before this..so turn to a few pages before this..so I've turned to SO..that means I've not turned to the page that I want so I continue to turn the pages until to SA..okay here I've turned to SA..the third letter is SAT..but I want to find SAV..so read the right side again..eh..this is the right pagehere has SAVE..but I've to find SAVO..that is to say the word that I want to find is probably on this page..so..SAVO..SAVOUR 5 so I've find a word that looks similar to the RH L word that I want to find..and it is SAVOUR., there is no E 6

so this word has..first line..'symbol' teaching us how to 'pronounce' 7 so I skip this first I'll read its meaning because in the 'passage' is a 'verb' 8 so I'll go to the 'verb' entry 9 here there are 2 meanings if it's used as a verb 10 the first meaning is 'appreciate the taste or flavour or character of

RP

FP

RG

FG

REN RD

FEN RGD

RD

FE

84 11 the Chinese meaning is to appreciate (yan' seung2) some taste (mef do6) or flavour (Jung' mei5) 12 also read the examples of usage..eh.."he'., next is how to read this..this should be..eh ..I'm not sure how to read this..just read this as 'our savour..he savoured the wine/the joke' 13 so it means he appreciates (ban2 seung2) the taste (mef do6) of wine 14 so the second meaning is..em..'suggest the..em.suggest the presence1 15 its Chinese meaning is having the presence (yi3 mef ) of some kind 16 eh..the example is 'such a proposal..eh.. savours of impertinence' 17 this proposal brings the presence (yi3 mef) of some kind 18 so I personally feel that should use the first meaning that is appreciate (yan' seung2) some kind of taste (mef do6) flavour (Jung' mef) 19 em .because this passage is mainly about food so therefore I think this meaning is more appropriate 20 so I'll write down its meaning..so is..eh., 'appreciate..appreciate the taste' that is appreciate (yan' seung2) some kind of taste (mef do6)

RD

FC

RX

FX

RX

FX

RD

FE

RD

FC

RX

FX

RX

FX

RD

CD

RT

CC

WD (writing down appreciate the taste, yan' seung2.. mef do6)

TMS

Here is entry savour from OALECD (1984) sa-vour (US » -vor) /'sem(r); 'sev»/ n [C, U] - at. taste or flavour (of Ith): suggestion (of a duality): (x»«»»ill:ia*:ft!«: