Patterns and Meanings : Using corpora for English language research and teaching [1 ed.] 9789027298911, 9789027222701

Citation preview

patterns and meanings

SCL

Studies in Corpus Linguistics Studies in Corpus Linguistics aims to provide insights into the way a corpus can be used, the type of ¼ndings that can be obtained, the possible applications of these ¼ndings as well as the theoretical changes that corpus work can bring into linguistics and language engineering. The main concern of SCL is to present ¼ndings based on, or related to, the cumulative e¤ect of naturally occuring language and on the interpretation of frequency and distributional data.

General Editor Elena Tognini-Bonelli Consulting Editor Wolfgang Teubert Advisory Board Michael Barlow (Rice University, Houston) Robert de Beaugrande (University of Vienna) Douglas Biber (North Arizona University) Wallace Chafe (University of California) Stig Johansson (Oslo University) M.A.K. Halliday (University of Sydney) Graeme Kennedy (Victoria University of Wellington) John La¤ling (Herriot Watt University, Edinburgh) Geo¤rey Leech (University of Lancaster) John Sinclair (University of Birmingham) Piet van Sterkenburg (Institute for Dutch Lexicology, Leiden) Michael Stubbs (University of Trier) Jan Svartvick (University of Lund) H-Z. Yang (Jiao Tong University, Shanghai) Antonio Zampolli (University of Pisa)

Volume 2 Alan Partington Patterns and Meanings Using corpora for English language research and teaching

Patterns and Meanings Using Corpora for English Language Research and Teaching

ALAN PARTINGTON

john benjamins publishing company amsterdam / philadelphia

8

TM

The paper used in this publication meets the minimum requirements of the American National Standard for Information Sciences – Permanence of Paper for Printed Library Materials, ansi z39.48-1984.

Cover design: Françoise Berserik Cover illustration from original painting Random Order by Lorenzo Pezzatini, Florence, 1996.

Library of Congress Cataloging-in-Publication Data Partington, Alan. Patterns and meanings : using corpora for English language research and teaching / Alan Partington. p. cm. (Studies in Corpus Linguistics, issn 1388-0373 ; v. 2) Includes bibliographical references and index. 1. English language--Research--Data processing. 2. English language--Discourse analysis--Data processing, 3. English language--Study and teaching--Data processing. 4. Computational linguistics. I. Title. II. Series. PE1074.5.P37 1998 420’.285--dc21 98036312 isbn 978 90 272 2270 1 (EUR) / 978 1 55619 343 9 (US)  (Hb ; alk. paper) isbn 978 90 272 2271 8 (EUR) / 978 1 55619 396 5 (US)  (Pb ; alk. paper) isbn 978 90 272 9891 1 (Eb)

© 1998 – John Benjamins B.V. No part of this book may be reproduced in any form, by print, photoprint, microfilm, or any other means, without written permission from the publisher. John Benjamins Publishing Co. · P.O. Box 36224 · 1020 me Amsterdam · The Netherlands John Benjamins North America · P.O. Box 27519 · Philadelphia pa 19118-0519 · usa

Contents

Acknowledgements 0 Introduction 0.1 Corpora and linguistic description 0.2 Corpora and language learning 0.3 Approaches to the use of corpora in language pedagogy 0.4 Guide to the contents of this book 0.5 Equipment and methodology

viii 1 2 4 5 7 9

1 Collocation and phrase patterns 1.1 Definitions of collocation 1.2 Collocation, text-type and style 1.3 Collocation and communicative competence 1.4 The ‘‘idiom’’ or ‘‘collocational’’ principle 1.5 Collocation and linguistic schemas (or schemata) 1.6 Routines and patterns in language acquisition studies 1.7 Types of collocation 1.8 Conclusion

15 15 17 18 19 21 23 25 27

2 Collocation and synonymy 2.1 Synonymy, translators and language learners 2.2 Definitions of synonymy 2.3 Discriminating among semantically similar items: the case of sheer 2.3.1 Semi–grammatical items—a problem for learners 2.3.2 The senses of sheer 2.4 How synonymous are the synonyms of sheer? 2.4.1 pure 2.4.2 complete 2.4.3 absolute 2.5 Conclusion 2.6 Suggestions for further research

29 29 29 33 33 34 39 39 42 44 46 47

vi

contents

3 True and false friends 3.1 Equivalence in translation 3.2 The evaluation of context 3.3 Look-alike words and false friendship 3.4 The problem for translators: a limit example 3.5 Comparing items across languages: correct and corretto 3.6 A further experiment: absolutely, completely, entirely and their Italian look-alikes 3.7 Conclusion

48 48 50 51 53 53

4 Connotation and semantic prosody 4.1 Connotation 4.2 Semantic prosody 4.3 Semantic prosody and dictionaries 4.4 Prosodies and persuasion: sharp dealings 4.5 The creation of prosodies 4.6 Conclusion 4.7 Areas for further research

65 65 66 69 72 74 76 77

5 Syntax 5.1 Investigating verb constructions 5.2 If constructions and language learners 5.3 Clause combinations 5.4 Expanding the investigation: Beyond the use of if as conditional 5.5 Further research

79 80 81 81 84 87

6 Cohesion in texts 6.1 Text reference 1: General noun phrases 6.1.1 The functions of general nouns 6.1.2 An example: a move as a general noun phrase 6.1.3 Suggestions for further research on general noun phrases 6.2 Text reference 2: Labelling noun phrases 6.2.1 The functions of labelling noun phrases 6.2.2 Examples: allegation and claim 6.2.3 Suggestions for further research on labelling noun phrases 6.3 Text reference 3: General verbs 6.3.1 Happen 6.3.2 Occur 6.4 Conclusion

89 90 90 92 96 96 96 97 101 101 101 102 104

56 62

contents

vii

7 Metaphor 7.1 Systematic metaphors and genre-typical metaphors 7.2 An example search 7.3 Theories of metaphor 7.4 Metaphor in business journalism: Data and methodology 7.5 Metaphor in business journalism: Results 7.6 Conclusion: The living and the dead 7.7 Further research

107 107 108 109 111 112 117 119

8 ‘‘Unusuality’’ 8.1 Definition and methodology 8.2 Quotations 8.3 The four mechanisms of change 8.4 Productive phraseology 8.5 Proverbs and sayings 8.6 Double sense and relexicalisation 8.7 Expressions and collocations 8.8 Plays on frequently co-occurring items 8.9 Collocation and the brain 8.10–Other possible research methodology 8.11–Conclusion

121 121 122 125 128 131 133 135 138 139 140 142

9 General conclusions 9.1 Some criticisms and limitations of corpus study 9.2 The role of this book

144 144 148

References

151

Index

159

Acknowledgements I am grateful to a number of colleagues and friends who provided intellectual input and/or encouragement and support at some stage in the gestation of this book. These include members of the Cobuild group at the University of Birmingham, particularly Gill Francis, Ramesh Krishnamurthy and John Sinclair, the members of the Forlì research team, namely, Guy Aston, Franco Bertaccini, Ruey Brodine, Laura Gavioli, Federico Zanettin, Daniela Zorzi-Calò, William Dodd and Laurie Anderson (these last two of Siena University). I am also endebted to Cesare Zanca and Guy Aston for their technical assistance with editing, and to Peter Levy and John Morley for their meticulous proof-reading, helpful comments and general moral support. I am especially grateful to Paola Corrado for her constant encouragement and belief and for her help with the Italian material. I would also like to thank Elena Tognini-Bonelli, the series editor, for all her advice. I am most grateful to Mike Scott for his kind permission to use and discuss the WordSmith Tools program in Chapter 7. Chapter 3 is a revised version of an article published in Perspectives: Studies in Translatology, 1995: 1, but the latter sections are new. Chapter 7 is a revised version of an article in Anglais de Spécialité 7/10 (1995) but several sections are new. Chapter 8 uses some material from an article in Textus IX (1996), but is a revised, extended and updated version.

0 Introduction The main part of this book consists of a series of case studies which involve the use of corpora and corpus analysis technology. Each case study sets out with a discussion of a problem area, with reference to both the descriptive and applied literature on the argument, and then goes on to suggest ways of exploiting corpus data to shed light on the area. Emphasis is given to what can be gleaned from studying data but cannot be learnt from other resources such as dictionaries, grammars, language textbooks and so on. In many cases the evidence of language use obtained from corpora is able to refine, or even correct, the information that other resources provide. A number of the case studies (principally those reported in Chapters 3 and 5) were carried out with the collaboration of Italian mother-tongue students at the SSLMIT (Faculty for Translators and Interpreters, University of Bologna at Forlì). The particular contribution the book makes to corpus studies is that it combines language description with suggestions for pedagogical application (where these are relevant). Moreover, it includes various different types of language description, ranging from investigations into word sense, through phraseology and syntax, text studies, idiom and metaphor to creative use. It deals with meaning at a variety of levels: denotational, connotational, syntactic, textual, metaphorical and pragmatic/cultural. What is common to all the studies, however, is the regard for the importance of context in interpreting every sort of meaning. The reason why corpus analysis can enhance other resources is that it not only constitutes a new technological device, but also provides a new philosophy for language description. Until the advent of computers, language description was largely carried out by a process of introspection on the part of the linguist, and there has indeed been a certain amount of opposition to corpus linguistics, which argues that data collection can never supplant the grammarian’s intuition.1 In reality, the corpus represents both a resource against which to test such intuitions and a motor which can help to generate them. Corpus research is generally carried out in the following manner. A researcher has an intuition about language, checks this against the data the corpus provides, and this checking process frequently suggests other avenues of research to be taken, often entirely unsuspected at the start of the process (the socalled ‘‘serendipity’’principle, Higgins 1991). Intuition and data collection work hand in hand:

2

alan partington Recent corpus users have accepted that corpora, in supplying first-hand textual data, cannot be meaningfully analysed without the intuition and interpretative skills of the analyst, using knowledge of the language (qua native speaker or proficient nonnative speaker) and knowledge about the language (qua linguist). In other words, corpus use is seen as a question of corpus plus intuition, rather than of corpus or intuition.2—(Leech 1991: 74)

0.1 Corpora and linguistic description W. Nelson Francis, the co-compiler of one of the earliest and most widely used computerised corpora, created at Brown University, defines a corpus functionally as ‘‘a collection of texts assumed to be representative of a given language, dialect, or other subset of a language, to be used for linguistic analysis’’ (1982: 7). The kinds of corpora currently in existence are extremely heterogeneous, and the best way to give a general description of them is perhaps to list some of the main uses to which they are put. Below, then, is a very brief overview of some of the main areas of corpus-based linguistic analysis. • Style and authorship studies. These generally attempt to identify distinctive characteristics of a particular author’s writings (see the Chapter Stylistic Analysis and Authorship Studies in Hockey 1980). A recent development in this area is forensic linguistics which analyses written documents or transcripts in the attempt to provide evidence in legal cases of disputed authorship (Coulthard 1993, 1994). • Historical studies. A number of corpora consisting of collections of texts, of various kinds and dating from various periods in history, are currently available. Examples include the Helsinki corpus (Kytö & Rissanen 1990), which contains texts from the Old English period up to Early Modern English, classified according to regional varieties (or dialects e.g. Old Scots) and the ARCHER corpus (Biber, Finegan et al. 1994), designed for genre research, containing texts from 1650 to the present day from various social or professional fields. Diachronic linguistics compares language from different periods in time to gather information on language change (Mair 1993; Partington 1993). • Lexis. The main sort of corpus-based research into lexis conducted using corpora investigates the frequency of words and word senses in different text types or language varieties and their collocational behaviour, that is, their patterns of combination with other words. This kind of study was very difficult to perform without the aid of the large quantities of data which corpora provide. Collocation is discussed at length below. • Syntax. The patterns of combination of words in phrases, clauses or even sen-

introduction

•

•

•

•

3

tences can be investigated using corpora. Such studies have shown how a word, even each individual sense of a word, appears in typical phraseologies. The close correlation between the different senses of a word and the structures in which it appears implies that syntactic form and meaning are interdependent in the sense that each helps define the other. Sinclair (1991a: 53–65) highlights this phenomenon in an investigation of the behaviour of the word yield. G. Francis (1993) discusses lexico-syntactic (usually known as lexico-grammatical) interdependence with a number of examples. Text. The description of linguistic phenomena at levels above that of the clause has, in contrast, received relatively scant attention, largely because the nature of the technology apparently facilitates the study of discrete lexical items and sequences rather than larger stretches of language. In addition, little progress has yet been made in designing programs to annotate corpora automatically for such features as co-reference, or speech act valency. One notable recent work is Stubbs (1996), who underlines the importance in text analysis of studying grammar and lexis (with the aid of corpora), taking his cue from Halliday (1985: 345) who argues that ‘‘the study of discourse . . . cannot properly be separated from the study of grammar that lies behind it’’. He looks, for example, at how the use of passive or ergative clauses can enable writers to be vague about agency, i.e. exactly who is doing what to whom in a text. Spoken language. Corpora of spoken language have been used to study, among other things: pauses (Stenström 1990), repeats and other non-fluencies (Stenström & Svartvik 1993), hedges, back-channel responses and softeners (Altenberg 1990).3 As regards the way speakers organise their discourse, Tognini-Bonelli has studied the pragmatic function of real (1993a), how it implies the existence of an alternative concept or entity which is left unexpressed and dismissed as unimportant. Although none of the studies in this book deal specifically with spoken language, attention is given to how the study of corpora can reveal writers’ strategies for organising texts. Translation studies. A number of ‘‘equivalent corpora’’ (i.e. corpora in two or more languages containing similar text types) have been exploited to be used as a source of linguistic, semantic or pragmatic information to aid the process of translation. An example is the PIXI corpus (Gavioli & Mansfield (eds) 1990), which consists of transcripts of interactions in Italian and in British bookshops. Zanettin (1994) describes ways of designing small-scale corpora (mini-corpora) of this type, giving suggestions for their pedagogical application. For an overview of the present and prospective role of corpora and concordancing in translation studies, see Baker (1993, 1995). Register studies. Corpora can also be used for comparing, not whole languages, but different varieties of the same language, The best known work on comparing

4

alan partington

sublanguages is probably that carried out by Biber and associates on register (Biber 1988, 1989; Biber, Conrad et al. 1994) who classify text types according to six communicatively defined ‘‘dimensions’’ (‘‘involved vs. informational production’’, ‘‘narrative vs. non-narrative concerns’’, ‘‘explicit vs. situationdependent reference’’, ‘‘overt expression of persuasion’’, ‘‘abstract vs. nonabstract information’’ and ‘‘on-line informational elaboration’’), and by Nakamura (1993), who describes methods of semi-automatically classifying texts according to type (see also Nakamura and Sinclair 1995). • Lexicography. This is the area of corpus research which has had the greatest impact on English language pedagogy. The use of the Cobuild corpus at the University of Birmingham for the preparation of the Cobuild dictionary (1987) is documented in Sinclair (ed. 1987b), and all modern learners’ dictionaries follow much the same methodological principles. The sheer wealth of authentic examples that corpora provide enables dictionary compilers to have a more accurate picture of the usage, frequency and, as it were, social weight of a word or word sense. There is no ‘‘standard size’’ for corpora. Some of the research outlined above has been carried out using large, commercial corpora, such as the British National Corpus (BNC), produced by an academic and industrial consortium consisting of Oxford University Press, Longman, Chambers Harrap, Oxford and Lancaster Universities and the British Library which holds around 100 million words, and the Bank of English corpus compiled at the University of Birmingham, which contains over 200 million. Both of these corpora are commercially accessible. Corpora of spoken language tend to be smaller but the Cancode corpus, under development by Nottingham University and Cambridge University Press and the BNC each contain about 10 million words of transcribed speech. Both the BNC and the Bank of English also have sizeable sections containing transcribed speech. However, much important work has used far smaller corpora. Two of the most influential corpora, the Brown corpus (see above) and the LOB (Lancaster, OsloBergen) corpus, compiled in the early 1960s each contain one million words of written American and British English respectively. The London-Lund spoken corpus has approximately 400,000 words, but is still very widely exploited.4

0.2 Corpora and language learning A number of language researchers and teachers have constructed their own minicorpora for particular purposes. Flowerdew (1993), for instance, argues that such specially designed corpora are far more relevant to many sorts of language teaching than larger general corpora. His example is a collection of Biology lecture texts,

introduction

5

used to teach the English of this particular sector to science undergraduates, in which word and structure frequencies are radically different from those of a large corpus of ‘‘general’’ English (he contrasts his corpus with the (then) ten-million word Birmingham-Cobuild collection of texts). Stubbs & Gerbig (1993) used a small corpus, consisting of a geography textbook, to show how common language functions such as change, causation and agency are handled in particular ways, some of which are typical of geographical language and some perhaps of the particular variety or genre of ‘‘textbook English’’, but which, in any case, deserve to be brought to learners’ attention. Other researchers have used corpora as a source of classroom material and activities. Mparutsa, Love & Morrison (1991) recount their experiences of using mini-corpora of economics, geology and philosophy texts in their courses at the University of Zimbabwe. Mparutsa (op. cit.) shows how a specially-compiled corpus of only 20,000 words provides useful information on two of the most troublesome lexical problems for non-native students of economics: general words used in a special way in a particular genre (e.g. demand in economics) and the sub-technical vocabulary used to establish subject-specific concepts through the text (e.g. impacts, inputs, heterogeneous and multilateral). The studies described in this book also make use of a specially compiled collection of texts. An excellent account of what is possible using small corpora in language teaching is given in Ma (1993).

0.3 Approaches to the use of corpora in language pedagogy In the ’70s and ’80s high hopes were held in the world of CALL (ComputerAssisted Language Learning) of exploiting advances in Artificial Intelligence to produce a true teaching machine. This has not yet come about, and, while we wait for developments, attention has largely shifted to using the machine, not as a surrogate teacher, but as an informant (Johns 1991a) or a pedagogue or educational slave (Higgins 1988), in other words, as a resource of information to be tapped by the learner. This information tends to be in the form of corpora and a wealth of different corpora are now becoming available on disc or on CD-ROM. These range from corpora consisting of collections of newspaper or magazine articles (as far as British English is concerned The Independent, The Telegraph, The Guardian, The Times, The Economist and New Scientist are all available on CD-ROM, non-current year editions being fairly economical) to collections of texts from easy readers which are more appropriate for some kinds of students. There are two general approaches to using these materials in the classroom. Teachers can either analyse the corpora themselves for material design or they can decide to introduce them into the classroom and train students in their use. In the first case:

alan partington

6

teachers might use corpus-based investigations to (i) determine the most frequent patterns in a particular domain; (ii) enrich their knowledge of the language, perhaps in response to questions raised in the classroom; (iii) provide ‘‘authentic data’’ examples; and (iv) generate teaching materials.—(Barlow 1996: 30)

In the second: teachers may also wish to have their students explore corpus materials, either in following a path of investigation determined by the teacher (so that the students come to understand a particular pattern of usage such as say versus tell or the collocations of bright) or in exploring an issue in a more open-ended way.—(Barlow 1996: 30)

Johns (1991a, 1991b) has described a number of ways teachers can create materials and exercises for use in the classroom, and he has also developed a theory of the place of what he calls ‘‘data-driven learning’’ (DDL) in the pedagogical universe. The essence of this kind of learning is the inductive acquisition on the part of students of grammatical rules or regularities through the process of analysing the patterns of language use of specially selected items as revealed through corpora (Tribble & Jones 1990; Johns 1991a). Johns outlines three general effects of adopting the DDL approach: [It] can have a considerable influence on the process of language learning, stimulating enquiry and speculation on the part of the learner, and helping the learner also to develop the ability to see patterning in the target language and to form generalisations to account for that patterning. The second main effect of DDL is on the role of the teacher, who has to learn to become a director and co-ordinator of student-initiated research [. . .] The third [. . .] is a revaluation of the place of grammar in language-learning and language teaching [. . .] The DDL approach [. . .] makes possible a new style of ‘‘grammatical consciousness-raising’’ (Rutherford 1987) by placing the learner’s own discovery of grammar at the centre of language learning.—(Johns 1991a: 2–3)

One of the advantages of this methodology is that the learners’ role becomes very active; they become a sort of language researcher: What distinguishes the data-driven learning approach is the attempt to cut out the middleman as far as possible and give direct access to the data so that the learner can take part in building up his or her own profiles of meaning and uses. The assumption that underlies this approach is that effective language learning is itself a form of linguistic research, and [. . .] offers a unique resource for the stimulation of inductive learning strategies—in particular the strategies of perceiving similarities and differences and of hypothesis formation and testing.—(Johns 1991b: 30–1)

introduction

7

This is perhaps a rather strong statement of the theory. In fact, as Johns explains elsewhere, there is still an element of the ‘‘middleman’’ inasmuch as the language data presented to the students is often mediated by the fact that the teacher has both chosen the object of research and edited the data to remove untypical or unrevealing examples. A pure example of an unmediated approach to corpus-based learning—of empowering students as researchers—is described by Jordan (1993). He outlines experiments in which computers were made available to students so that learners could interrogate corpora on their own, with practice worksheets designed both to familiarise them with the program and to suggest types of lexico-grammatical research to be carried out (e.g. comparing the contexts of for and since, of hope and expect, of yet, still and already). The role of the teacher in this case has become that of ‘‘co-ordinator of student-initiated research’’ (1993: 22). Finally, Aston suggests that approaches to using corpus resources in the classroom can be divided into two general areas, those for reference and those for browsing: On the one hand, they might be treated as a reference tool, which could be looked up to provide examples and therefore clarify doubts on particular problems which had arisen in other language activities. From this perspective the corpus could be seen as complementing the grammar, the dictionary, and the encyclopaedia. On the other hand, a corpus might be treated as a source of activity in itself, a hypertext to be browsed in, where the user passes from one text or concordance to another, and where, rather than being determined by a preselected goal, progressive discoveries occur on a negotiated step-by-step basis.—(Aston 1997)

From the pedagogical point of view, the studies reported in this book are offered as a demonstration of how it is possible to interrogate a medium-sized corpus for purposes such as those described above. In each case study described, the area involved seems one in which at least one of Barlow’s goals (see quotations above) can be met; the teacher-researcher’s awareness of the language can be enriched, syllabuses and materials can be refined and developed, and learners can explore data as a means of developing their study skills and understanding of the language.

0.4 Guide to the contents of this book Chapter 1 contains a definition of concepts important in the studies, particularly that of collocation, and outlines the linguistic philosophy which informs them. The next three chapters report studies into various aspects of word and phrase sense. Chapter 2 describes how corpus analysis can be used by learners and translators

8

alan partington

to choose the best item among a group of near synonyms for a particular context. Dictionaries and thesauruses are often not capable of supplying information which is detailed enough to allow this. Chapter 3 contains a somewhat similar study, but with still more relevance to translation practice. It shows how corpus analysis can be employed to study the phenomenon of ‘‘false friendship’’, of determining when seeming translation equivalents are ‘‘false friends’’ and when they are more reliable. Chapter 4 is an investigation into what Louw has called ‘‘semantic prosodies’’, that is, into how some items appear preferentially with words of a particular type— e.g. the word dealings was found to show a strong tendency to occur with negative items (with shady, corrupt, etc. and expressions such as allegations of, investigations into), whereas ‘‘synonyms’’ such as deals or relations do not. This information is vital for non-native speakers to understand not only what is grammatically possible in their language production but, in Hymes’ (1971) terms, also what is appropriate and what actually happens. However, an awareness of semantic prosodies can also be extremely useful for native and non-native alike in uncovering a text producer’s hidden attitudes. In Chapter 5 the focus is extended beyond the phrase. As an example of how looking at lexis can provide information at various language levels, from the syntactic to the discoursal, an analysis of if constructions is described. Of the circa 400 if constructions examined, less than 40% were in the form of any of the famous three conditionals dear to EFL textbooks and didactic grammars. A number of possible explanations based on corpus evidence are suggested. In Chapter 6 the focus is enlarged still further to look at textual phenomena—in particular, text reference. General nouns and anaphoric nouns are studied and the results compared to the statements made about them by previous authors. Also examined is the behaviour of a class of little-studied items, what the present author has called ‘‘general verbs’’, including happen and occur. These items are used by writers to refer to extended stretches of text, both preceding and following, when this passage of text is describing a process or event. Chapter 7 discusses metaphor. Given the nature of the corpus, it is possible to examine single genres and compare them to others in order to discover which sets of metaphor (or ‘‘systematic metaphors’’: Lakoff & Johnson 1980) are typical of which genres. Particular attention is paid to business and financial journalism. In the course of this analysis, some light was thrown on the controversy between conflicting linguistic views of the nature of metaphor itself. In Chapter 8 attention is turned to the creative use of language. A sub-corpus of newspaper headlines was compiled in order to study what the present author terms ‘‘unusuality’’, that is, the conscious attempt of a text producer to upset and somehow exploit the normal phraseological patterns of proverbs, sayings, quotations and

introduction

9

so on. The main part of the chapter contains a discussion of numerous examples of the phenomenon from headlines. In the final part an attempt is made to explain why it happens in terms of wider linguistic theory.

0.5 Equipment and methodology Most of the linguistic analysis performed using computerised corpora is born out of a statistical methodological philosophy, the search for—and belief in the importance of—recurring patterns. It is based on the twin concepts of frequency (a factor of (past) observation) and probability (a factor of (future) predictability). In other words, if something is seen to happen frequently in a language, then it is significant. It is significant precisely because this frequent occurrence, or regularity, can be used as the basis for predicting how other, as yet unanalysed, chunks of language will behave, and, in the end, for hypothesising a description of how the entire universe of discourse under study (whether it be a register, dialect or other subset of language) is constructed. Very often these language patterns are not immediately obvious in the course of simple introspection, but they can become more apparent through the medium of the concordance. A concordance, or rather KWIC (KeyWord In Context) concordance, is a list of unconnected lines of text, which have been summoned by the concordance program from a computer corpus, that is a collection of texts held in a form which is accessible to the computer. At the centre of each line is the item being studied (keyword or node). The rest of each line contains the immediate co-text to the left and right of the keyword. Such a list enables the analyst to look for eventual patterns in the surrounding co-text, which proffer clues to the use of the keyword item. The search for such clues is the basis of the methodology in the case studies reported in the following chapters, which contain numerous examples of concordance lists. The following is a concordance of the word stark, which shows how it appears regularly before certain nouns: 1 ancellor of the Exchequer, faces a stark choice between a penny off tax or 2

entire career has revolved around stark choices and moments of daring brav

3 the mid1990s. The prediction is in stark contrast to Mr Horton's belief tha 4 lmost all the Muslim conquests, in stark contrast to the failure of medieva 5 ntribution to 1930s housing, is in stark contrast to the cottagey suburban 6 d an unbridled aggression, were in stark contrast to the shortcomings of St 7 ntribution to 1930s housing, is in stark contrast to the cottagey suburban 8 ccess as scrumhalf and captain, in stark contrast to his first taste of the 9 e management. The profits were `in stark contrast to others in the construc

10

alan partington

10 gestures ( nutus ). This stark contrast is not invention, but it 11 ow had enough''.

It is all in stark contrast to the two pages Neues De 12 but fell back to close at 47p _ a stark contrast to the year's high of 234 13

the change was made. It seems in stark contrast to the crackdown on radic

14 ormal living conditions''.

In stark contrast, yesterday's statement fr 15 ednesday, their back line in such stark disarray, so frequently outsmarted 16 g to prevent an arrest. This is a stark example of constructive criminal l 17 ore importantly, his productions' stark expression of human vanity perhaps 18 the stock market is now providing stark illustration of how equities and c 19 he changes are an eloquently stark illustration of the massive reorie 20

bears the burden.

From these stark images he switches style completel

21 arge expanses of solid colour and stark images, often depicted in black. T 22 can police Colonel who is affably stark mad.

There is not room for all 23 egan stripping again and was seen stark naked running in and out of traffi 24

ANTHONY MOULD Ishbel Myerscough. Stark nudes and portraits. 173 New Bond

25 ugh not everything that blurs the stark opposition of the one and the six 26 ll of this doesn't mean he wasn't stark raving mad, and just putting on. B 27 cocked hat. All this happens on a stark set, bare but for two leather armc 28 anner, in Sarajevo, witnessed the stark transformation of a bustling commu 29 under of Scotland United, takes a stark view: `We can take Westminster, or

Most of the pieces of research reported below were carried out using MicroConcord from Oxford University Press (Scott & Johns 1993; Murison-Bowie 1993) to prepare concordances of various lexical items contained in a corpus of English compiled by the author of this volume, along with a number of colleagues, at the SSLMIT, University of Bologna at Forlì. This corpus contains about five million words of running text. The MicroConcord system can automatically sort the concordance lines by alphabetical order of the first word to the left, or alternatively the first word to the right, of the keyword. This means that frequent collocational patterns are apparent. In general (assuming the language being studied is English), if the keyword is, for example, a noun, and the user wishes to know what adjectives it collocates with, then they will ask the system to sort the lines in alphabetical order of the word to the left of the keyword (to ‘‘left-sort’’): 1 ke the threat of youth to stir ageing bones. With the England selectors due 2 an school cites evidence from ancient bones in Europe and North Africa to s 3 aptor Stuart Paterson offers the bare bones of Zola's story of illicit love 4 vox and the like. Relying on the bare bones of the semiacoustic guitar, sna 5 92 / RECORDS / The sounds of breaking bones: Andy Gill reviews the week's n 6 t hospital on Kauai, most with broken bones and one with a heart attack. Th 7

d to get up. Mr King suffered broken bones around his skull, a fractured c

introduction

11

8 nstruction worker, suffered 11 broken bones in the base of his skull, a bro 9 ernal injuries and a number of broken bones. He was described as `comfortab 10 k, vertebrae, feet, kneecaps, collar bones _ bar his legs. But permanent i 11 motional flesh on these diagrammatic bones. It does convey, though, that B 12

that until now had seemed to be dry bones. This was the Vivaldi whom Bach

13 ched tightly over his prominent face bones. The bonus seemed to hold an il 14

tainted by Western decadence. Human bones and scraps of clothing are stil

15 after police found hundreds of human bones scattered around a property in 16 e's plenty of life left in those old bones yet. Jane Chapman, with Django 17 range satisfaction at making his old bones happy. Even the parents had sto 18

amounts can be obtained if the soft bones in some varieties of fish (like

Thus, in the concordance from my corpus, the noun bones is preceded by these adjectives: ageing, ancient, bare, breaking, broken, collar, diagrammatic, dry, face, human, old and soft. The words bare, broken, human and old all occur more than once after the keyword, and by alphabetical left-sorting these occurrences appear next to each other in the concordance list. This proximity makes it easier for the user to see which collocations are the most frequent in their concordance. If, on the other hand, the user is studying not nouns but, say, a set of adverbs and wishes to know what adjectives follow them, they will sort by the word to the right (to ‘‘rightsort’’). Thus, in the concordance from my corpus, the adverb mightily is followed by these adjectives: complex, constrained, displeased, feared and suspect: 1

crowds to watch players striving mightily against one another than to watc

2 tra's fine treatment of what is a mightily complex score, and by Margaret 3 ition. plainly the wartime legacy mightily constrained the policy options 4 a major attraction, are otherwise mightily displeased. A baleful corporate 5 e is Khaled again, still grinning mightily, exuding bonhomie and energy hal 6 re that Ross Perot's return, once mightily feared by the Democrats, will ma 7

and his cause, which he advanced mightily, securing for it a permanent and

8 n the guilty?' is a recurring and mightily suspect phrase) is mixed in with 9 much- criticised defenders rising mightily to the occasion. While the Bills 10 f, a characteristic that appeals mightily to the Celtic nature.

Last w

The system also allows the user to arrange the lines according to categories of grammar or sense e.g. to separate bank as noun, from bank as a verb, or bank as a place where money is kept from bank as the side of a river. This sorting must be done manually. Chapter 2 contains examples of lists sorted this way. The program also has a ‘‘View’’ facility which enables the user to see the page of text from which the line derives and a further facility allows him or her to scroll earlier or later pages of the text. These facilities were useful in a number of the case studies, since the amount of co-text available for each single citation of the keyword

12

alan partington

available on the screen in the original concordance is often not sufficient to understand how it is being used. There are also a couple of ‘‘wildcard’’ facilities. The first is called by typing an asterisk which stands for a part or the whole of a word. Thus, if a user types ‘‘take * place’’, the concordance will include any occurrences in the corpus of take a place, take the place, take my/your/his etc place as well as take place. This facility is extremely useful if a user wishes to see all occurrences of a word or phrase irrespective of inflections. For instance, if a user types ‘‘call* off’’, the concordance will include any occurrences of call off, calls off, called off and calling off. The second wildcard is called by typing a question mark which stands for a single letter. Thus, by typing ‘‘eye?’’, the user will obtain a concordance of eye and eyes. Concordances can also be saved onto hard or floppy disk. Most concordance programs will ask the user to stipulate how many characters of text they wish to save for each concordance line. Around 100 characters is normal if one wishes each concordance ‘‘line’’ to fit on a single line of an A4 sheet, but more, say 200-250, will be necessary if one wishes to capture the whole sentence the keyword appears in. Once saved, the concordance can be accessed using a word processor (this author used Word for Windows) simply by calling up the filename the concordance has been saved under. In this way the concordance lines can be studied and, if necessary, edited using the Word Processing facilities. The corpus itself was compiled as follows. OUP provide in the MicroConcord package a couple of floppy disks, one containing a million words of articles from The Independent newspaper and the other a million words of text consisting of a variety of academic writing. These were transferred onto the hard disk of the computer, following the instructions provided. A further three million words of text were transferred from the commercially available CD-ROM containing all the articles printed in The Independent and The Independent on Sunday in the year 1992. The software which runs The Independent CD-ROM provides commands which permit the user to save batches of articles to either hard or floppy disk. Batches of about 50,000 words of text were saved. There is no particular reason why this size should be chosen. However, very long files can be cumbersome when called up by a word processor, which is a useful procedure when contextual information about a text is sought. It was also necessary to codify the files that were saved so that those containing roughly similar types of texts could be accessed as a whole by the concordancer, and others containing different types could be excluded. This is vital if the aim of research is to compare different genres (see Chapters 5 and 7). The concordancer software allows the user to access material by filename, and so each file was given a name which included a suffix indicating the kind of texts it contained. Taking the newspaper files, the suffixes are: HOM for home News, FOR for International or

introduction

13

Foreign news, ART for Arts and Features, BUS for Business and Finance, SPO for Sports news. These categories coincide with those used by The Independent itself for classifying its contents. The categories of the academic section of the corpus are: SCI for Scientific (actually popular science) writings, MED for Medical texts, BEL (belief) for Philosophy and Theology, ARS for the Arts and SOC for Society (i.e. texts in sociology). In each of the case studies, an indication is always given of which sections of the corpus—newspaper, academic or combined—were interrogated. A ‘‘sister’’ newspaper corpus of Italian was also compiled by the same methods. Two million words were downloaded and saved onto hard disk from the CD-ROM of Il Sole 24 Ore, an Italian quality newspaper. This was the data source used in the comparative study reported in Chapter 3. For some of the case studies, examples of the use of lexical items were also taken from the CD-ROMs of The Independent of the year 1992 and The Telegraph (Daily and Sunday) and The Times (Daily and Sunday) both of 1993. Each of these contains considerably more text than the corpus and was therefore more likely to contain occurrences of very rare items and so could be used to test corpus findings. MicroConcord can access CD-ROM material directly if this is in the form of text files, but it is often more convenient to make a collection of all the newspaper articles containing the item to be analysed (using the editing facilities provided by the CD-ROM software) and then save them into a file on disk. This file can then be accessed by MicroConcord to produce a concordance list. This was the method generally used in these studies. Newspaper texts clearly constitute the lion’s share of the corpus. It may be argued that they are hardly representative of language as a whole. On the other hand, as Biber has convincingly shown in his studies of register and text types (Biber 1988; Biber, Conrad et al. 1994), there is no such thing as ‘‘English as a whole’’. All language production belongs to one genre or another, and the English language, like any other, is a collection of genres, none of which deserves the title of ‘‘general English’’ more than any other. If this is so, then newspaper texts will serve as well as any as the basis of linguistic investigation. They also have a few particular advantages over other text types. They are the most widely read of long texts—almost everyone has considerable experience of newspaper language. Moreover, it must be remembered that newspapers consist of not one but a large number of different text types and, in fact, the newspaper section of the corpus is divided into five sections, or mini-corpora as described above (home news, arts and features and so on). As is well known, different kinds of newspaper texts are prepared in different ways (see Bell (1991) for a detailed analysis of how newspapers are produced). For example, many kinds of hard news articles (including front-page news, sports match commentaries, news flashes) are prepared in a very short time. Cowie (1992)

14

alan partington

suggests that the tight constraints on this kind of writing might mean that it has certain features in common with spoken language. Opinion articles may have less rigorous time limits, whilst other texts, e.g. soft news human-interest stories or pieces on general culture may take quite a time to complete and are stored in a newspaper’s library to be used when space allows. These differences are all reflected in the kind of language the various types of article contain.

Notes 1. See W. N. Francis’ discussion (1982: 7–8) of his personal experience. 2. A recent dispute between Owen (1993) and G. Francis & Sinclair (1994) of the Cobuild corpus research team highlights this point well. Owen writes that corpus-grammarians claim to eschew intuition since, in theory, it ‘‘plays no part in the compilation process since categories will emerge from the data’’ but that in practice, they ‘‘take a few short cuts’’ (1993: 179). G. Francis & Sinclair answer that there is no need to choose between intuition and authenticity: ‘‘It is not necessary [. . .] to abjure one or other kind of evidence. Certainly in Cobuild there is a heavy reliance on intuition [. . .] On occasions when someone’s intuition does not line up with the corpus evidence, there is a problem to be resolved’’ (1994: 191). 3. These last three were found to compose 9.4 per cent of all word-class tokens in a spoken corpus (London-Lund), which made them more frequent than prepositions, adverbs, determiners, conjunctions or adjectives. 4. Brown, LOB and the London-Lund corpora are available for research purposes on CD-ROM from ICAME (International Computer Archive of Modern English, The Norwegian Computing Centre for the Humanities, Bergen, Norway). The BNC is available for similar purposes on CD-ROM for Unix and accessible over the Internet (http://info.ox.ac.uk/bnc). Cobuild (Bank of English) provides a lookup service by anonymous telnet (titania.cobuild.collins.co.uk).

1 Collocation and Phrase Patterns . . . sont des mots qui vont très bien ensemble . . . (Beatles song).

1.1 Definitions of collocation The term ‘‘collocation’’, as is well known, was first coined in its modern linguistic sense by the British linguist J.R. Firth, along with the famous explanatory slogan: ‘‘you shall judge a word by the company it keeps’’. In an article entitled Modes of meaning (Firth 1957), he outlines how the study of ‘‘meaning by collocation’’ can contribute to a formal and contextual, as opposed to conceptual, approach to wordmeaning: Meaning by collocation is an abstraction at the syntagmatic level and is not directly concerned with the conceptual or idea approach to the meaning of words. One of the meanings of night is its collocability with dark, and, of dark, of course, collocation with night.—(Firth 1957: 196)

Later writers on collocation have picked up different aspects of Firth’s ideas. A number of different but related definitions of ‘‘collocation’’ have been forthcoming. Sinclair, who was a student of Firth’s at London University, sees it as follows: Collocation is the occurrence of two or more words within a short space of each other in a text.—(Sinclair 1991a: 170)

We might call this the ‘‘textual’’ definition. One item collocates with another if it appears somewhere near it in a given text. Under this definition, it is a consequence of the linearity of language, or, conversely, if we view text as a process rather than product, it is the principal method, together with syntax, with which this linearity is constructed. Another definition is given by Leech in his discussion of ‘‘Seven Types of Meaning’’, one of which is ‘‘collocative meaning’’:

16

alan partington Collocative meaning consists of the associations a word acquires on account of the meanings of words which tend to occur in its environment.—(Leech 1974: 20)

This could be referred to as the ‘‘psychological’’ or ‘‘associative’’ definition. It is part of a native speaker’s communicative competence (see below) to know what are normal and what are unusual collocations in given circumstances. Through lifelong exposure to a language, native speakers acquire what Firth calls ‘‘expectancies’’ (1957: 195) of which items commonly co-occur with which others in texts. The contribution of collocation, in psychological terms, to meaning is also emphasised by Aitchison, who says that ‘‘humans learn word-meaning from what occurs alongside’’ (1994: 21). The learner, child or adult, faced with an unknown word looks to the co-text to gain clues as to what the unfamiliar item might mean. Meaning is function in context, as Firth used to say. Finally, Hoey highlights another aspect of the concept: collocation has long been the name given to the relationship a lexical item has with items that appear with greater than random probability in its (textual) context. (Hoey 1991: 6–7)

We might call this the ‘‘statistical’’ definition of collocation. It is a good working definition of the concept for those studying corpus linguistics, where large quantities of text can be made available for computer analysis. The co-occurrence of two items becomes interesting if it seems to happen for a purpose, and especially if it is repeated, if there are ‘‘patterns of collocation’’. The habitual associations of a word with other items can thus be studied both by calling up concordances of that word (see previous section) and by obtaining lists of its most frequent collocates. Firth himself gives the example of the word time whose common collocates include saved, spent, wasted, frittered away and also presses and flies and even the word no. By these means, it is possible to build up a picture of the way in which each and every lexical item in a language behaves. The majority of modern learners’ dictionaries and grammars avail themselves of computer facilities for lexical analysis and include information on collocation. ‘‘Collocation’’ is usually used to refer to the co-occurrence of two single words. However, Firth himself originally used the term to refer to the co-occurrence of items at all grammatical levels, not just the word level. Even patterns such as assonance and alliteration were subsumed under the category of collocation. In fact, many of Firth’s examples in his discussion of the phenomenon are taken from poetry, including limericks, and Carroll’s nonsense verse. If we concentrate on higher levels of language, there is theoretically no qualitative difference between word with word, word with phrase, phrase with phrase, even phrase with clause and clause

collocation and phrase patterns

17

with clause collocation. All can be studied as phenomena of patterning in text, a field which is beginning to be referred to as ‘‘Combinatorics’’. This wider definition of ‘‘collocation’’ informs the case studies in this book.

1.2 Collocation, text type and style It is not difficult to see how the textual and psychological definitions of collocation, cited above, are connected. Their interrelation makes it possible for a text receiver to judge whether a particular collocation in a particular text is usual or unusual. However, it must be stressed that there is a distinction between the statistical information about ‘‘the language as a whole’’ and the individual’s psychological knowledge of what constitutes normal collocation. Collocational normality is dependent on genre, register and style i.e. what is normal in one kind of text may be quite unusual in another. Firth makes this point when he talks of ‘‘general or more usual collocations’’ as opposed to ‘‘more restricted technical or personal collocations’’ (1957: 195). A large contribution to the study of both literature and language as a whole can be made by ‘‘the study of the usual collocations of a literary form or genre or of a particular author’’ (1957: 195). For a non-literary illustration we can return to Firth’s example, cited above, of the word time. All the collocates he gives of this word (saved, spent, wasted, frittered away etc.) were found to be far less common in the sub-corpus of sports journalism than, say, half, full, extra, injury as well as the first, the second, the third and so on. These latter items can thus be said to be normal collocations of time in this particular text type. Sinclair (1966) also stresses the relationship between collocation and register. He points out how collocations such as vigorous depressions and dull highlights may seem odd out of context but that, placed in their register-specific habitats of, respectively, meteorology and photography, they are quite normal. Moreover, norms of collocation can be exploited and upset for purposes of style. In very general terms, as McIntosh (1966: 193) points out, there are the following four distinct stylistic modes: normal collocations and normal grammar, unusual collocations and normal grammar, normal collocations and unusual grammar, unusual collocations and unusual grammar. As Carter notes (1987: 54), producing language at the two extremes of the scale is risky. Too much respect for normal collocation and grammar produces language which is ‘‘too familiar and thus banal (e.g. ‘This is guaranteed to meet your special requirements’)’’. Too little respect and there is the risk of producing texts which are ‘‘unfamiliar and thus indecipherable (‘The ants with and swore the bald-headed carpet sweeper’)’’. Fortunately ‘‘between the two extremes is a dimension in which more individual and creative effects can be produced’’ (Carter 1987: 54). The last part of this work will be

18

alan partington

dedicated to a study of some of the effects that the use of unusual collocation can produce.

1.3 Collocation and communicative competence The knowledge of which collocations are normal in which environments is, as has been mentioned, part of a native speaker’s communicative competence, as defined by Hymes (1971). Hymes’ ideas arise from a dissatisfaction with Chomsky’s (1965) binary division of language use into competence and performance and a conviction that competence covers a much wider range of skills and knowledge than the internalisation of the grammatical system as in Chomsky’s definition. Hymes divides competence into four distinct kinds: the knowledge of what is formally possible given the constraints of the language system, of what is feasible, of what is appropriate and of what is actually performed i.e. what normally occurs.1 The first of these corresponds to Chomskian competence, whereas all the other types of competence are context-related, that is, they are knowledge systems which permit language users to make certain decisions about language use depending on the situation they find themselves in and the requirements they have. As an analogy, these four kinds of competence can be seen as an ordered set of refining mechanisms which enable the speaker or writer to make finer and finer choices among the infinity of possibilities made available for the purpose of communication by a language. Collocation choices are made at the latter stages of the refining process. Knowledge of what is normal collocation in a particular environment can thus be seen to be an important part of the fourth of these aspects of competence, knowledge of what actually occurs. But, at the same time, if a speaker chooses to use a collocation which is unusual in a particular circumstance, for surprise, dramatic or humorous effect, he or she makes a choice at one of the ‘‘higher’’ levels of refinement, those relating to what is feasible and what is appropriate. However, as we shall see in the final chapter of this volume, most collocational decisions made at the ‘‘higher’’ levels of competence, are still made with respect to what normally occurs. Thus, to take an example from literature, Dylan Thomas’ poetic inventions ‘‘All the sun long’’ and ‘‘All the moon long’’ depend for their effect on the more usual all day long and all night long. Moreover, there is no total agreement even among native speakers as to which collocations are acceptable and which are not. In the field of collocation, as Carter (1987: 55) remarks, ‘‘questions of acceptability are much more difficult to determine than the decision over what is grammatical or ungrammatical’’. He suggests using ‘‘techniques of informant analysis in which the intersubjective intuitions of a group of native-language speakers are statistically measured and a line drawn

collocation and phrase patterns

19

between what can generally be allowed and what can not’’ (Carter 1987: 55). However, in the light of the dependency of collocational acceptability on register and style, such a demarcation would have to be conducted with great care, making sure enough context was supplied to render informants’ decisions meaningful.

1.4 The ‘‘idiom’’ or ‘‘collocational’’ principle A number of authors, including Bolinger (1976), Pawley & Syder (1983), Sinclair (1987a, 1992) and Nattinger & DeCarrico (1989, 1992) see collocation as one of the two main organising features of text. The other, what Sinclair calls the ‘‘open choice principle’’, sees language production as a continuous series of open-ended choices, ‘‘a series of slots which have to be filled from a lexicon’’ (Sinclair 1987a: 320), the only restraints being grammatical i.e. that only items from certain word classes may appear in a given slot. However, in practice, we find that possible slot choices are massively reduced. Constraints other than the grammatical are in play in texts, as Pawley & Syder note: native speakers do not exercise the creative potential of syntactic rules to anything like their full extent [. . .] indeed, if they did so they would not be accepted as exhibiting nativelike control of the language. The fact is that only a small proportion of the total set of grammatical sentences are nativelike in form [. . .] in contrast to expressions that are grammatical but are judged to be ‘‘unidiomatic’’, ‘‘odd’’ or ‘‘foreignisms’’.—(Pawley & Syder 1983: 193)

In Sinclair’s terms, these further constraints express the ‘‘idiom’’ or ‘‘collocational principle’’. This principle states that openness of choice is not available to the same extent at every point along the syntagmatic progression of an utterance, but that ‘‘the language user has available to him a large number of preconstructed or semi-preconstructed phrases that constitute single choices, even though they appear to be analysable into segments’’ (Sinclair 1987a: 320). Bolinger (1976) calls these chunks of language ‘‘prefabrications’’ or ‘‘prefabs’’. It is by no means clear, at first sight, why all this should be so. If it is the case that the language user needs to store in his memory both the single vocabulary items and a lexicon of (semi-) preconstructed phrases, the number of which is probably in excess of the number of vocabulary items used productively by the individual, then it would seem counterproductive to store linguistic information in this way. The reason why this happens must be that the effort saved by making far fewer slotfilling choices in real-time—i.e. in the process of communication—easily outweighs the disadvantages of unwieldy storage. As Ladefoged has said:

20

alan partington The indications from neurophysiology and psychology are that, instead of storing a small number of primitives and organizing them in terms of a (relatively) large number of rules, we store a large number of complex items which we manipulate with comparatively simple operations. The central nervous system is like a special kind of computer which has rapid access to items in a very large memory, but comparatively little ability to process these items when they have been taken out of memory. (Ladefoged 1972: 282)

More recently, Nattinger & DeCarrico too have noted that: In formulating performance models of language processing, researchers endeavour to offer direct descriptions of psychological categories and processes, attempting to describe languages in terms of how they are perceived, stored, remembered and produced. These researchers feel that the storage capacity of memory is vast, but that the speed for processing those memories is not (Crick 1979: 219), so that we must learn short cuts for making efficient use of this processing time.—(Nattinger & DeCarrico 1992: 31)

Principal among these short cuts is the use of prefabs. A still more powerful reason for the employment of semi-preconstructed phrases probably lies in the way it facilitates communication processing on the part of the hearer. Language consisting of a relatively high number of fixed phrases is generally more predictable than that which is not (Oller & Streiff 1975). In real-time language decoding, hearers need all the help they can get. Redundancy in communication is often explained in this way and the collocational principle probably has the same functional origins. If the raison d’être of the idiom-collocation principle is to save processing time and effort, then it would tend to be most typical of on-line, spontaneous discourses, that is to say, conversation. We might expect preconstructed and semi-preconstructed phrases to be less common in writing, where time constraints tend to be less rigid. As Nattinger & DeCarrico put it, ‘‘at first, it would seem that the premeditated quality of written language might preclude any prefabricated lexical phrases at all’’ (1992: 81). However, in very many genres of writing, pre-cooked expressions are still diagnostic, vital elements. We need only think of legal documents, scientific/medical papers, business reports and so on. There are two reasons for this. Many kinds of lexical items, including prefabs, function as powerful indicators of register, and in most circumstances it is important for a writer to signal the register to which the text belongs. Secondly, although it may be less pressing than in conversation, there is still a need in written texts to balance new information with old information, novelty with habit, (prefabs contributing to the second items of these pairs) to cut down processing effort, especially if the text is long.2 ‘‘Written language, just as spoken, draws from a stock of ready-made phrases,’’ Nattinger &

collocation and phrase patterns

21

DeCarrico (1992: 81) conclude, ‘‘which exist in a continuum from the entirely fixed to the more variable’’. Discourses, then, are created by a combination of the idiom and open choice principles. Language tends to be chosen in blocks from the mental lexicon, but these passages are still linked together, they still have a syntagmatic relation to each other and to grammatical items. The two principles work together in any text. Nor is it always easy to know, at any point in a text, which of the two principles is operating most strongly. As Bolinger points out: in language it is so hard to be sure whether we are dealing with something freely and freshly constructed from its least elements or something assembled from rather large chunks consigned to us whole. It seems the brain stores both the parts and the wholes, and we retrieve them when we need them [. . .]—(Bolinger 1975: 56).

There is nothing new in drawing attention to the existence of multi-word units (Zgusta 1987, Baker & McCarthy 1990). ‘‘Every linguist makes room in his scheme of things for lexical units larger than words,’’ Bolinger writes. ‘‘He calls them idioms’’. What is new is that: we are now in a position to recognize that idiomaticity is a vastly more pervasive phenomenon than we ever imagined, and vastly harder to separate from the pure freedom of syntax, if indeed any such fiery zone as pure syntax exists.— (Bolinger 1976: 3)

Clearly Bolinger is using ‘‘idiomaticity’’ to refer to a wider phenomenon than many other linguists, who tend to reserve it to indicate items with some metaphorical content (though we can square the two definitions if we bear in mind Barthes’ famous statement that all language is dead metaphor). To avoid confusion, in this book, the terms ‘‘collocation’’ and ‘‘collocation principle’’ are employed.

1.5 Collocation and linguistic schemas (or schemata) Highly relevant to the present discussion are the studies carried out by Pawley & Syder (1983) on the concept of lexicalized sentence stems and by Barlow & Kemmer (Barlow & Kemmer 1994; Barlow 1996) on the similar notion of the linguistic schema, both of which explain the existence and functions in natural language of semi-preconstructed phrases. Barlow, for instance, finds purely syntactic descriptions of language, especially those of the generative schools, inadequate to explain how language works:

22

alan partington we wish to be able to describe the collocations that are found in language. Some words simply like to occur near each other, a fact that is unaccounted for within the generativist paradigm, which has focussed on creativity in the sense of a free association of lexical items, governed only by general constraints of grammar.—(Barlow 1996: 15)

He also emphasises that collocation is no peripheral phenomenon, suggesting in fact that ‘‘most of language consists of semi-regular, semi-fixed phrases or units’’, a fact which the standard generative schools cannot cope with since these semi-fixed phrases ‘‘do not fit in particularly well in the lexicon (because the units are too large and contain too much that is regular)’’ whilst they also ‘‘do not fit in the syntax (because they are in some sense a unit [. . .])’’ (Barlow 1996: 15). A lexicalised sentence stem consists of a sentence or part of a sentence in which one or more of the structural elements is a class rather than a particular discrete item. Pawley & Syder illustrate the notion with the example of the conventional expressions of apology: I’m sorry to keep you waiting. I’m so sorry to have kept you waiting. Mr X is sorry to keep you waiting all this time.—(Pawley & Syder 1983: 210)

These are variations on the grammatical frame: NP be-TENSE sorry to keep-TENSE you waiting. Such a frame is a sentence stem, whilst its various realisations are termed inflections. Any eventual additional components (as ‘‘all this time’’ above) are called expansions. A schema is the form component of a ‘‘form-meaning pairing’’ which shares some of the qualities of a fixed phrase but which also contains variable parts capable of capturing context dependent information. Barlow (1996: 15) cites the example to let oneself go in which ‘‘the individual components have a recognisable meaning, but the expression as a whole means more than the sum of its parts’’. In to let oneself go there are two variables: the verb let which changes in agreement with its subject and the next word in the phrase which can be any item from the closed set of reflexives. But this variability is contained within an idiomatic expression-frame which has the structure ‘‘let + reflexive + go’’. This frame is the schema. To lose one’s way is another example, which has the structure ‘‘lose + possessive + way’’. Again lexical and phrasal principles are combined in the same units. The fixed parts of the structure define the general meaning of the phrase, which is different from the

collocation and phrase patterns

23

meaning of the individual lexical items, whilst the variable parts link the meaning to the context of situation (that is, who exactly is ‘‘letting her/him themselves etc go’’ or is ‘‘losing his/her/their etc way’’). The term ‘‘schemas’’ (or ‘‘schemata’’) is borrowed from psychology and artificial intelligence (roughly similar concepts are also called ‘‘frames’’ or ‘‘scripts’’) where it describes ways in which the mind organises experience and acquires new learning. As Brown & Yule say: Schemata are said to be ‘‘high-level complex (and even conventional or habitual) knowledge structures’’ (van Dijk 1981: 141), which function as ‘‘ideational scaffolding’’ (Anderson 1977) in the organisation and interpretation of experience.—(Brown & Yule 1983: 247)

A schema is built up in the mind by repeated experiences of situations of a similar type, say ‘‘going to a restaurant’’, which then become general background knowledge and also serve to supply general expectations of what can happen and how people should behave in that situation. But they are never deterministic, of course, there is always room for some variation. A linguistic schema is constructed in an analogous way, by the brain being exposed to many instances of a particular language structure, which is stored as a whole and becomes a model for production, but which also contains knowledge of which elements are variable according to context.

1.6 Routines and patterns in language acquisition studies There is, then, a psycholinguistic explanation for the prevalence of preconstructed and semi-preconstructed phrases in language. Their existence reflects the way language is acquired by the human brain and a number of studies have thrown light on the role they play in both first and second language acquisition. Krashen & Scarcella (1978) noted the presence of ‘‘formulaic’’ speech in learners’ utterances and distinguished between routines, or completely fixed chunks of language (for example: How do you do?), and patterns, which are defined as only partly analysed phrases, that is phrases which follow a set pattern but have one or two open slots to be filled (for example: How much did ___ cost?). These can, for our purposes, be equated respectively with the preconstructed and semi-preconstructed phrases discussed above. For Krashen & Scarcella such phrases are not central to the language acquisition process. They claim instead that their presence in learners’ speech results from the pressure, generally exerted by teacher on pupil, to communicate at too early a stage, that is, before the learner has had time to break

24

alan partington

them down into their constituent units. Other psycholinguistic and pedagogical researchers disagree. Ellis points out (1985: 167–9) that they are found in the speech of all varieties of learners, both classroom and non-classroom, child and adult, first and second language learners. As possible mechanisms of acquisition and production, Ellis also discusses pattern memorisation whereby the learner ‘‘attends to the input and [. . .] identifies a number of commonly occurring whole utterances in terms of the contexts in which they are used’’ and pattern imitation which involves ‘‘the deliberate and methodical copying of whole utterances or parts of utterances used in the speech of an interlocutor’’. Speakers are motivated to carry out these activities since set phrases tend to perform very useful functions, e.g. Can I have ____ ? or Thank you very much. Later on, many of the phrases memorised stay as they are because they continue to perform their particular function, whereas others are analysed into smaller parts as the learner’s syntactic skills develop, and can thus be employed in more varied circumstances. The parallels with schema theory are clear. Bolinger (1976: 8) sums up this process of pattern analysis by suggesting that ‘‘learning goes on constantly—but especially with young children—in segments of collocation size as much as it does in segments of word size’’ and that ‘‘much if not most of our manipulative grasp of words is by way of analysis of collocation’’. He draws analogies from morphology to support his argument. If speakers are capable of isolating a morpheme like un- from inside words to employ them creatively in other circumstances, then there is every reason to think that they also abstract whole words from lexical phrases they have memorised for use in new contexts. Peters (1983) has carried out studies of both first and second language acquisition and has formulated principles for deciding which phrases are unanalysed and which are available for the manipulation of their components. She suggests that there are two approaches to language learning which operate simultaneously—‘‘the gestalt approach’’, in which children attempt to use whole prefabricated utterances in socially appropriate contexts, and the ‘‘analytic one-word-at-a-time approach’’, in which children construct sentences from ‘‘scratch’’. In other words, the collocational and the open choice principles which, as we have seen, play their part in the creation of discourse are also the dual principles responsible for language acquisition. Finally, it has already been noted that the memorisation of the form of a phrase takes place together with the memorisation of its associated function(s). It is also apparent that once the learner begins to analyse a phrase into its constituent parts, the particular meanings and connotations it had can change drastically. Bolinger (1976: 7) discusses various combinations of intensifying adverbs with the word hurt. He notices that a phrase like badly hurt tends to be used when the hurt is physical, but when the intensifier is stronger, as in terribly hurt or cruelly hurt, the hurt

collocation and phrase patterns

25

is probably moral or emotional. Evidence from an analysis of the newspaper corpus described above suggests that the picture is more complicated still, that each combination of intensifier with hurt has one connotation or the other. All occurrences of badly hurt in the corpus do indeed refer to physical hurt. But the use of a stronger intensifier does not always imply the description of an emotional state. Whereas this is the case for deeply hurt, really hurt and terribly hurt, it is not true for seriously hurt or severely hurt which both indicate physical harm. There seem to be no general rules here; each combination has its own connotation (see also Chapter 4 below on semantic prosodies). If further evidence of this were required, the one occurrence in the corpus of the phrase hurt badly (with the intensifier following hurt) expresses emotional discomfort.

1.7 Types of collocations We have already noted that some collocational phrases are fixed and some allow degrees of variation. A fully fixed phrase is one which allows no syntactic transformation and no internal lexical variation. Sinclair (1991a: 110–11) cites the example of the phrase of course whose constituent words cannot be shifted around, or added to, or altered in any way. It effectively functions as a single unit and it is only the vagaries of English orthography which prevent it being spelt as one word along the lines of insofar as, an invariable phrase in which three out of four of its constituents are in fact written as one. Compare also in spite of which, with three orthographic words, does the same work as the single word despite. Nattinger & DeCarrico call these items ‘‘polywords’’ and note how some of them maintain a regular grammatical form (are ‘‘canonical’’, in their terminology), for example for the most part, beside the point, while others do not have the typical structures of English (are ‘‘non-canonical’’), such as time and again, so far so good and so on. Other types of fully fixed phrases include proverbs and sayings, quotations and those idioms which have a frozen form (e.g. on the wagon). They also include the category of fully restricted collocation, that is, phrases in which the appearance of one item implies the co-occurrence of another. For example, if we meet the word stinking before an adjective, this can normally only be rich, or if we meet the word blithering it is almost certain that idiot will follow. Variations in phraseology come about in a number of ways. As Sinclair notes, ‘‘many phrases have an indeterminate extent’’, in other words, it is not always possible to say precisely what is properly part of the phrase and what is an optional addition: As an example, consider set eyes on. This seems to attract a pronoun subject, and either

26

alan partington never or a temporal conjunction like the moment, the first time, and the word has as an auxiliary to set. How much of this is integral to the phrase [. . .]? (Sinclair 1991a: 111)

Many phrases allow what Sinclair calls ‘‘internal lexical variation’’. There seems to be little to choose, he says, between in some cases and in some instances. Other phrases allow some variation in word order (see paragraph below). We might add that some phrases are susceptible to lexical insertion. Even highly restricted collocational phrases such as vested interest(s) can have words inserted in them to produce, for example, vested financial interest or vested political interests. Carter (1987) proposes a number of clines in the criteria which are relevant in determining how fixed or how free particular lexical patterns are. One of these is the cline of syntactic structure or word order, similar to that described above. As an example of fixed order he cites phrases such as go it alone and the more the merrier. He points out that the more irregular (or non-canonical) a phrase’s syntax, as in the above examples, the more likely it is to be a fixed phrase. As an example of a phrase with a more flexible pattern he cites break someone’s heart which can take on a number of forms including he broke her heart, they’re heart-broken, she’s a right little heart-breaker. Another of Carter’s clines is that of collocational restriction. The most restricted are phrases such as stinking rich and blithering idiot where the occurrence of one item is conditioned by the other. Total conditioning however is probably quite rare, except for the special category of ‘‘irreversible (or ‘‘ordered’’) binomials’’, which include cash and carry, bread and butter, ups and downs and so on.3 At the opposite end of the scale, we have what Carter calls ‘‘unrestricted collocation’’, which ‘‘describes the capacity of particular lexical items to be open to partnership with a wide range of items’’. Most common lexical items (e.g. fat, bright, head) are in this category. In between the two extremes are the items which have a ‘‘semi-restricted collocational range’’, that is, they normally collocate with a limited number of other lexical items. Carter’s examples include the item harbour (as a verb) which collocates with doubts, uncertainty, a grudge, suspicion. There is less difference between the categories of ‘‘unrestricted’’ and ‘‘semirestricted’’ collocational potential than might at first sight appear. Whilst it is true that a word like head—one of Carter’s examples of an item with unrestricted collocational ability—collocates very widely, corpus evidence nevertheless demonstrates its collocational preferences, in the sense that it is found more often with particular items than others. It collocates frequently with bang, scratch, turn, nod, and so on, in non-idiomatic uses, with lose one’s, come to a, go to one’s, fill one’s, go over one’s etc in idiomatic expressions. The word head is not unusual in this. In fact, every lexical item enters into particular collocational relations with the rest of

collocation and phrase patterns

27

the lexis of a language, a behaviour which can be studied and described in terms of frequencies and preferences.

1.8 Conclusion In this chapter I have given a brief history of thought about collocation and phrase patterning. The chapter began with a number of overlapping definitions of collocation and went on to outline its role in discourse building and comprehension and its importance in language acquisition, both first and second language. It ended with a brief overview of the different types of collocation which have been proposed in the literature. The conclusion was reached that every lexical item in the language has its own individual and unique pattern of behaviour. The next two chapters contain examples of this, and they show how even words with very similar meanings are used in particular phraseological patterns. Chapter 2 looks at the behaviour of a group of intensifying adjectives often treated as more or less synonymous (sheer, pure, complete and absolute) and demonstrates how concordancing can compare and contrast their patterns of behaviour and therefore highlight differences in their use. Chapter 3 compares another group of near synonyms in English (correct, right, proper, just) with their look-alike translation items in another language (in this case Italian) and shows how concordancing can help translators choose the most appropriate item for translation by matching the behaviour patterns of words in two different languages. Subsequent chapters look at how concordancing can reveal information (often otherwise obscure) on various non-denotational typesof meaning—connotational,syntactic, textualand metaphorical—which lexical items can express. Where relevant, I return to the themes contained in this chapter, in particular, what the data analysed and eventual discoveries made might imply in terms of language learning and acquisition.

Notes 1. ‘‘If an adequate theory of language users and language use is to be developed, it seems that judgements must be recognized to be in fact not of two kinds but of four. (The two kinds referred to here are ‘grammaticality, with respect to competence’, and ‘acceptability, with respect to performance’) [. . .] I would suggest, then, that for language and for other forms of communication (culture), four questions arise: (a) (b)

whether (and to what degree) something is formally possible; whether (and to what degree) something is feasible in virtue of the means of implementation available;

alan partington

28 (c) (d)

whether (and to what degree) something is appropriate (adequate, happy, successful) in relation to a context in which it is used and evaluated; whether (and to what degree) something is in fact done, actually performed, and what its doing entails.’’—(Hymes 1972: 281)

2. In general, the kinds of texts which rely heavily on novelty and where prefabs are comparatively rare tend to be shortish texts, poems, advertising slogans, newspaper headlines and so on. 3. These combinations are called irreversible, and, in fact, when the two elements which compose them are found in the reverse order the resulting phrase will have an entirely different sense. The corpus shows that salt and pepper refers to the condiments on everyone’s table, whereas pepper and salt refers to the colour of someone’s hair. Black and white is used in reference to physical colours, as in black and white photograph, whereas white and black occurs in more figurative senses, such as reference to ‘‘skin colour’’, as in white and black people.

2 Collocation and Synonymy Yet the sentence groaned under the sheer tonnage of this freight train of substantives (Corpus).

2.1 Synonymy, translators and language learners The phenomenon of synonymy is a central interest for both the semanticist and the language learner. For the former, synonymy is an important member of the theoretical set of logical relations existing in language. For the latter, there is a good deal of evidence to suggest that vocabulary is often best acquired by analogy, in other words, remembered as being similar in meaning to previously acquired items (Rudska et al. 1982, 1985; see also Carter’s discussion (1987: 170–3) of word acquisition using semantically related grids). In addition, what we might term ‘‘definition through synonym’’ is a central feature of most dictionary organisation (Ilson 1991: 294-6). For motives of stylistic variation, non-native learners and translators have a pressing need to find lexical alternatives to express a particular concept, especially in writing. Harvey & Yuill (1994) found that searches for synonyms accounted for over 10% of dictionary consultations when learners were engaged in a writing task. However, given the rarity of absolute synonymy, learners also need to know which of the particular synonyms given by dictionaries and thesauruses is the most suitable for any given context. Harvey & Yuill reported that, in over 36% of synonym searches, learners reported that the entry did not give them the information they needed. No other type of dictionary look-up—which included searches for syntactic, semantic, collocational and register information—met with such a low level of success. Comparing the collocational behaviour of so-called synonyms in concordance data may supplement dictionary information and help learners decide in what circumstances substitution of one item for another is possible in a text.

2.2 Definitions of synonymy What exactly is synonymy? There would seem to be different kinds of it as well as different degrees. First of all, as Lyons, following the German philosopher Gottlob

30

alan partington

Frege, reminds us, it is not co-referentiality. Two items with the same referent are not necessarily synonyms since ‘‘synonymous’’ means ‘having the same sense’ not ‘having the same reference’ (1977: 199). This disallows pairs such as Michael and Professor Halliday or yew tree and Taxus baccata as well as the examples often cited from Frege: the Morning Star and the Evening Star (both having as referent the planet Venus), and from Husserl: the victor at Jena and the loser at Waterloo (both Napoleon). Nevertheless, the items in any of the above pairs might be synonymous on any given occasion (an occasion when they have the same sense as well as referent). In fact Lyons elsewhere (1981a: 148-9) makes the distinction between complete synonymy and absolute synonymy. Items are said to be ‘‘completely synonymous (in a certain range of contexts) if and only if they have the same descriptive, expressive and social meaning (in the range of contexts in question)’’. They can be described as ‘‘absolutely synonymous if and only if they have the same distribution and are completely synonymous in all their meanings and in all their contexts of occurrence’’. To test whether two items are synonymous in a given context, he appeals to Leibnitz’s principle of the identity of discernibles (i.e. of interchangeability), which he translates as ‘‘Two things are the same if one can be substituted for the other without affecting truth’’ (salva veritate) (1977: 160). He then goes on to postulate a third kind of synonymy, ‘‘descriptive synonymy’’ which he compares with ‘‘complete synonymy’’ in the following way: the selection of one lexeme rather than another may have no effect on the message being transmitted. In this case, we can say that the intersubstitutable lexemes are completely synonymous. The selection of one rather than the other may change the social or expressive meaning of the utterance, but hold constant its descriptive meaning (if it has descriptive meaning) in which case, we can say that the intersubstitutable lexemes are descriptively synonymous.—(Lyons 1977: 160)

Instead of ‘‘descriptive synonymy’’, Cruse uses the term ‘‘cognitive synonymy’’. Two utterances are ‘‘cognitively synonymous’’ if they fulfil the same truth conditions even though a part of the first utterance has been substituted by something else in the second: X is a cognitive synonym of Y if (i) X and Y are syntactically identical, and (ii) any grammatical declarative sentence S containing X has equivalent truth-conditions to another sentence S1, which is identical to S except that X is replaced by Y. (Cruse 1986: 88)

Later in the same book, he seems to relax somewhat the requirement of syntactical identity in discussing the pair of utterances:

collocation and synonymy

31

–I just felt a sharp pain.

and –Ouch!—(Cruse 1986: 271)

He claims ‘‘there is a sense in which the content of the message conveyed by these two utterances is the same, or at least very similar’’ but they differ in what he calls ‘‘semantic mode’’, the first being in ‘‘propositional mode’’ the second in ‘‘expressive mode’’. The relaxation of the requirement of syntactic identity opens the way to another kind of synonymy. Consider the two potential utterances: –You make me sick.

and –Will you ever grow up?

It is conceivable that in some contexts these two utterances could perform the same function—the same speech act, to use Austin’s (1962) terminology—presumably that of insulting or putting someone down. This kind of synonymy is clearly highly context dependent and would be a subcategory of Lyons’ ‘‘complete synonymy’’. We might call it ‘‘illocutionary synonymy’’. Moreover, throughout this discussion of synonymy, it must be remembered that speakers always have the option of refining any of the relations between lexical items. Almost any two items can be set up as synonyms by a speaker in the course of an ongoing discourse. Hockett (1958) gives the comic example of the wife who teased her husband by saying ‘‘that’s a nice shade of blue’’ whenever they happened to see something green, because he had once described a blouse she considered to be green as blue. Conversely, close synonyms are frequently treated as opposites, or at least as being in some sort of opposition (e.g. No it’s not roasted, it’s broiled) (Cruse 1986; Malmkjaer 1993). Cruse contrasts his cognitive synonymy with the concept of ‘‘absolute synonymy’’. As Lyons, he sees absolute synonymy in language as being extremely rare, perhaps even non-existent. As nature abhors a vacuum, languages—or rather language users—continually differentiate between items which are potentially synonymous. In terms of communicative efficiency it is a waste to have more than one item meaning exactly the same thing. Cruse, moreover, has this very interesting point to make about the search for absolute synonymy: two lexical units would be absolute synonyms if and only if all their contextual rela-

32

alan partington tions were identical. It would of, course, be quite impracticable to prove that two items were absolute synonyms by this definition, because that would mean checking their relations in all conceivable contexts [. . .] However, the falsification of a claim of absolute synonymy is in principle very straightforward, since a single discrepancy in the pattern of contextual relations constitutes sufficient proof.—(Cruse 1986: 268)

Although it clearly would not be possible to examine all the ‘‘contextual relations’’ of a pair of items, by utilising corpus data, it is possible to examine large numbers of their ‘‘co-textual relations’’, principally their collocational patterns. The language user does not need to worry about never finding absolute synonymy. Some items are nearer to being absolute synonyms of any particular item than others, that is, they are substitutable on more occasions in more contexts, and this is extremely valuable information for linguists, translators and language learners. In still more practical terms, what the learner or translator usually wants to know is whether term x can replace term y in a given circumstance, in his or her hic et nunc, as they are speaking or writing. The following section shows how such information can be obtained through a concordance-based examination of the differences in the co-textual relations of a group of semantically similar items. Finally, both Lyons and Kempson (1977: 40) link synonymy with other semantic relations, especially with the concept of entailment. ‘‘Synonymy’’ writes Kempson ‘‘is defined by logicians as mutual entailment’’. As an example she cites John is a bachelor and John has always been unmarried as two mutually entailing phrases.1 The question of entailment, even the non-mutual kind, is highly relevant to this discussion. An example of non-mutual entailment might be as follows: ‘‘If x is a cat’’ entails ‘‘then x is an animal’’, but ‘‘If x is an animal’’ does not entail ‘‘then x is a cat’’. The relationship of cat to animal is one of hyponym to superordinate, and it is always possible to replace a hyponymous term in a phrase with its superordinate without altering the truth value of the phrase, while doing the opposite may well do so. The Cobuild dictionary (1987 edition) often explicitly provides superordinate terms for its entries (see the example below of absolute given as a superordinate of sheer). The dictionary thus implies that the superordinate term is found in all the senses or uses or circumstances that the hyponym is found in, and has others of its own as well. Other dictionaries (including the 1995 second edition of Cobuild itself) seem not to make a distinction between items treated as synonymous and those treated as superordinates and indicate them all simply as ‘‘alternatives’’ to the entry item. Concordances from the corpus can be studied for clues as to whether phrases which include an allegedly hyponymous term have the same meaning if that term is replaced by the superordinate. Once again this kind of information is invaluable for translators and language learners seeking stylistic variation.

collocation and synonymy

33

2.3 Discriminating among semantically similar items: the case of sheer 2.3.1 Semi-grammatical items: a problem for learners From the general discussion above, we can move on to a specific problem for language learners. If the learner wishes to choose between relatively frequent lexical (as opposed to grammatical) items, modern learners’ dictionaries give extensive and reliable descriptions. These dictionaries are themselves compiled largely on the basis of corpora. It is when they need to make a choice among a set of related semigrammatical words (words which belongs to one of the categories generally described as being ‘‘lexical’’—noun, verb, adjective, adverb—but which tend to play a grammatical role in the sentence) that concordance data becomes especially useful as an addition to dictionary information. Examples would be the adverbs rather and somewhat which belong to a group classified by Quirk et al. (1985) as ‘‘downtoners’’ since they moderate, or tone down, the force of the adjective or verb they modify (a kind of grammatical function). Such items are notoriously difficult for dictionaries to deal with largely because they do not have a clear denotational meaning as do, say, elephant or cautiously. They can only really be described in terms of what they do to the phrases they find themselves in. Cobuild 1987 gives ‘‘somewhat’’ as a synonym of rather (but not vice-versa). However the definition of rather is ‘‘to a certain, limited, or slight extent’’, whereas the definition for somewhat is ‘‘to a fairly large extent or degree’’. LDOCE 1987 is undoubtedly more consistent, defining rather as ‘‘to some degree’’ and somewhat as ‘‘by some degree or amount’’, but these definitions are so vague as to be meaningless. It gives ‘‘rather’’ as a synonym of somewhat but not vice-versa—the other way round from Cobuild. LDOCE 1987 omits to tell the reader what is perhaps the most important distinction between them: that somewhat is of a more formal register than rather. What precisely being of a more formal register implies can be seen from their concordances. First of all, somewhat is much less frequent, suggesting that its use is restricted to special environments. Moreover, it collocates frequently with ‘‘difficult’’ or rarefied adjectives or adverbs—circuitous, disingenuously, fatuous, foolhardy, fusty, and so on. Rather is more ‘‘democratic’’ and appears with many more common-or-garden items.2 The item sheer is also a semi-grammatical word. Although it is an adjective and adjectives are usually defined as lexical items since they constitute an open set, its action in the phrase is that of intensification, a grammatical function. It is classified by Quirk et al. as belonging to the word class of ‘‘emphasizing intensifying adjectives’’. Other items which can perform this function include pure, complete, utter

alan partington

34

and absolute. What follows below is an attempt to measure the degree of collocational overlap (Partington 1992) that exists among these words. In other words, we will try to discover in what circumstances they can be interchanged in texts. At various stages in the course of the study, the concordance information is compared to that found in three learners’ dictionaries, Cobuild (1987 and 1995 editions) and LDOCE (1987 edition3). 2.3.2 The senses of sheer LDOCE 1987 defines three separate senses of sheer as follows: adj. 1 pure; unmixed with anything else; nothing but: He won by sheer luck / determination / The sheer size of the country (= the simple fact that it is so big) causes tremendous communications problems / It would be sheer folly to buy a car—we wouldn’t be able to run it. 2 very steep; (almost) straight up and down; PERPENDICULAR; a sheer cliff. 3 very thin, fine, light in weight, and almost transparent: ladies sheer stockings.

Thus pure is given as an equivalent of sheer when the latter is used in sense 1, i.e. a semi-grammatical, intensifying use. Cobuild 1995 defines the relevant sense of the word sheer as follows: You can use sheer to emphasise that a state or situation is complete and does not involve or is not mixed with anything else. His music is sheer delight . . . Sheer chance quite often plays an important part in sparking off an idea . . . acts of sheer desperation.

This definition gives complete as similar in meaning. In this dictionary’s so-called ‘‘extra column’’, the word pure is explicitly given as a synonym. In the 1987 edition, absolute is given as a superordinate of sheer. In an attempt to discover in what circumstances each of these words—pure, complete and absolute can be used in the same contexts as sheer, their collocational behaviour in the five-million-word corpus was examined. Sheer used as an intensifier appears 92 times in the corpus.4 In a large proportion of these, it collocates immediately to its right with some expression of magnitude or weight or volume: 1 rmat, the easy-to-read print, the sheer volume of reliable information, the 2 otages the play in the other: the sheer monotonous volume of Charlie's diat 3 loor—a way of suggesting that the sheer weight of them would be too much fo 4 s unprepared for the shock of the sheer weight of noise from the opposition

collocation and synonymy

35

5 glish rugby's greatest asset: its sheer weight of numbers. `If we've got ou 6 reform, in order to deal with the sheer numbers of unemployed, landless pea 7 In the industrial revolution, the sheer number of artefacts got out of hand 8

same materials. To cope with the sheer number of drawings—comparing the BM

9 began to succumb yesterday to the sheer magnitude of the task facing it. By 10 yet they counter this with their sheer mass. They read like condensed equi 11 arths what symbolism buries, the sheer scale of destruction. Once the 350, 12 lotteries will be drowned in the sheer scale of prizes: the bigger countri 13

and the Banque de France is the sheer scale of the selling. We are talkin

14

used to this sort of thing: the sheer scale of chaos and inhumanity goes

15 ay, everyone's talking about the sheer quantity of football on television. 16 ich reflects that here it is the sheer multiplicity of specimens rather th 17 rocesses. If only because of its sheer size, it is plausible tha 18 t I realise the magnificence and sheer size of the stadium. It brings on a 19 t the sentence groaned under the sheer tonnage of this freight train of su

The only items which occur immediately to the left of the keyword are the (15 times including line 18, where it appears three words to the left), its (twice), their (once). In the great majority of examples to the right of the ‘magnitude’ word we find of, followed by a noun phrase. The typical phraseology here, then, is ‘‘the sheer (magnitude word) of (noun phrase)’’. LDOCE gives an example—‘‘The sheer size of the country’’—with this structure, but Cobuild 1995 does not. The magnitude words include: weight and scale (3 times each), size, volume, number, numbers (all twice). In another large group, sheer collocates with items expressing ‘force’, or ‘strength’ or ‘energy’: 20 rieux real but already past. The sheer energy of Bussell's first act gives 21 But even this doesn't match, for sheer exertion, Calypso's approximation p 22 ngs in seismic areas, said: `The sheer force of an earthquake is less impo 23 me fell apart. Nevertheless, the sheer force of his presence on the field 24 or and imposed his will with the sheer force of his personality; Cliff Mor 25 ility to conclude deals. Through sheer force of personality he has managed 26 umbers, it also demonstrates the sheer charismatic force of Lydon's fund o 27

oup match. `He was 18 1/2 st of sheer muscle,' Leonard said. `I'd never p

28 ich other qualities can outweigh sheer brute strength. It was hard to argu 29

in his heyday have done, mixing sheer, brutal power with devious elastici

30 er on the near post by virtue of sheer power. It was like an instant hango 31 icantly by population growth and sheer pressure on existing farmlands (alb 32 s widespread, due in the main to sheer pressure of numbers of small-scale 33 lem seems to be in part that the sheer imaginative effort of thinking abou 34

coach journey to Essex.

The sheer fury of Southend's running kept Spu

35 least to see for themselves that sheer physique—England's tour literature 36 tes that make it vulnerable. The sheer pace of its expansion is part of th

36

alan partington

37 ity of the Brucknerian pace, the sheer intensity of the mountainous climax 38 rs like a broken puppet.

The sheer physicality of this piece may seem 39

The instrumental virtuosity and sheer physical energy required to carry t

Sheer is again sometimes preceded by the, but less often than in the ‘magnitude’ group (here 10 from 20). We also find for (l.21), through (l.25), by virtue of (l.30), due in the main to (l.32). The phraseology ‘‘sheer (noun phrase) of (noun phrase)’’ is common but not as overwhelmingly so as in the previous group. The right-collocating nouns include: force (5 times), power, pressure, personality (all twice). A third group contains right-collocates which express some sort of idea of ‘persistence’: 40 he game and the third went wide. Sheer persistence by Jonathan Couves, six 41 retirement, the reasons why, the sheer irreversibility of it. Then, lo and 42 red saves by Hans Segers and the sheer obstinacy of the Dons—who refused t 43 menal ball-carrying but also his sheer indomitability. Small wonder Alan D 44 e did succeed, sometimes through sheer insistence, in creating excitement. 45 ys to Leave your Lover', was his sheer reliability. You press Sanborn's bu 46 charm, determination, humour and sheer integrity got him through so many b 47 ble combination of cockiness and sheer resolve on the right side of a midf 48 ment, bold stroke-play—while the sheer consistency of Hugh Morris, now unr 49 Their music is now loved for its sheer character and individuality, and fi 50 t the climactic encounter out of sheer cussedness; that I couldn't was

do

51 ip it into some kind of life. By sheer hard work they do manage to make it 52 is late start his enthusiasm and sheer hard work meant that things moved q 53 nition,' Gradi said, `because of sheer hard work and determination to succ

Here the phraseology ‘‘the sheer (noun phrase) of (noun phrase)’’ is no longer dominant. Instead we find the keyword preceded by through (l.44), out of (l.50), by (l.51), because of (l.53). We can add to these with the, through, by virtue of and due in the main to which were found to the left of sheer in the second group. When sheer collocates with a noun expressing ‘force’ or ‘persistence’, there is very frequently a sense of causation or agency, that is, something is ‘sheer enough’ to cause something else to happen. Even in occurrences which do not contain an explicit expression of agency or causation like through or by virtue of, this sense is still clear. Often, for example, the noun phrase containing sheer functions as the subject of a verb in a causative relationship e.g. The sheer energy of Bussell’s first act gives the second [. . .] (l.20), his [. . .] sheer integrity got him through [. . .](l.42), his enthusiasm and sheer hard work meant that things moved quickly (l.52). If we expand the context a little, many more examples will become available, e.g. by expanding line 40, we find Sheer persistence by Jonathan Couves, six minutes into the second half, increased England’s lead [. . .].

collocation and synonymy

37

Neither the Cobuild dictionaries nor LDOCE make any explicit reference to this use of ‘‘sheer + noun’’ in causative expressions, which is probably its most frequent and characteristic use. On the other hand, LDOCE and Cobuild 1987 (but not Cobuild 1995) contain examples of this use (my emphasis in each case): –Many of the audience walked out through sheer boredom (Cobuild 1987) –He won by sheer luck/determination (LDOCE) –The sheer size of the country causes tremendous communications problems (LDOCE)

Sheer also collocates with nouns expressing ‘strong emotion’: 54

early Rolling Stones songs, for sheer foot-tapping fun. My Father's Verti

55

Norwich that he encountered the sheer joy of playing the game. `I loved t

56 's mixture of sophistication and sheer joy in life, Viola and Sir Toby. Hi 57 ike a baptism of fire. `It's the sheer panic you go through in deciding wh 58 cation of its continuity, by its sheer inspiration. Architecture Pa 59 University, prolonging it out of sheer enjoyment. Her enthusiasm for the I 60 performers results in moments of sheer exhilaration, a feeling of exaltati 61

bold Beckettian way, but out of sheer terror lest the tiniest bit of plot

62 gnored ethnic boundaries, and by sheer terror—killings, forced collectivis 63 ame breath as his father—I mean, sheer jealousy. And of course Bron natura

and with nouns expressing some kind of ‘personal quality’, including ‘beauty’: 64 d the ear with its potential for sheer beauty. The adjustable acoustic fou 65 ll `sat transfixed by the film's sheer beauty and its evident sympathy and 66 is a proven Cavaradossi, and for sheer beauty and poise this version takes

or ‘nastiness’ of some kind: 67 its reluctance to please, in the sheer brutality of its forms. The paintin 68 n which true pathos jostled with sheer emotional thuggery. Julian Jensen ( 69 p to fight out of hunger and the sheer hardness of street life. In Jimmy W

or a mixture of the two: 70 , he didn't have the weight, the sheer glamour of evil, to make a true vil

If ‘folly’ can be considered a ‘personal quality’, then these examples can be added: 71 tence, corruption, stupidity and sheer folly which has dominated half a ce 72 c ties are being broken, this is sheer madness; it reminds me of the atmos 73 dend without a recovery would be sheer folly. Mr Forte knows that, but hop

38

alan partington

Noun phrases expressing the qualities of ‘competence’ or ‘expertise’ (or lack of them) are also relatively common: 74 g can replace the skill, wit, or sheer professional expertise of a talente 75 s is a stolid affair as yet, its sheer choral and orchestral competence dr 76 o miss both available boats. The sheer virtuosity and textural ingenuity o 77 r any other `ism', but plenty of sheer good writing and strongly chiselled 78 the government financial process—sheer gamesmanship to offer succour to th 79 ce that they betray weakness and sheer lack of class despite (Berlusconi's 80 the clearest evidence yet of how sheer slackness and mismanagement led to

Sheer sometimes appears in a prepositional phrase introduced by for. We have seen six examples so far, of which the following three have much in common: 21 But even this doesn't match, for sheer exertion, Calypso's approximation p 66 is a proven Cavaradossi, and for sheer beauty and poise this version takes 82

with a swagger By ROB STEEN For sheer gall it took some beating. At the s

The phrases containing the keyword all explain the reason for which something stands out above other similar items. The last two contain versions of an idiomatic expression which has the phraseological pattern ‘‘for sheer (noun), (noun phrase) takes some beating’’. It is noticeable how, in the ‘emotion’ and ‘physical quality’ groups, the noun phrase containing sheer frequently (15 examples out of 28) premodifies a member of a series of nouns: 56 's mixture of sophistication and sheer joy in life, Viola and Sir Toby. Hi 58 cation of its continuity, by its sheer inspiration. Architecture Pa 65 ll `sat transfixed by the film's sheer beauty and its evident sympathy and 66 is a proven Cavaradossi, and for sheer beauty and poise this version takes 67 its reluctance to please, in the sheer brutality of its forms. The paintin 69 p to fight out of hunger and the sheer hardness of street life. In Jimmy W 70 , he didn't have the weight, the sheer glamour of evil, to make a true vil 71 tence, corruption, stupidity and sheer folly which has dominated half a ce 72 g can replace the skill, wit, or sheer professional expertise of a talente 75 s is a stolid affair as yet, its sheer choral and orchestral competence dr 76 o miss both available boats. The sheer virtuosity and textural ingenuity o 77 r any other `ism', but plenty of sheer good writing and strongly chiselled 79 ce that they betray weakness and sheer lack of class despite (Berlusconi's 80 the clearest evidence yet of how sheer slackness and mismanagement led to 81

of Scorsese's contribution, the sheer fatuousness of Francis Coppola's `L

collocation and synonymy

39

Finally, it collocates with items expressing ‘fortune’, good or bad, as noted by LDOCE in its example ‘‘He won by sheer luck’’: 84 be the last role of his life. By sheer chance Beineix has created a metaph 85 anners which appalls us, but the sheer bad timing. We have come all this w 86 g planes. Once launched, it was `sheer bad luck' one of the missiles hit t

A final point is that in every single concordance example sheer appears in attributive or premodifying position, never appearing in a predicative form such as ‘‘the x was sheer’’.

2.4 How synonymous are the ‘‘synonyms’’ of sheer? As has already been mentioned, the language learner or translator needs to know the precise circumstances in which a lexical item can be substituted by a similar item. In reality, however, the choice of item is often extremely complicated. Given the fact that the selection to be made is between semantically related items, the substitution of one by another may not change the meaning of the phrase in any evident way, but one item may well be more appropriate than another at that point. In other words the learner/translator must know the collocational habits of the related items in order to achieve not just semantic feasibility, but also collocational appropriacy. With this in mind, the behaviour of pure, complete and absolute are examined. 2.4.1 Pure LDOCE’s entry for pure comprises six senses, of which senses one and five are the only ones relevant to a comparison with sheer: 1 not mixed with anything else: ‘‘Is this sweater made of pure wool?’’ ‘‘No, it’s 60% wool and 40% acrylic.’’ ¦ pure silver ¦ a horse of pure Arab breed. [. . .] 5 [A no comp.] infml complete; thorough; only; By pure chance/coincidence my boss was flying on the same plane as me / The error was due to carelessness pure and simple (=only carelessness)

The wording of the definition of sense one—‘‘not mixed with anything else’’ recalls that of the relevant sense of sheer which was defined by the same dictionary as ‘‘unmixed with anything else’’. Sense five of pure implies the word’s use as an emphasiser. It is interesting that a number of alternatives to pure are given for this sense (‘‘complete’’, ‘‘thorough’’, ‘‘only’’) but they do not include sheer.

40

alan partington

In the Cobuild 1995 extra column, ‘‘sheer’’ is given explicitly as a synonym of sense six (out of eight) in its entry for pure, which is worded as follows: Pure means complete and total. The old man turned to give her a look of pure surprise . . . To sleep on my own and not hear the boys snore or grunt was pure bliss

Sheer would be a safe substitute in at least the second of the example phrases here, since ‘‘bliss’’ belongs to the semantic field of ‘strong emotion’. Cobuild 1995’s sense one for pure is as follows: 1 A pure substance is not mixed with anything else, EG . . . a carton of pure orange juice

which resembles the dictionary’s wording of the relevant definition of sheer—‘‘Sheer means complete and not involving or mixed with anything else’’ (my emphasis). However sheer could not substitute pure in any of the examples given in the two dictionaries for this sense (‘‘pure wool’’, ‘‘pure Arab breed’’, ‘‘pure orange juice’’, but not sheer wool *, sheer Arab breed * or sheer orange juice *). These dictionary definitions seem to suggest that the relationship between sheer and pure is one of hyponymy, that sheer is synonymous with only one sense of pure—that of emphasising adjective. For the dictionary user, the fact that Cobuild (both editions) indicates synonyms in the extra column against a particular sense is very helpful. To improve things still more, when learners’ dictionaries give a synonym for an entry, they might also indicate precisely which sense(s) of that item was synonymous with the entry, e.g. LDOCE, rather than simply giving pure as a synonym in the entry for sheer, could indicate that this holds only for its sense number five of pure, and likewise, the Cobuild dictionaries might indicate that only its sense number six of pure has this function. But it still remains to be seen whether this sense of pure is really synonymous with sheer. The corpus has 135 examples of the use of pure. It is found in a larger variety of syntactic environments than sheer, which was only found in attributive position (i.e. preceding the noun head it modifies). This, of course, reflects the fact that pure has several more highly lexicalised senses in addition to being an emphasiser. In many other occurrences pure is used attributively, but sheer can still not take its place: • to preserve the pure doctrine of religion [. . .] • neither a Jew nor a Christian, but he was of the pure faith who submitted to God (muslim). • if God is pure goodness [. . .]

collocation and synonymy

41

• and tried a variety of pure chemicals. • contained only about 1 per cent of pure penicillin. • His nature was pure gold.

In the first three examples, pure has a religious-moral semantic content, which sheer does not share. The others however, mean ‘‘unmixed with anything else’’, yet sheer still cannot be substituted. This is probably because sheer is not used to modify material substances (except in one of its rare lexicalised senses i.e. sheer silk). Neither of the dictionaries gives this information. On the other hand, Cobuild 1995’s definition for the first sense of pure does in fact mention ‘‘substance’’—‘‘a pure substance is not mixed with anything else’’. The following is a list of the examples where pure not only means ‘‘unmixed with anything else’’ but in which it also has a certain amount of emphasising function. In these cases we might predict a degree of overlap with sheer. 1 Yes, it's good isn't it? It was a pure accident, but when I noticed I had i 2 ion of the Italian lira: that was pure Bundesbank. Even after the half- poi 3

reaction attributed evolution to pure chance and held the opinion that com

4 ipe, find an open window, and, by pure chance, enter the bedroom of the fab 5

outcome which may be a matter of pure chance: if such driving happens not

6 he reactions of spectators. Here, pure cinematic skill must carry the day, 7 m a Planetarium-like skyscape—was pure comedy. In the midst of this, action 8 hed for. But think about the line—pure Disney. It's mood shattering, unreal 9 10

Although, Tognoni added, ``it is pure experiment', FIFA is not averse to i way beyond debt management into pure financial speculation.

There are

11 > Bodybuilding: Charlie rules on pure flex appeal: Joe Lovejoy enters the 12 ight have invoked was quelled by pure funk—in Hoomey's case, at least—of w 13 rnational stocklending market is pure guesswork. What is not in doubt is t 14 s of flowers that, in taste, are pure Holiday Inn. The Chalk Garden contin 15 o state unpleasant truths out of pure honesty. He was sensitive, vulnerabl 16 elief, finally, to hear a bit of pure humbug in which the kilt and sporran 17 ould have been prevented. It was pure idiocy for countries to destroy each 18 Second Division football, it was pure Jim Dunk. On television, the lugubri 19 iction the image would have been pure kitsch _ a girl riding a horse, her 20

forget the . . . hatred . . . a pure malice against us.' Someone else mur

21 in on a lacklustre job -- it was pure marketing hype. The stock market had 22 of a musical soundtrack which is pure movie magic and builds to a shudderr 23 m and Paul Gambacinni. There was pure nostalgia from Barbara Castle, 84, t 24 d for the National Anthem out of pure orneriness. After a year, he left sc 25 s going to be. Now I stay out of pure patriotism. In any given day, I sudd 26 h the progress of the drama; the pure pleasure of voices in ensemble soari 27 vincing many that he is the best pure rusher in today's game, perhaps the

alan partington

42

28 t the crab, an emblem for him of pure selfishness, a life of unbroken take 29 terday dismissed the reports as `pure speculation'.

September's 7.6 pe 30

viewers who couldn't see Hill's pure talent for the belching clouds of vu

31 n both be reduced to a matter of pure taste. The mainstream critics' thank 32 g maul was, however, a moment of pure theatre, and the crowd rose to him. 33 those privileged to watch it was pure theatre. He knew it. His own appearr 34 onging for Astrov, she expressed pure vulnerability. I did not feel that I 35

the client out of the salon was pure West End theatre. Raymond was a one-

36 either by pure functionalism nor pure whim. As Outram says: `It doesn't ma 37 harles Simpson, machinating with pure wickedness), while the action has be 38

his own patrician drawl: it was pure Willie Mufferson. What is clearly tr

There is clearly no trace of pure collocating with items expressing either (i) ‘magnitude’ or (ii) ‘force’ or (iii) ‘persistence’, with which sheer was found to collocate most frequently. Pure, however, does collocate at times with items belonging to some of the same semantic classes as sheer. In examples 3, 4, 5 it collocates with chance, in line 17 with idiocy. It occasionally co-occurs with nouns expressing some personal quality—both positive and negative ones—as in lines 11, 15, 28, 34 and 37. Now and again it collocates with items indicating emotion—lines 12, 23, 26 and perhaps 36—but the emotions seem to be less extreme than those which often appear with sheer (which included joy, terror, panic and so on) and the frequency of such collocations appears to be much lower. Unlike sheer, pure also collocates with proper nouns (lines 2, 8, 14, 35, 38). An examination of the phraseology of these lines shows that pure is frequently preceded by the verb be in some form or other—16 times in the 38 lines (only 7 times in 92 occurrences of sheer). Conversely, pure only occurs 7 times in some kind of prepositional phrase expressing causation (lines 3, 4 and 5—where ‘‘chance’’ is the causal agent—15, 24, 25, 36), nor does it appear in prepositional phrases introduced by for (explaining the reason why something is outstanding), both common phraseologies for sheer. These observations seem to show that, whereas sheer is used to explain causal relations—something happens because of sheer something (force, persistence, emotion etc), or something is outstanding due to sheer something—pure is more typically involved in describing a state—an item simply is pure something—which would explain its frequency of occurrence with ‘‘be’’. Translators or language learners would find such general facts extremely useful. 2.4.2 Complete Cobuild 1995, if we recall, includes the item complete as part of its definition of sheer. On the other hand, LDOCE does not indicate complete as related to sheer and

collocation and synonymy

43

Quirk et al. place it in a different word class, that of the ‘‘amplifying intensifiers’’ (sheer was counted among the ‘‘emphasizing intensifiers’’). As an adjective, complete appears 382 times in the five-million-word corpus, which makes it four times as frequent as sheer and three times as frequent as pure. It is used both attributively and predicatively, and like pure, shows a far greater syntactic flexibility than sheer. Only those examples where complete premodifies a noun are considered. Sheer frequently appeared in the environment ‘‘the sheer (magnitude or force word) of (noun phrase)’’. Complete sometimes appears in the construction ‘‘the complete (noun) of (noun phrase)’’: 1 treet _ made more striking by the complete absence of the South African Pol 2 hese abnormalities range from the complete absence of limbs to loss or gain 3 atment. This is so because of the complete absence of perspective. Chronolo 4 is reported to be negotiating the complete withdrawal of the 1,500 Soviet m 5 ctor of the new company, said the complete transfer of the Japanese company 6 on to strive for agreement on the complete destruction of nuclear artillery 7 accompli, effectively heralds the complete break-up of Deloitte Haskins &an 8 the most important countries, the complete disintegration of Deloitte Haski 9

a big step yesterday towards the complete abandonment of Communism when it

10 n cannot join the EMS before the complete removal of capital controls in F 11

from novels by Tom Wolfe to The Complete Book of Portable Power Tool Tech

12 hich, they felt, didn't give the complete picture of American music. `By n 13 by picture, and to scale, of the complete oeuvre of the great French ninet

The noun following complete is not a magnitude or force word. There is a very high frequency of words expressing ‘absence’, ‘change’ or ‘destruction’.5 It also collocates in these other structures with ‘absence’ words: 14 world people for them to request complete anonymity when discussing him. < 15 re saying that there has to be a complete ban on overtime and that they sh 16 the doctor-narrator claims that `complete indifference is the only humanit 17 nd mystery, the consequence of a complete news black-out by the authoritie

in these structures it collocates with ‘change’ words: 18 p> As for Llanelli, they offer a complete contrast. Their side is settled, 19 exercise should be extended to a complete revamping of the Community's ins 20 ng the videotaped interview as a complete substitute for the child's evide

and in these with ‘destruction’ words:

alan partington

44 21

probably are, suggest an almost complete collapse of its electoral appeal

22 German industry would be left in complete ruins if the raids continued unh

None of these collocational sets were found in connection with sheer or with pure. Complete is never used with the causal function which distinguishes sheer. Like pure it tends to describe a state, but seems to collocate with different items. A language note in LDOCE refers the reader to an explanation of ‘‘intensifying adjectives’’. It claims that complete, along with absolute, total and utter can be used ‘‘in front of words which express very strong feelings’’ (examples given include agony, astonishment, bliss, despair, joy, ecstasy), what might be called ‘‘hyperbolic nouns’’. Since sheer was sometimes found to premodify nouns expressing strong feeling, we might imagine that in this use at least complete and sheer might be interchangeable. However, in the nearly 400 examples of complete in the corpus, not one collocates with a noun of strong emotion. The strong emotion nouns cited often appear with an intensifying adjective but they collocate as follows: agony with terrible and exquisite; astonishment with deep and substantial; bliss with exceeding and rare; despair with blind, deep, terrible, ponderous and utter; and joy with boundless, delirious, simple and sheer (twice). Ecstasy is not found with any intensifiers, only with a few classifying adjectives such as erotic, sexual and spiritual. In writing at least, a large variety of adjectives are used to intensify emotions, and examples like blind despair or boundless joy show that many of them are linked to a particular noun forming a fairly fixed collocation. It may be true, however, that items like complete, absolute and utter occur with emotion nouns more frequently in the spoken language. 2.4.3 Absolute Absolute, indicated by Cobuild 1987 as a superordinate of sheer, appears as a noun modifier in the corpus 134 times. In a good proportion of these, it is used in a technical-political sense e.g. absolute majority (8 times), monarch (4), sovereign/ sovereignty (5), authority (4) etc, which we can discount for the purposes of a comparison with sheer. Even in less technical uses such as the following: 1 the clergy to assume and maintain absolute control of their schools led a n 2

own citizens, and its desire for absolute control of information. The cost

it has a lexical sense of ‘unchallenged’, ‘unquestioned’ that sheer does not. These lines contain the construction ‘‘absolute (noun) of (noun phrase)’’ as do the following: 3 es but yet again demonstrates his absolute command of the tough and tender

collocation and synonymy 4

45

a week. `The big four were under absolute pain of death to keep quiet,' a

5 efore currencies have reached the absolute limits of their bands, since by 6 t promptly said there would be an absolute howl of anger in parliament. Rea 7 s by this Finnish master is their absolute command of musical movement, and 8 of training jumpers. Here was the absolute antithesis of the light-framed, 9 ones and the triad are proper and absolute foundations of music has resonat 10 cept the 25 per cent quota as an absolute minimum of television programmin 11 est scale. As they controlled an absolute majority of shares, the club was 12 n debate, long entrenched in the absolute positions of the opposing activi 13

emissions is well known. Though absolute amounts of carbon dioxide emitte

14

beyond my understanding. He has absolute peace of mind; faith before whic

15 breviated here: a monolithic and absolute view of the world, with its acco 16 ges in performance, not the absolute level of performance, that are m

The first noun phrase never contains a ‘magnitude’ or ‘force’ word. In these lines, absolute sometimes has the lexical senses ‘unchanging’, ‘not relative’ (lines 9, 10, 12, 15), not found with sheer. The following are all cases in which absolute shows a similar collocational behaviour to sheer. There are two examples where absolute is emphasising items from the ‘competence’ or ‘expertise’ semantic set: 17 d-winner, even though it rewards absolute excellence in literary terms. Si 18 ly lies in an ability to combine absolute precision and clarity with the i

There are also two lines in which absolute collocates with a noun expressing ‘persistence’ or some related quality: 19 rk, unlimited self-sacrifice and absolute determination.

They started 20 his career on the Turf, that is, absolute integrity, kindliness and under

As was often found to be the case with sheer, the phrase including absolute is part of a list. As far as the other concordance lines are concerned, it is noticeable that absolute collocates occasionally with hyperbolic nouns. The quality expressed by these nouns can be either positive or negative: 17 l shift by Labour.

`It is an absolute bloody shame that the Labour Par 18

a quarter past seven and it was absolute chaos. The place was devastated:

19 ,'' Brown added. (Silence.) `The absolute cheek.'' (Lone snigger.) `Readin 20 nds at a hefty 20. The boy''s an absolute disgrace to his class. 21 ealth spokesman, said: `It is an absolute disgrace that the commercialisat 22 ss and honest self-appraisal. An absolute gem. Would that one could say th

alan partington

46 23

and the most collected—he''s an absolute genius,'' says Coleman, who has

24 this pub and they think it''s an absolute godsend. It''s usually the local 25 tic psychodrama, Bitter Moon (an absolute hoot), who claim that it was alw 26

me or indeed anybody else.

`Absolute horror, and then recognition tha

27

promptly said there would be an absolute howl of anger in parliament. Rea

28 ved, going where they had to go. Absolute outrage. Grant Simon, who had wo

Although sheer was sometimes found with nouns expressing strong emotions, absolute is found with a far wider range. Of the adjectives examined here, it is the one which intensifies already hyperbolic nouns with greatest regularity. A number of the uses of sheer are also found to be typical of absolute, but the latter has a wider variety of syntactic and collocational behaviour. From this evidence, is Cobuild 1987 justified in claiming that absolute is in a superordinate relationship to sheer? Only in part. Sheer appears in phraseologies in which absolute does not, notably in patterns expressing size and force. More important still, there is no evidence at all that absolute is employed to express causation, the function which distinguishes sheer most markedly.

2.5 Conclusion This chapter began with a discussion of the importance for language learners and translators of being able to choose among a selection of near synonyms. The way to make such a choice is to acquire as much information as possible about the contexts in which each item is used, and it was suggested that an effective way of doing this was by studying its concordance. In terms of the example discussed here, sheer was found to be involved in a number of typical phraseological patterns, which tended to correspond to a particular function (of expressing causation), which set it apart from a number of near synonyms. Each of these synonyms was also found in its own particular set of phraseological patterns, and this fact supports the argument reported at the end of Chapter 1 that every lexical item in the language has its own individual and unique pattern of behaviour. If then the situation concerning sheer and its ‘‘synonyms’’ is frequent, it is no surprise that translators and language learners experience great difficulties in the search for equivalents in texts. Dictionaries such as LDOCE and Cobuild thus do well to give numerous examples since these can provide information about typical uses and are just as useful as the definitions given for entries. On the other hand, they need to be particularly careful in indicating which other items, and which senses of these items, can be considered synonymous (or superordinate or hyponymous) to the entry word. By the same token, the thesaurus—which

collocation and synonymy

47

simply gives lists of synonyms—is positively dangerous for the non-native speaker.

2.6 Suggestions for further study The procedures and techniques described in this chapter are productive, in the sense that they can be used to study other groups of sematically similar items. This kind of research—using corpus data to distinguish the precise uses of similar words or phrases—is one of the easiest and most interesting exercises for students who are learning concordance analysis. Depending on their level of language competence, they might begin by comparing the concordances of grammatical items, such as for and since, or phrases such as try to and try *ing (i.e. try to open and try opening etc), or prepositions with a superficially similar meaning such as over and above. Concordances of semantically similar lexical items, such as look, see and watch, can also be studied, and students will inevitably discover differences in use which are not contained in grammars and dictionaries. The Cobuild publication Confusable Words (Carpenter 1993) can be used to glean ideas for items to be compared.

Notes 1. In a footnote Kempson adds: ‘‘I am assuming that sentences such as these are synonymous, though it has sometimes been doubted whether two sentences are ever truly synonymous’’, thus implying that although the two phrases might be descriptively equivalent, there is the possibility of their differing in regard to other kinds of meaning. The distinction between descriptive meaning as defined by truth conditions and other kinds of meaning she relates to the distinction between semantics and pragmatics. 2. A didactic rule of the type ‘‘do not use somewhat with words like good, bad, happy, sad’’ may be useful to some learners. The whole concept of register can be seen from a lexical point of view i.e. that restrictions of register can be interpreted as limits on the collocational possibilities of an item. The existence of rules of register (albeit fuzzy rules) like the one just given raises the theoretical possibility of writing a grammar of register. 3. The latest (1995) edition of this dictionary, considerably revised with respect to the earlier version, gives no cross referencing between the emphasising adjectives under study in this section, with the exception that ‘‘pure chance/greed/hell etc’’ is defined as ‘‘complete chance etc’’. While this avoids the risk of leading users astray, it does not help them in the search for stylistic alternatives. 4. Two other occurrences were found in the sense of ‘very steep’. 5. The related adverb completely was also found to collocate with items from these three semantic fields (see following chapter).

3 True and False Friends And Tony would make one an absolutely correct safari suit, a proper one with all the pockets in exactly the right place (Corpus). In the preceding chapter, it was shown how concordance data could reveal details of context, of typical collocational and phraseological behaviour, which enabled the researcher to distinguish among items in the same language which are closely related sematically. In this chapter, somewhat similar techniques are employed to examine a well-known problem in translation: evaluating when items which are generally described as being ‘‘true’’ (as opposed to ‘‘false’’) ‘‘friends’’ are reliable translation equivalents and when, instead, another item from the same semantic field would be a better choice. Taking examples from English and Italian, two cases are considered: (i) the ‘‘look-alike’’ items correct and corretto; (ii) the set of adverbs absolutely, completely and entirely and their counterparts assolutamente, completamente and interamente. To make comparisons, items were concordanced from two parallel corpora: the newspaper section of the English corpus and the Italian corpus (which consists exclusively of newspaper texts: see Introduction). The starting hypothesis is that a study of a large number of uses in context of such pairs of items, made possible by a concordancer, can provide more information about the textual environments in which pairs of items constitute translation equivalents than is available through traditional means such as mono- and bi-lingual dictionaries, thesauruses and so on. Before describing the two case studies, it is necessary to discuss what is meant by translation ‘‘equivalence’’, and what the status of ‘‘false friends’’ is.

3.1 Equivalence in translation In the field of translation studies, the concept of equivalence comes under regular scrutiny. Gentzler (1993: 4) lists equivalence/non equivalence as one of the standards for analysing translation which he criticises as being too abstract and limiting. He also cites previous authors who, he claims, rejected the concept of equivalence

true and false friends

49

as being impractical, notably Quine (1960: 27) and Holmes who is quoted as making the following point: Put five translators onto rendering even a syntactically straightforward, metrically unbound, imagically simple poem like Carl Sandberg’s ‘‘Fog’’ into, say, Dutch. The chances that any two of the five translations will be identical are very slight indeed. Then set twenty-five other translators into turning the five Dutch versions back into English, five translators to a version. Again the result will almost certainly be as many renderings as there are translators. To call this equivalence is perverse.—(Holmes 1973–44: 68 in Gentzler 1993: 95)

Sager, on the other hand argues that ‘‘fundamental to all theories of translation is the concept of equivalence’’ (1994: 142). Moreover, ‘‘units of translation and the search for equivalents for these units must lie at the heart of any theoretical or practical discussion about translation’’ (1994: 222). The contrast in opinions, however, is probably more apparent than real. Sager is clearly talking about a concept of translation equivalence which is much more subtle and complex than the primitive sort which Holmes is reacting against in the passage above. He goes on: There is, however, considerable diversity in the interpretation of what is meant by this concept. It is generally recognised that the relationship of a source and a target text is one of cognitive, pragmatic and linguistic equivalences.—(Sager 1994: 142)

‘‘Cognitive equivalence’’ is equivalence of informational content in two languages. ‘‘Pragmatic equivalence’’ implies that, whatever contextual changes need to be made to a text, its significance is the same for the target community as the source community. ‘‘Linguistic equivalence’’ implies the maintenance of form at various levels, those of text type and of structure at textual, sentence, phrase and word level. The problem, then, is not whether translation seeks to achieve equivalence but the inherent difficulty of defining and evaluating what translation equivalence consists of: How these equivalences are achieved and how they operate is, however, far from clear. Equivalence between the two documents involved in translation can also be stipulated at different levels and there is further diversity in the evaluation of what is considered successful equivalence.—(Sager 1994: 142)

It is generally accepted that complete equivalence of message, function and form between source and target text is rare, to say the least. For this reason, many theorists prefer to talk of the adequacy of a translation:

50

alan partington There is also a problem concerning the use of the term ‘‘equivalence’’ in connection with translations. It implies that complete equivalence is an achievable goal, as if there were such a thing as a formally or dynamically equivalent target-language version of a source language text. The term is, of course, usually intended in a relative sense [. . .] But the concept of ‘‘adequacy’’ in translation is perhaps a more useful one. Adequacy of a given translation procedure can then be judged in terms of the specifications of the particular translation task to be performed and in terms of users’ needs.—(Hatim & Mason 1990: 8)

These last two items—the specification of the translating task and the users’ needs —as perceived by the translator—can be subsumed under the concept of the purpose for which the translation is undertaken. And the purpose the translator has preset himself or herself will define the kind of equivalence which is aimed at. In some translation tasks the maintenance of cognitive equivalence is of vital importance but the form of the target text is, of necessity, radically different. This would be true of, say, summary translation—a résumé of the foreign press, for example. In other cases, the target text is meant to have an entirely different function from the source text but it is necessary to maintain the cognitive and/or the formal equivalence between the two texts. Translations carried out in examinations fall into this category. Conversely, in some forms of literary translation, cognitive and formal equivalence are sacrificed in the attempt to maintain the pragmatic. On the other hand, when the two texts have what Sager calls ‘‘full equal status’’ the specification of the translation task is to attempt to maintain ‘‘full intentional and cognitive identity [. . .] as well as very close correspondence at the word, phrase and sentence level’’ (1994: 223). Compromises are inevitably made at each level in order to keep all three aims in play.1

3.2 The evaluation of context Having established that, in any act of translation, the translator is seeking to maintain some sort of equivalence, not necessarily on all three levels but on one at the least,2 the problems are still only just beginning, as Sager suggested in the passage above. In particular, how do translators know that what they are producing is an adequate equivalent? Halliday (1992a) attempts to answer this question by stressing the importance of the appraisal of context in translation practice. Translation, he argues, is a ‘‘meaning-making activity’’ with the peculiarity that it is the ‘‘guided creation of meaning’’. He defines meaning itself, after Firth3, as ‘‘function in context’’, and equivalence of meaning is therefore ‘‘equivalence of function in context’’. Now,

true and false friends

51

Linguistics cannot offer a theory of translation equivalence. There can be no such general theory. What it can offer, on the other hand, is a theory of context. (Halliday 1992a: 16)

Looking to explain the decision-making processes of translation by relating them to the system of functional grammar, Halliday argues that the translator is: aware that an item X in the source language has a potential equivalence of items A, B, C, D, E, F in the target language. Note that the term ‘‘item’’ is a general term: it might be a morpheme, or a word, or a phrase, or any other piece of wording in the language [. . .] Secondly, you are aware that these are not free variants, but they are contextually conditioned. By ‘‘contextually conditioned’’ I do not mean that in a given context you must choose A and cannot choose B or C, but that if you choose A or B or C then the meaning of that choice will differ according to what the context is.—(Halliday 1992a: 16)

Halliday equates context with the grammatical concept of rank, or constituent unit. At morpheme level, the choice of one morpheme rather than another in the source language will depend on the meaning/function of that morpheme at the word level. The choice of one word rather than another will depend on the significance of that choice at phrase level. The choice of phrase will depend on its meaning at clause level. Clause level choice depends on the sentence, the sentence on text and finally text level choice will depend on the context of situation, that is, the social and pragmatic significance of the resulting text in its target cultural environment (Ulrych forthcoming). Happily, in the context of this chapter, Halliday gives examples of the first of these processes from English and Italian. When can the Italian morpheme -tore be safely chosen as an equivalent of the English -er? The question can only be answered by looking at word level—writer and driver can be translated as scrittore and conduttore but no such word as pranzatore exists to translate diner. This chapter—a detailed investigation of the behaviour of pairs of ‘‘look-alike’’ items in the related languages English and Italian—is to be considered as a study in functional equivalence at the phrase level; in Hallidayan terms, of how the translator can use knowledge of the phrase context, acquired through concordancing, to refine the choice of word-rank translation equivalents.

3.3 Look-alike words and false friendship The phenomenon of so-called ‘‘false friends’’ is a familiar one to translators. They are defined as: words or expressions which have the same form in two or more languages but convey different meanings. They are often associated with historically or culturally related

52

alan partington languages such as English, French and German, but in fact false friends also abound among totally unrelated languages such as English, Japanese and Russian. (Baker 1992: 25)

They come about in two different ways. They can be loan words or they can exist in two languages because these languages share a common ancestor. Their development in the two languages might take very different paths. As Baker says: ‘‘once a word or expression is borrowed into a language, we cannot predict or control its development or the additional meanings it might or might not take on’’ (1992: 25). She cites the example of the Japanese loan word feminist, which, in that language is used to describe ‘a man who is excessively soft on women’, a far cry from its meaning in English. Similarly if two languages borrow a word from a third, there is no guarantee that they will use it to describe the same thing. A good example would be magazine and the French magasin (meaning ‘shop’) both borrowings of an Arabic word meaning ‘conserve’, ‘store’. Baker discusses false friends as one example of non-equivalence among languages and regards them as a possible problem area for translators and language learners. An interesting point she makes is that if the meanings of the false friends in the two languages are far apart, they are unlikely to cause problems, but if the two meanings, as often happens, have something in common, there is a danger of confusion: Some false friends are easy to spot because the difference in their meanings is so great that only a very inexperienced translator is likely to be unaware of it [. . .] An experienced French or German translator may, however, confuse English sensible with German sensibel (meaning ‘sensitive’), or English sympathetic with French sympathique (meaning ‘nice/likeable’).—(Baker 1992: 25–6)

In other words, few learners or translators have problems with the more obvious false friends, such as (taking examples from English and Italian) camera and camera (‘room’) or stamp and stampa (‘print’ or ‘the press’). But they can have problems with words whose meanings in the two languages overlap. This is the case of words such as sanity (‘mental health’) and sanità (‘health in general’) in which one term is more restricted in meaning than the other.4 Another case is that illustrated by finally and finalmente (meaning both ‘finally’ and also ‘at last’) where one of the two has several senses, only one of which corresponds to the meaning of the other. A third case arises when the two terms have a similar meaning but belong to different registers or have a different cultural valence (for example Baker invites us to compare the use of a term like au fait in English, where it has particular prestige value, with that in French).

true and false friends

53

3.4 The problem for translators: a limit example On occasion, the relationship between the two items can be highly complex. The two nouns sanction and sanzione have a similar meaning in English and Italian of ‘legal punishment’. As for the related verbs, the Cobuild dictionary (1987) defines sanction as follows: If an institution or person with authority sanctions an action or practice, they give their approval of it so that it can officially be done

Three reputable, up-to-date Italian dictionaries which were consulted,5 all gave ‘approve’ as the first sense of the verb sanzionare. However, a concordance list of the forms of the verb sanzionare show clearly that ‘approve’ is not its first sense, at least in the particular Italian corpus which was used (see Introduction: Methodology and Equipment), but ‘penalise’ (that is, more or less the opposite meaning). Of the 20 occurrences of some form of sanzionare, 17 were used with this second meaning, as in the following: (1)

(1a)

La mancata presentazione nei termini di tale relazione non è sanzionata in alcun modo, né sotto il profilo amministrativo, né sotto quello penale, né tantomeno determina l’incommerciabilità dell’ immobile. Tuttavia [. . .] Literally: The failure to present such a report cannot be penalised in any way, either in civil or criminal law, nor does it entail that the property cannot be sold. However [. . .]

The conflict between the dictionary definition and the corpus evidence is probably to be explained in terms of semantic shift. It is probable that sanzionare is a word which is undergoing a (very drastic) change in meaning—perhaps under the influence of the plural noun sanzioni, meaning ‘sanctions’ which is found more and more in the Media. The kind of disaster awaiting a translator who was unaware of this fact, who trusted the dictionary definition of sanzionare and unwittingly rendered in English a word that means ‘penalise’ with a word that means ‘approve’, or vice versa, rendered in Italian a word meaning ‘approve’ with one generally taken to mean ‘penalise’, needs no emphasising.

3.5 Comparing items across languages: correct and corretto As already remarked, two words of similar form from different languages can have a very similar meaning or a very different meaning. Taking examples from

54

alan partington

English and Italian—task and tasca (‘pocket’) are totally unrelated in meaning, whereas impossible and impossibile are very closely related indeed. There is a cline in the relationship of similar-form words running from the one extreme (of false friendship) to the other (of near total reliability). But it is those words which have a similar meaning which cause the greatest problems for translators and learners. In other words, two items which seem to mean much the same thing in two languages—what we might call ‘‘true friends’’—may well in reality be used in subtly different ways and in different contexts in each of the languages, and so there is the danger that the expectations of the non-native user may, on occasion, be betrayed. In the introduction to their exercises on false friends, Aresca & Schmid (1989: 61) first of all give some examples of unproblematic words, that is, English words whose meaning can easily be guessed by Italian speakers, before they go on to point out that one has to be careful in the case of other items—the false friends. The unproblematic words they cite are impossible, correct and romantic.6 Whilst it is true that the general meaning of these items would be easy for Italian speakers to understand, does this mean that it is always safe to translate them with the Italian forms impossibile, corretto and romantico? The second pair of these—correct and corretto—were chosen to be studied by means of their concordances from English and Italian corpora.7 In this way it was hoped to discover how they interacted with their respective linguistic contexts. The first fact to be noted was that corretto in its various forms (corretto, corretta, corretti, corrette) is relatively more common in the Italian texts than is correct in the English. Since corretto can also mean ‘corrected’, care was taken to remove all occurrences of this use from the Italian concordance. Nevertheless, in one and a half million words of Italian, corretto occurs 118 times (equivalent to 80 occurrences per million words (o.m.w.) of text), in four million words of English, correct appears 120 times (or 30 o.m.w.). There is a considerable difference in the collocational behaviour of the two words, as far as this can be compared across two languages. Taking correct first, it collocates (taking a span of three words to the left and three to the right) with response and weight (6 times each), procedures and interpretation (4 times each), entry-ies, answer, time, date and method (3 times each), diagnosis, moment and way (twice each). Included in this list are only those items which collocate with the keyword with a frequency of more than once. The word corretto on the other hand collocated with procedura (‘procedure’: 10 times), interpretazione (‘interpretation’: 7 times), comportamento (‘behaviour’) and prospettiva (‘perspective’/‘prospect’) (5 times each), modo (‘way’ /‘method’: 5 times), operazione (‘operation’: 4 times), uso (‘use’) and gestione (‘handling’ /‘management’) (3 times each). The only lexical overlaps in these corpora, then, are

true and false friends

55

the items interpretation, procedure(s) and way/method with their Italian counterparts interpretazione, procedura/e and modo. From this evidence, it is possible to hypothesise that a good number of the occurrences of correct are unlikely to be rendered in translation by corretto. In some cases this is clearly due to the fact that correct and corretto are involved in different expressions, for example correct weight is rendered by peso forma (‘fitness weight’), and correct entry would probably be translated by the phrase primo estratto (‘the first to be drawn’) which takes for granted the fact that non-correct entries are ignored. In other cases the picture is less clear cut. In the English material the lexicosemantic field of ‘answer’,‘response’ is well represented, but is totally absent from the Italian material. The behaviour in the Italian corpus of the word risposta—the item given as translation of both response and answer in the Paravia’s (1989) bilingual dictionary—was examined in order to discover exactly what it does collocate with. Although it occurred 270 times in singular and plural forms, it never collocated with corretta/e. It did, however, collocate with adeguata (‘adequate’: twice), concreta (‘concrete’: twice), precisa (‘precise’) as well as with positiva, negativa and affermativa (several times each). Oddly, there was no occurrence of either risposta esatta or risposta giusta which a number of Italian native speakers referred as their preferred translation of correct answer/response in the contexts found in the corpus. Correct collocates with a number of items from the lexico-semantic set of ‘time’ —the words time, date and moment. The phrase correct time is functionally equivalent to l’ora esatta (literally, ‘the exact hour’). As far as moment is concerned, there were no occurrences in the Italian corpus of momento corretto, but there were 12 occurrences of the phrase momento giusto which would seem to have an equivalent meaning. There were no examples of data corretta (equivalent to correct date) but there was one occurrence of data esatta although this might have a slightly different meaning (compare correct date with exact or precise date). It would seem from this evidence, then, that the Italian item corretto does not collocate with words from the lexico-semantic set of time, but that other items from the same semantic field, like giusto and esatto, do collocate with such words. If we now look at those items that corretto collocates with, we find still more differences in behaviour between the supposedly ‘‘good friends’’ corretto and correct. Corretta collocates a number of times with prospettiva, whereas the English word is not found at all in the corpus with perspective. On the other hand perspective is found to occur with proper and true (twice each). Corretta also collocates with gestione, whereas neither handling nor management are found to occur with correct. Handling however is found with deft, good and mature, whilst management collocates with prudent (7 times), competent and proper (3 times each). It is neces-

56

alan partington

sary to treat the data with some caution at this point. Native speakers of English would be unlikely to classify phrases such as correct handling or correct perspective as unacceptable. On the other hand, the corpus evidence would suggest that there are other collocations which are preferred. In contrast, corretto with gestione and prospettiva seem to be highly preferred collocations. The choice of most suitable equivalent is also a question of the awareness of such stylistic phenomena. Perhaps the most revealing common collocate of corretto is comportamento (‘behaviour’). Although appearing in the (four-million word) corpus nearly 200 times, behaviour collocates only once with correct (in a sporting environment). In comparison corretto/i co-occurs five times with comportamento/i which occurs 170 times in the much smaller (1.5 million word) Italian corpus. These results may have come about entirely by chance, but other evidence suggests that the Italian word is more often applied in the sense of ‘morally appropriate’ that is the English version, even though this is still one of its minor senses—and this would account for the higher incidence of the Italian word with words like comportamento. Here are some examples: (2) (3) (4) (2a) (3a) (4a)

assicurando ai tedesco-orientali corrette chance di convivenza con i loro fratelli occidentali [. . .] aggiungendo però che naturalmente i Centri faranno una corretta concorrenza alla rete tradizionale di distribuzione. Noi vogliamo dialogare con i giornalisti ma vogliamo anche che l’opinione pubblica sia informata in modo corretto. ensuring that the East Germans have a proper chance of living together with their Western brothers [. . .] adding however that the Centres would naturally offer fair competition to the traditional distribution networks We want dialogue with journalist but also that the public opinion be informed in a way that is right and proper.

In each case the translation of corretto is unlikely to be correct.

3.6 A further experiment: absolutely, completely, entirely and their Italian look-alikes From the above, we can see how even words of similar form considered to be excellent friends are not always reliable translation equivalents. The English item correct is often better rendered by giusto or esatto which would, on appearance, seem to be true friends of just and exact. Conversely corretto, as we have seen, is often best translated by right, proper, fair or some other item from the same lexico-semantic

true and false friends

57

set. Indeed, it seems probable that this phenomenon is quite common since pairs of related languages frequently both have a number of items which belong to a particular lexical field but the match between the senses of the items across the two languages is not a perfect overlap. As a further example, the overlap in sense and collocational behaviour among another set of items was examined: the group of English intensifying adverbs absolutely, completely, entirely, and their Italian lookalikes assolutamente, completamente, interamente. A subset of adverbs was chosen because, as Johansson points out, their collocational behaviour can be particularly difficult to fathom:8 Adverbs are no doubt the most heterogeneous of the traditional word classes. Syntactically, the patterns of co-occurrence are less marked than for other classes of lexical words [. . .]—Johansson (1993: 46)

Of these, assolutamente and absolutely proved the most interesting couple. That they are not always translation equivalents becomes clear in the translation of the following text of an interview published in Il Sole 24 Ore. Towards the end of the interview, the head of an American company dealing in ‘‘crisis management’’ is asked where he thinks future demand for his company’s product will grow most quickly. He replies that when Western companies take over East European ones and begin laying off workers, they will have to call in ‘‘crisis managers’’ to assuage local opinion: (5)

Quando una multinazionale come General Electric acquista la Tungsram in Ungheria, deve gestire un piano di licenziamento a cui dipendenti e l’opinione pubblica non sono assolutamente preparati (my italics).

(5a)

Literally: When a multinational like General Electric takes over Tungsram in Hungary, they need to organise a redundancy plan for which the employees and public opinion??

For which the employees and public opinion are not absolutely prepared, which is the literal rendering, would be a serious mistranslation. A version containing is totally unprepared or is not ready at all would render the sense of the Italian text. To investigate the degree of mismatch between the look-alikes absolutely and assolutamente, concordance lists of the two words were prepared, along with those for completely and entirely and those for completamente and interamente. It was immediately noticeable that absolutely was the only one of the three English items to collocate regularly with adjectives having some kind of in-built superlative sense —that is, items expressing some kind of very strong attitude or opinion (and which we might call ‘‘hyperbolic adjectives’’ in parallel to the class of hyperbolic nouns mentioned in the previous chapter). This function of absolutely has already been

58

alan partington

noted (Altenberg, 1991: 137). It collocated with appalling (twice), gutted (3 times), delighted (2), outstanding (2), hilarious, disgraceful, fantastic and bloody hopeless. Neither completely nor entirely collocated with any frequency with hyperbolic adjectives. Another group of lexical items collocating with absolutely belonged to the semantic field of ‘highly important’. These included essential (5 occurrences), crucial (3) and vital (2) none of which co-occurred in the corpus with either of the other adverbs. The related item necessary was found with absolutely on 19 occasions. The only other intensifier to premodify it was entirely on two occasions. One more important sub-group seemed to consist of items expressing ‘certainty’/‘clarity’, including certain (2 occurrences), sure (1) and clear (9). The first two usually have personal subjects whilst the latter is often used when the subject is impersonal as in the following example from the corpus: (6)

It is important that the tasks agreed on are absolutely clear.

Other particular features of this intensifier include the following. It is the only one amongst those studied here to intensify the words nothing, no and not (Altenberg, 1991: 137) with 18, 5 and 7 co-occurrences respectively. There was a reasonable balance between ‘‘favourable’’ and ‘‘unfavourable’’ items amongst absolutely’s collocates e.g. delighted, marvellous and wonderful etc as against shocking, intolerable and appalling and so on. Two important lexicosemantic categories, those expressing the absence of something—a quality, object or behaviour (such as, for example, devoid, lacking and empty) and those expressing a change of state (e.g. different, altered, rebuilt and perhaps new) are very thinly represented among the collocates of absolutely in the corpus. (These types of words, on the other hand, made up a much higher proportion of the collocates of both completely and entirely —see below). From the evidence of the Italian corpus, assolutamente can very occasionally be used to premodify superlative items. It co-occurred with catastrofiche (‘disastrous’) and eccezionale (‘exceptional’, ‘wonderful’). However, these were the only examples in the entire corpus. We can conclude that this use is much rarer than is the case of its English counterpart. Moreover, assolutamente was found to collocate very frequently with adjectives which are non-hyperbolic interno (‘inside’), protetti (‘protected’), fisiologici (‘physiological’), secondaria (‘secondary’, ‘minor’), modesto (‘average’, ‘unexceptional’), estranea (‘extraneous’, ‘unrelated’), invenduti (‘unsold’) and many others. On this evidence, assolutamente does not have the primary function of intensifying hyperbolic items, as is the case for absolutely. Assolutamente was found to premodify items expressing ‘necessity’, which is one of the subgroups absolutely was found to co-occur with. The Italian corpus included

true and false friends

59

assolutamente necessaria and assolutamente prioritaria (‘absolutely essential’). It was also found as an intensifier of no, non and nulla, roughly equivalent to English no, not and nothing. However, there are differences in use between the two languages. It would be unusual to find, in English, absolutely no where no was the answer to a question rather than a determiner, as in no food in the house. It is, on the other hand, quite usual to find assolutamente with no (= ‘no’) in answer to questions. There were two examples of this in the Italian corpus: (7) (8)



,

(7a) (8a)

‘‘Do you regret your choice?’’ ‘‘Absolutely . . . (no).’’* ‘‘Isn’t there also a problem about the quality of the product—that the German one is better?’’ ‘‘Absolutely . . . (no).’’*

It is even possible for it to intensify si (‘yes’): (9)

>—

(9a)

Literally: ‘‘Are you satisfied with the result you got in the end?’’—‘‘Absolutely . . . (yes)’’*

The other sub-group found to collocate frequently with absolutely consisted of items expressing ‘certainty’/‘clarity’. The following is a list of all the examples of absolutely with certain, clear and sure from the corpus: 1 was `very good for the saver. I'm absolutely certain the old rules discrimi 2 n in itself because she's been so absolutely certain of herself. I've found 3

qualifying. `But we have made it absolutely clear that there is no questio

4 scheme given in the Preface seems absolutely clear and definite, this is no 5 teaching of the church is seen as absolutely clear and devoid of problemati 6

for several days. I have made it absolutely clear from the outset that I w

7 nvolved in the patient's care are absolutely clear about the current policy 8 Mr Brown said: `The rules make it absolutely clear that public funds could 9 rs and the British government are absolutely clear that we will not allow A 10 t that the tasks agreed upon are absolutely clear. This make misunderstand 11 ting up. Not a bit of it. He was absolutely sure about it. `My greatest re

There is evidence in the Italian corpus that assolutamente can also be used to premodify ‘certainty’ words. It appeared with certo (‘certain’), detto (literally ‘given’, ‘certain’, ‘to be taken for granted’): (10)

e, secondo il portavoce della Bundesbank, non è assolutamente certo che i governatori riescano ad eliminare, nel corso della settimana, le divergenze [. . .]

alan partington

60 (11)

Ciò significa che in presenza di un’ elevata situazione di liquidità non è assolutamente detto che la Banca d’ Italia sia a priori disponibile a drenarla [. . .]

(10a) and, according to the Bundesbank spokesman, it is not ?? that the governors will manage, in a week, to eliminate the differences . . . (11a) Which means that, in a situation of high liquidity, it is not ?? that the Bank of Italy is necessarily willing to drain it . . .

In both cases the phrase contains the negative particle non. In complete contrast, not a single one of the phrases containing absolutely plus certain/clear etc. listed above had a negative sense. If, however, the corpus is searched for the items clear, certain and sure when premodified by an intensifier in -ly, the following examples are found with entirely: 1 which is fighting the bid, is not entirely clear to the City. All City insu 2 thousand years ago was never made entirely clear but genetic studies of pop 3 ulously depilated show. It is not entirely clear to me why the hairremovers 4 sed that even these facts are not entirely clearcut. Last week, Mr Cornes a 5

worth taking, though she was not entirely certain what she could gain from

6 y: The Movie. Police can never be entirely sure that a criminal will be tem

There are no examples of entirely collocating with one of these items in a nonnegative expression. It would seem likely then that, to translate a phrase like non è assolutamente certo, a version using entirely would be more suitable than one with its look-alike absolutely. In general, the corpus contained many more examples of assolutamente being used in a negative expression than absolutely. In particular, the former collocated frequently with adjectives which contain some negative particle, chiefly in-, im-. It was found to premodify impensabile (‘unthinkable’), impreparati (‘unprepared’), inadeguato (twice: ‘inadequate’), incapace (‘incapable’), inderogabile (lit. ‘not dispensible’, ‘imperative’), ineccepibile (‘faultless’), inedite (lit. ‘unpublished’, ‘novel’, ‘unique’), inefficiente (‘inefficient’), invenduti (‘unsold’), scriteriato9 (‘injudicious’) and estraneo (twice: ‘extraneous’, ‘unrelated’). In addition, there were a number of phrases with the pattern ‘‘non + assolutamente + adjective’’, for example: (12)

proprio per questo molti programmi, anche di notevole valore commerciale, non sono assolutamente protetti

and the phrase we have already come across: (13)

un piano di licenziamenti a cui i dipendenti e l’ opinione pubblica locale non sono assolutamente preparati.

true and false friends

61

which would probably be translated into English by an intensifier + negative adjective (in these case, unprotected and unprepared). In all, about 30 per cent (20 out of 62) of the occurrences of assolutamente were found to be of the types described in this paragraph. In contrast, only 5 of the 155 occurrences of absolutely collocated with an adjective containing a negative particle (im-, in-, ir- or un-). When, however, completely was examined, a much larger proportion of such items among its collocates were unearthed: 31 out of 200. On this evidence, completely may often be a better translation for the Italian word, in this kind of negative phrase, than is absolutely. For instance, the two sentences containing assolutamente reported above might well be translated as follows: (12a) for this very reason, many programmes, even ones of considerable commercial value, are completely unprotected (13a) a redundancy plan for which the employees and local public opinion are completely unprepared

On the subject of negative construction, the item entirely, on the other hand, was found with great frequency in the phraseology ‘‘not/never + intensifier + adjective’’, for example: (14) (15)

He was already drunk and not entirely pleased to be visited by journalists Football was never an entirely respectable sport . . .

Entirely appeared with these negative particles 66 times in 250 occurrences, compared to completely’s 25 out of 200 and absolutely’s 6 out of 155 (this figure does not count the three occurrences of absolutely not where the negative particle follows the intensifier). Moreover, whenever entirely was found in the vicinity of an adjective which is strongly positive in sense, it was usually possible to predict that the phrase construction would be negative—in other words, it rarely intensifies ‘‘good’’ qualities but often negates them, as in the following (the only three examples of entirely happy in the corpus): 1 credits, for several years. Never entirely happy as a desk man, he ab aband 2 mpatible. Few performers would be entirely happy touring the country as, sa 3

to whom it is unavailable is not entirely happy. Locke's belief in the rel

In contrast, the Italian word interamente appeared with a negative only once out of 35 occurrences (by sheer coincidence modifying corretto)—non è interamente corretto. This, along with the sharp difference in relative frequency of the two items, may be evidence that once again there is a less than perfect match between the English and Italian words. Again, for the following corpus example:

alan partington

62 (16)

Kolesnik afferma che seguire il destino della famiglia di Stalin significa capire più profondamente gli eventi di quel periodo. Non ci sembra. Il quadro è già interamente definito.

(16a) Kolesnik claims that tracing what happened to Stalin’s family means being able to understand better the events of that period. I don’t agree. The picture is already ??

A possible translation of definito is clear or well-defined. In any case, it is possible to assign the adjective to the semantic class of certainty, whose members, as we have seen before, do not often collocate with entirely in a non-negative phrase. Absolutely, which does so, would be a more suitable alternative in this context, and so The picture is already absolutely clear would be an acceptable version in the above case. One final observation. Mention has already been made of the fact that both completely and entirely were seen to collocate frequently with words expressing ‘absence’ and those expressing a ‘change of state’. A selection of collocates of completely expressing ‘absence’ from the Independent corpus includes: devoid, drained, empty, dry, hollow, forgotten, lost, ignored, omitting, stripped, naked, anonymous, bald, painless, symptomless and valueless. Those expressing ‘change’ include: alter, change, changed, different (20 times), new (7 times), rebuilt, reformed, reorganized, revamp, rewriting and transformed. A few which include both concepts are destroy, demolish and collapse. The lists for entirely are similar. ‘Absence’: absent, blameless, deprived, devoid, eliminate, free of (twice), forgotten, missing and without (5 times); ‘change’: different (6), distinct (2), new (13), recomposed, reorganised and restored. These are both semantic fields with whose members absolutely collocates very rarely. In contrast, assolutamente collocates with the ‘absence’ items libero da (‘free from’) and invendute (‘unsold’) and the ‘change’ words diverso (‘different’), riabilitato (lit. ‘rehabilitated’, but here ‘restored’) and nuovo (twice, ‘new’). A suitable translation of phrases containing these items would probably not contain absolutely, but either completely or entirely.

3.7 Conclusion It is not hard to envisage in the near future a computer system dedicated to translators, in which corpora in various languages are held on CD-ROM and are in a form permitting direct access by a concordancer. Multilingual databases have long been used in research (see for example the Council of Europe Multilingual Lexicography Project (Sinclair 1991b)) and similar resources should soon be available for personal computing. These language corpora might also be usefully divided into various subcorpora each containing texts of a particular genre or text type, e.g. medical texts,

true and false friends

63

legal texts and so on (for the design and use of such sub-corpora see Mparutsa et al. (1991) and Zanettin (1994)). This would permit a direct comparison of a particular genre in one language with the same one in another. The studies reported in this chapter show how even words of similar form considered to be excellent friends are not always reliable translation equivalents. Indeed, it seems probable that this phenomenon is quite common since related languages frequently both have a number of items which belong to a particular lexical field but the match between the senses of the items across the two languages is not a perfect overlap. Put another way, there is tentative evidence to suggest that the number of wholly reliable true friends between even closely related languages is probable fewer than is generally imagined. This helps to explain why bilingual dictionaries, which, for reasons of space, tend to pay little attention to context, are often of limited use to translators, or even lead them astray. Moreover, if we bear in mind Baker’s point, quoted above, that words which enter a language take their own independent paths and we cannot predict the way they will change, then this relative paucity of totally reliable look-alikes should come as no surprise. And while this fact may on occasion cause problems for the translator, it is evidence of the vitality of natural languages, of how linguistic systems are self-regulating organisms, which, over time, adapt to suit changing environments.

Notes 1. The situation is not always as bleak as this sounds. The various levels of equivalence are, of course, not independent of each other, in practice each one tends to back the other up. Maintaining formal equivalence, for example, tends to facilitate the maintenance of cognitive equivalence. In other words, matching the form of a source language structure with its natural equivalent in the target language generally enables the translator to maintain the sense, though there are usually many exceptions in the course of any translation. 2. Providing the translator is honest. There is the limit case of what might be called ‘‘propaganda translation’’ such as that found in newspapers where journalists report the words of some foreign politician they or their political patrons find disagreeable. In these cases, the function of the ‘‘translated’’ text, to discredit the source, is certainly in contrast with that of the original. The cognitive content may well be deliberately manipulated and the form will tend to alter too—the second language version is usually a summary. However, some attempt at cognitive equivalence will have to be made, and where it is not, it can be argued that we are no longer dealing with translation but free variation, or less kindly, invention. 3. See Linguistic Analysis and Translation in Palmer (ed), 1968. 4. There are a number of private chemist shops in Italy which misleadingly call themselves ‘‘Sanity Center’’.

64

alan partington

5. Zingarelli N. 1991. Bologna: Zanichelli. Devoto G. e G. Oli 1991. Firenze: Le Monnier. Palazzi, F. and G. Folena 1992. Torino: Loescher. 6. ‘‘L’inglese e l’italiano hanno molte parole che si somigliano e di cui si può facilmente ‘indovinare’ il significato, per esempio impossible, correct, romantic e così via. Bisogna però stare attenti a non fidarsi sempre delle parole che ci sembrano familiari […]’’ (Aresca & Schmid 1989: 61). 7. Only the newspaper sections of the English corpus (four million words) were used in order to facilitate comparison with the Italian texts, which are all from the newspaper. 8. The collocational behaviour of various types of adverbs, but especially intensifiers, has been much studied, before and after the advent of corpora. See Greenbaum (1970, 1984), Backlund (1973), Allerton (1987), Altenberg (1991), Partington (1992), Johansson (1993). 9. The initial s of this word is a negative particle deriving from the Latin ex-.

4 Connotation and Semantic Prosody Mr Hunt [. . .] retaliated by accusing the Opposition of peddling rumour, innuendo and smear (The Daily Telegraph, 1993).

4.1 Connotation The two previous chapters have been principally concerned with showing how concordance data can give insight into the conceptual (or ‘‘denotational’’)1 meaning of words by revealing their context of use, that is, how they behave in their phraseological environment. Another aspect of meaning which can emerge from a study of context of use is the connotational significance of lexis. The term ‘‘connotation’’ (often in the plural: ‘‘connotations’’) has been defined in several ways, e.g. ‘‘the vaguer associations of a word for a group or individual’’ (Cook 1992: 8); the ‘‘secondary implications’’ of an item (Lyons 1977: 278). In traditional semantics, connotation is usually seen as ‘‘lying outside the core meaning’’ (Backhouse 1992: 297), and even in some way ‘‘incidental to language rather than an essential part of it’’ (Leech 1974: 15), since connotative meaning is not specific to language but is also shared by visual art, music and even—some would say especially—smell. It is therefore an important concept in those branches of macro-linguistics (socio- and psycholinguistics, for example), which attempt to analyse the relationships between language and the real world. In particular, since it refers to ‘‘various aspects of the communicative value of linguistic terms’’ (Backhouse 1992: 297), an awareness of the connotational value of lexis is a vital part of the communicative competence (Hymes 1971) of a speaker—the knowledge of what is the right thing to say at the right time in the right circumstances. However, the term connotation is used to refer to at least three distinct phenomena. In the first, markers of particular speech varieties have what Backhouse calls social or situational connotations. Class, regional origin, age, sex, and relationship between speakers may all be connotated by particular lexical or grammatical choices. For example, an expression such as absolutely awful is readily identifiable as belonging to an upper-middle class variety. Beefy is more colloquial than robust,

66

alan partington

and the choice by a speaker of one rather than the other will reflect the formality of the register he or she is using. Connotation may also be a matter of what a lexical item denotes within a culture. Leech (1974: 15) points out how the word woman has in the past been burdened with such attributes as ‘‘frail’’, ‘‘prone to tears’’, ‘‘cowardly’’ and ‘‘irrational’’, along with rather more positive qualities such as ‘‘gentle’’, ‘‘compassionate’’ and ‘‘sensitive’’. Such cultural connotations are clearly liable to modification as a society’s values change. Lastly, choosing to use certain lexical items implies a favourable or unfavourable evaluation by the speaker towards what they describe. This is termed expressive connotation. Such choices tend to be highly personal and relatively more voluntary than those involved in social connotation. It would be difficult to sustain the argument that this kind of connotation lies outside the core meaning of the lexis. A word like pig-headed only exists because it has an expressive connotation of disapproval, there is nothing ‘‘secondary’’ about this implication. Similarly, the sole purpose of the term venerable is to put old age2 in a good light, and that of callow to express disapproval of youth. Backhouse’s paradigm—‘‘I am firm, you are stubborn, he is pig-headed’’—is a particularly clear example of expressive connotation. Distinctions between the connotations of terms with similar denotational meanings are, however, often more delicate and difficult to grasp than this. Connotation is thus a problem area for learners. And since, as was said, it is an important mechanism for the expression of attitude, it is of paramount importance that learners be aware of it in order to grasp the illocutionary intent of messages they receive. In this section we look at how the contextual clues inherent in corpus data can reveal some of the subtle ways in which expressive connotation is used. Connotations are generally taken for granted among native speakers, and when connotation is as intrinsic to a word as it is with pig-headed or callow, it is often a question of luck whether it will be revealed through corpus data. There is, however, one particularly subtle and interesting aspect of expressive connotation which can be highlighted by corpus data. This is the phenomenon of semantic prosody.

4.2 Semantic prosody Often a favourable or unfavourable connotation is not contained in a single item, but is expressed by that item in association with others, with its collocates. A clear example of this is the word commit which, as the following concordance sample shows, collocates with items of an unpleasant nature. There were 41 citations altogether in the combined corpus, and every fifth one is reproduced here:

connotation and semantic prosody

67

1 with grievous bodily harm if they commit a deliberate foul resulting in ser 2 n if he is committing or about to commit an act likely to endanger life, an 3 eception and one of conspiracy to commit deception. The deception charges r 4 of those who drink alcohol do not commit offences of violence thereafter. B 5 il for an IRA bombing she did not commit, or the accused in the MatrixChurc 6 offence of hacking with intent to commit serious crime _ including all seri 7 ill eventually starve to death or commit suicide. Although skeletally thin 8 e perfect environment in which to commit these offences,' Detective Inspect

The unfavourable connotation can be seen to reside not simply in the word commit but over a unit consisting of commit and its collocate (offences, serious crime, foul, etc). Another word which has an unfavourable semantic prosody is the adjective rife. There were 24 citations of this word in the corpus, of which every third one is listed here: 1 edy, rundown areas where crime is rife and the misery of unescapable povert 2 nd other deficiency diseases were rife. He stayed with Glaxo all his workin 3

director. Uncertainty is already rife. Several months ago, when rumours ci

4 my void of privatisation laws and rife with corruption. Delegates also vote 5

is more popular, so mistakes are rife. The change has been gradual since t

6 ripe for takeover: Speculation is rife in the building society industry. Pa 7

group last month. Speculation is rife that it will soon sell the reconstin

8 racy: Votebuying and violence are rife as Thais prepare for polling day, wr

The term ‘‘semantic prosody’’ was first coined by Sinclair to describe this phenomenon. He writes (1987b; 1991a) that the items happen and set in are habitually associated with unpleasant events. Of set in he comments: The most striking feature of this phrasal verb is the nature of the subjects. In general they refer to unpleasant states of affairs. Only three refer to the weather; a few are neutral, such as reaction and trend. The main vocabulary is rot (3), decay, ill-will, decadence, impoverishment, infection, prejudice, vicious (circle), rigor mortis, numbness, bitterness, mannerism, anticlimax, anarchy, disillusion, disillusionment,slump. Not one of these is desirable or attractive.—(Sinclair 1987b: 155–6)

The result of keeping this ‘‘bad company’’ is that the use of the item set in is enough, in itself, to signal that some undesirable process is being described. Moreover, since the item is imbued with an ‘‘unfavourable prosody’’, it cannot, in normal circumstances, be used in a favourable environment. A phrase like good times set in would be a highly marked, probably humorous use. Similarly, Stubbs (1995), who analyses 40,000 examples across 120 million words from the Cobuild corpus,

68

alan partington

shows how the lemma cause has a strongly unfavourable prosody in that its collocates are overwhelmingly unpleasant, including accident, concern, damage, death, trouble. The word provide, on the other hand, had a favourable prosody in the Cobuild corpus material, collocating with the semantic fields of ‘care’, ‘food’, ‘help’, ‘money’ and ‘shelter’. The term ‘‘prosody’’ is borrowed from Firth (1957), who uses it to refer to phonological colouring which spreads beyond segmental boundaries. Semantic prosody refers to the spreading of connotational colouring beyond single word boundaries. In an article entitled Irony in the text or insincerity in the writer?: The diagnostic potential of semantic prosodies, Louw (1993) investigates how writers sometimes diverge from ‘‘the expected profiles of semantic prosodies’’, that is, how they upset these normal collocational patterns. He suggests that, when they do so consciously, it is usually with ironic intent. He quotes an example from Small World by David Lodge: The modern conference resembles the pilgrimage of medieval Christendom in that it allows the participants to indulge themselves in all the pleasures and diversions of travel while apparently bent on self-improvement.—(Louw 1993: 164)

Citing data from the Cobuild corpus, Louw shows that bent on usually collocates with unfavourable items—destroying, harrying, mayhem. In upsetting its semantic prosody, making it collocate with the evidently favourable self-improvement, Lodge is here searching for an ironic effect. Louw also argues, in a section entitled Prosodic clash and possible insincerity, that writers can also diverge from a prosody by accident, in which case the reader may detect a difference between what the writer is apparently saying and what he/she really believes. He cites the example of a speaker who is apparently praising the professional standards of the University of Zimbabwe as follows: ‘‘it is symptomatic of the University of Zimbabwe which has such a high reputation that there are fifteen links between departments in the University here and equivalent departments [. . .] in Britain’’ (1993: 169). Corpus data shows that the prosody of symptomatic is heavily unfavourable, which suggests that the speaker may not privately think so highly of the University he is discussing. Louw suggests that looking out for clashes between private use and normal prosodies may help in uncovering the secret aims of advertisers or propagandists. Finally, Louw also suggests that lexicographers have not, in the past, been able to deal satisfactorily with semantic prosodies, because they have tended to remain ‘‘hidden’’ to the lexicographer’s ‘‘naked eye’’. The modern fusion of corpus linguistics and lexicography is making the study of prosodic profiles (and how they are exploited by writers) feasible:

connotation and semantic prosody

69

Semantic prosodies have, in large measure and for thousands of years, remained hidden from our perception and inaccessible to our intuition3 [. . .] At present (computer-held) corpora are just large enough to allow us to extract profiles of semantic prosodies from them. From these prosodies we may then extract the narrow band of irony and insincerity they contain.—(Louw 1993: 173)

Later in this section, an attempt is made to investigate Louw’s claims in greater detail, to isolate other examples of the exploitation of semantic prosody and to analyse precisely how writers manipulate them to achieve particular perlocutionary effects. Before this, however, in the light of Louw’s claim that lexicographers in the past have not been fully aware of the extent of semantic prosody but that modern corpora provide new opportunities of studying the phenomenon, a brief examination is carried out into how this aspect of meaning is treated by comparing two groups of dictionaries (both learners’ and non-learners’), one dating from before, one after, the advent of computerised corpora. The investigation makes use of data from both the corpus and from The Independent (1992) and The Telegraph (1993) CD-ROMs (see Introduction).

4.3 Semantic prosody and dictionaries The following non-learners dictionaries were included in the study: The Merriam Webster New Universal (MWNU 1967 edition) Webster’s New World Dictionary, (1988) The Shorter Oxford English Dictionary (SOED 1993). Two sets of learners’ dictionaries were also examined. The first set consisted of dictionaries produced at the end of the 1980s: The Collins-Cobuild (Cobuild 1987) The Longman Dictionary of Contemporary English (LDOCE 1987) The Oxford Advanced Learner’s (OALD 1989). The second set consisted of learners’ dictionaries all published in 1995: The Collins-Cobuild 2nd edition (Cobuild 1995) The Cambridge International Dictionary of English (CIDE) The Oxford Advanced Learner’s Dictionary—new edition (OALD 1995) Longman Dictionary of Contemporary English (LDOCE 1995).

70

alan partington

These last four were all compiled with the aid of large computerised corpora of texts. Both LDOCE 1987 and the first edition of Cobuild 1987 were also produced with the aid of corpora,4 but the 1995 dictionaries were all built on the basis of relatively much larger ones.5 The first item examined was set in which, it will be recalled, according to Sinclair had a most unfavourable prosody in the Cobuild corpus. None of the non-learners’ dictionaries give any explicit reference to the prosodic behaviour of the item6 and nor does OALD 1989. On the other hand, both LDOCE 1987 and, not surprisingly, Cobuild 1987, of which Sinclair is the general editor, do give such indication: If something unpleasant sets in, it begins and seems likely to continue—(Cobuild 1987)

All the 1995 learners’ dictionaries make more or less explicit reference to this item’s unfavourable prosody. OALD 1995 has added the indication ‘‘of rain, bad weather, infection etc’’ after the headword. Another item felt intuitively to have a highly unfavourable prosody is the verb peddle and its related nouns pedlar and peddler (the second of these is usually referred to as a U.S. equivalent of the first), when these words are used in a metaphorical way, that is, not referring to an ‘itinerant hawker’. The behaviour of these words was examined first of all in the combined corpus. There were thirteen occurrences of these words, of which every other one is listed below: 1 ubt that IMG has a reputation for peddling all four sides of the square. Th 2

s Page 2 IND 06 MAR 92/Sinclair peddles electric innovation: Susan Watts

3 ter to do with their time than to peddle mischievous insinuations which sim 4

Theatres began to fold, the porn peddlers moved in, new names hit the stre

5 ferent. What consolation will the pedlars of these cynicallymanufactured cr 6 illegal payments from an influencepeddling scheme by his campaign treasurer 7 ated businesses has been actively peddling the investment bank for the past

Peddle was also found to appear almost exclusively in highly unfavourable environments in the CD-ROM texts of The Telegraph (1993) and The Independent (1992). Below are a few examples: (1) Mr Hunt, the recently-appointed Employment Secretary, retaliated by accusing the Opposition of peddling rumour, innuendo and smear. (Telegraph) (2) Mr Hart said the results gave the lie to claims that educational standards were declining and should silence those who peddled doom and gloom. (Telegraph) (3) I am surprised that senior Opposition politicians can find nothing better to do with their time than to peddle mischievous insinuations which simply distract attention from the important issues of the day. (Independent)

connotation and semantic prosody

71

Of the 1980’s learners’ dictionaries, only LDOCE 1987 indicates ‘‘usu derog’’ i.e. usually derogatory. The others give, as the first sense of the word, some variation on ‘‘to sell goods in small quantities’’ (OALD 1989). In the newspaper texts, as has been seen, this is not the most common use, but dictionaries have a tendency to place concrete uses before more abstract ones.7 The second, metaphorical, sense is defined by OALD 1989 as ‘‘advocate or promote (ideas, a philosophy, a way of life)’’. Cobuild 1987 says of peddler ‘‘someone who is a peddler of particular ideas, often expresses these ideas to other people’’. However, such a definition seems to wholly miss the point. The items peddle and pedlar (peddler) are employed precisely in order to avoid the neutral or perhaps even favourable colouring of ‘‘express’’, ‘‘advocate’’ or ‘‘promote’’, words which seem to convey a touch of gravitas to whatever idea is being mooted. The use of peddle tends to tell the reader at least as much about the opinions and beliefs of the text producer as about the actual topic, implying their opposition or even hostility to someone’s ideas: (4) This venal piece of political opportunism peddled by Mr Waldegrave, in which citizens are told they have rights without responsibilities, has led to a moral crisis within the profession. (Telegraph) (5) look at the German National Socialist Workers’ Party (NSDAP), whose early literature was as full of bourgeois-bashing anti-capitalism as anything peddled by the Socialist Workers Party today. (Telegraph) (6) but at least it is a useful antidote to the contrary view still widely peddled in academic and feminist circles that anyone who dares to suggest that there might be substantial differences in skills and attributes between the sexes [. . .] (Telegraph)

Finally, an example which indicates in very explicit terms that this word has an unfavourable semantic prosody: (7) Sometimes he touched an unexpected note of wry, self-deprecating irony. ‘‘. . . one of the problems that I face, as someone who peddles hope, is the presumption against one’s credibility and integrity.’’ Someone who peddles hope? He wasn’t confessing cynicism so much as modestly doing himself down, after the English fashion. (Corpus)

Three of the four new 1995 dictionaries include some explicit indication of the unfavourable character of this word. Cobuild 1995 definitionsinclude what it calls ‘‘pragmatic information’’, i.e. indications about the conditions of normal usage of entries. In the case of peddle, the user is told that this item is ‘‘used showing disapproval’’. CIDE tells the user that this word is ‘‘esp. disapproving’’. OALD 1995, on the other hand, gives no explicit reference to its prosody, its definition being simply ‘‘to spread or promote sth., eg an idea or a rumour, in an attempt to get it accepted’’. However, the examples it supplies show the word in a highly

72

alan partington

unfavourable environment: ‘‘peddling malicious gossip’’, ‘‘peddling his crazy ideas’’. Strangely, LDOCE 1995 no longer uses the tag ‘‘usu derog’’, and there is no concrete indication of the word’s unfavourable prosody. The non-learners’ dictionaries were, without exception, silent about the prosody of peddle. Louw’s hope was that prosodies should begin to receive their just attention from dictionary makers. The evidence from set in and peddle suggests that progress has been made as far as learners’ dictionaries are concerned. On the other hand, there is no evidence here that semantic prosody is included for consideration in non-learners’ dictionaries, even the most modern ones.

4.4 Prosodies and persuasion: sharp dealings The lexical items mentioned so far display highly regular prosodies. Other items, however, may have much less apparent prosodic behaviour, that is to say, they have a particular colouring on many occasions but not on all. Sinclair found this to be the case with the word happen which shows a tendency to collocate with unpleasant events but this characteristic is not binding, it is occasionally found to collocate with neutral or event pleasant occurrences. When the semantic prosody of an item is not deterministic, as with happen, it may not be apparent even to a native speaker’s intuition, but corpus data may show up its statistical tendencies. The behaviour of the word dealings is analysed below. It could be contended that if the semantic prosody of an item is not deterministic, then speakers and writers are free to use it as they will. However, an awareness of what are marked and unmarked uses of lexis is extremely valuable. Moreover, as Louw argues, a text receiver needs to utilise the knowledge of semantic prosody to perceive irony in the message or to judge the degree of sincerity of its proponent. Access to information on prosody from the corpus is particularly important for nonnative speakers, since they are both more liable to miss such irony and to be more vulnerable to the hidden intentions of the producer than native speakers, who probably have some sensitivity to it at a subconscious level. The dictionaries give no explicit indication that dealings has any particular colouring. The LDOCE 1987 definition is ‘‘personal or business relations’’. Cobuild (both editions) says ‘‘Someone’s dealings with a person or organisation are the relations they have with them’’. The 1987 edition however gives some revealing examples ‘‘Ford insists that Carter’s dealings with him have been totally correct’’ and ‘‘He was questioned about his past business dealings’’. One common use of dealings is as a technical word indicating transactions on the Stock Exchange. Examples of this sense from the corpora were ignored. In its remaining uses from the corpus, and The Independent and Telegraph CD-ROM texts,

connotation and semantic prosody

73

it exhibited a strongly unfavourable prosody, but much less explicitly than was the case for peddle. Its concordance contained a striking paucity of modifiers expressing value judgements. In the 58 corpus examples, only four such modifiers were found: shady, illegal and behind-the-scenes which are unfavourable, and cordial which is favourable (but see below). However, if the wider co-text is taken into consideration, it becomes clearer that dealings generally indicates some unattractive or dishonest activity. It is very frequently found in the company of words and phrases expressing dubious legality, such as investigation into, inquiry into, revelations about, allegations about, exposé of someone’s dealings, as well as the following: (8) (9) (10)

Behind the mask we may discern traits of craftiness in his business dealings. the council intends to take a dim view of private dentists not behaving ethically in their commercial dealings. rugby union is a ‘‘profession’’ which dwells partly in the half-light of whispers and backroom dealings and partly in the daylight.

People also deny, decline or refuse to discuss their dealings. Nor is the disreputable colouring of dealings restricted to the business world.The corpus supplied the rather odd collocation sexual dealings. Taken by itself, of course, there is nothing to show whether the phrase refers to something felicitous or infelicitous, but the choice of dealings should alert us to the probability of the latter, and if we widen the context, we indeed find the following: (11)

It is the work of a writer for whom, in successive fictions, the theme of sexual dealings between people of different races has necessitated the representation of violence. Rapes and murders occur in this area [. . .] (Corpus)

Thus, although the journalist (one hopes) is not suggesting that inter-racial sexual contact is a bad thing, he or she is noting that the writer in question associates it with violent activity. The phrase dealings with is common, and the party named after the preposition is usually seen as pretty unsavoury e.g. Hizbollah, the Mafia, Iraq, extremists, an unscrupulous Press (corpus). In fact, it would appear that it is possible to imply that someone’s business affairs are shady simply by labelling them dealings, and by the same token, imply that a person or group of people are somehow unpleasant or criminal by talking of dealings with them. This claim is hard to prove, precisely because the taint or accusation is implicit, but consider: (12) (13)

but as more revelations have become known about her immediate family’s dealings with the Mafia [. . .] (corpus) Its spokesman said that it has no dealings with extremists. (Telegraph)

74 (14)

alan partington Labour’s dealings with the unions ‘‘made Jurassic Park look like a tea party’’, he added to Tory glee. (Telegraph)

The unfavourable prosody of dealings thus helps to imply the people involved in them are up to no good: (15)

(16)

The two witnesses are said to be a retired Swiss shopkeeper who had business dealings with Mr Ryan in the early 1970s, and a British businessman who had dealings with Mr Ryan in Libya. (Corpus) But Andrew Neil, the editor of the Sunday Times, denied his newspaper had any dealings with Mr Genoud and said no payment had been, or would be, made to him. ‘‘We have not made any payments to this man,’’ he insisted. (Corpus)

Thus writers either tell us something about their own attitudes or try to tamper with ours in a subtle fashion. If a journalist tells us that someone became a millionaire through business transactions, we are likely to have a different conception of him than if we are told he got rich through property dealings. Interestingly, a phrase of the kind dealings with Mr. . . is often to be found in reports of judicial cases where the person is implicated in something wicked or unsavoury (as with the Mr Ryan example above, who is accused of being an IRA bomber). In this final, rather complicated example, we return to the only case of dealings modified by a favourable adjective. (17)

The Japanese leader’s aim in his talks today with Jacques Delors, the Commission President, and John Major, the current chairman of the European Council, must be to use the mutual admiration between Japan and Britain to foster more cordial dealings between his country and the rest of the Community. A Japanese spokesman yesterday denied that there was anything more to the Thatcher meeting than a wish to stay in touch with a respected former prime minister. But he added: ‘‘We expect to develop our relations with the EC based on our good relations with the UK.’’

The word cordial would seem to suggest that all is sweetness and light. By contrast, the appearance of the following dealings seems to imply that the real relationship between Japan and the E.C. is less than perfect, and the suspicion of subterfuge is confirmed by the next sentence’s hint that the Japanese and the former British P.M. are involved in secret negotiations.

4.5 The creation of prosodies Another interesting example of the creation of attitude through prosody is the use in newspapers (and elsewhere) of the word fundamentalist (and fundamentalists,

connotation and semantic prosody

75

fundamentalism). This is an ‘‘over-the-fence’’ word, that is, a word used to describe an outsider group: the groups referred to by The Independent and The Daily Telegraph as fundamentalists would presumably not use the word to describe themselves. The word has gained—or been given—a highly unfavourable connotation, as can be perceived from the set of its collocates. It precedes in the corpus material noun heads such as aggressors, backlash, cults (another ‘‘over the fence’’ word), fanatics, fervour, guerrillas and warlords. It right-collocates with abusive, armed, crazy, crude, extremist, militant, murderous, primitive and truculent. These items seem to come in the main from the semantic fields of ‘violence’, ‘mindlessness’ or ‘underdevelopment’. The people the word is most frequently used to describe are Islamic or Muslim fundamentalists, although there are also a few references to Christian and Catholic fundamentalism. However, what makes this item of special interest is that it seems to be possible to apply it creatively to denote any group the writer dislikes. A writer can refer to a group he or she disagrees with as fundamentalist, in the hope that the bad connotation of the word will infect that group by a process of contagion or prosodic spread. There are, to cite an extended example, a number of references to Green fundamentalists in an exchange of letters in the newspaper corpus. The letters are in response to an article which appears on 1st June 1992 and begins (my emphasis): (18)

WITH preparatory meetings due to begin today at the Earth Summit in Rio de Janeiro, more than 200 scientists and intellectuals, including 54 Nobel Prize winners, have signed a declaration attacking Green fundamentalists who oppose industrial and scientific progress, writes Nicholas Schoon. Toxic chemicals and radioactivity are inescapable facts of modern life, the academics say.

This sparks a reply from a reader which begins: (19)

Sir: Your report on the declaration by certain scientists and intellectuals condemning green fundamentalism at a Rio fringe meeting (‘‘Irrational Greens condemned’’, 1 June) is depressing. The report probably simplified and condensed complex arguments in the declaration [. . .]

and which is itself answered a couple of days later by another reader: (20)

Letter: Earth Summit must reject green fundamentalism Sir: Andrew Watterson’s depression (Letters, 2 June) at the prospect of scientists warning against swallowing green fundamentalism whole is as nothing when set against the depression induced by his own fulminations against logic and reason [. . .] The sort of mindless nonsense put about by fundamentalist greens is more likely to make the situation worse than better.

alan partington

76

It is to be noted how the term green fundamentalism has been inserted by the subeditor into the letter’s title. Three further letters to the paper all have the title Earth Summit must reject green fundamentalism even when the letter’s contents are proGreen: (21)

Letter: Earth Summit must reject green fundamentalism Sir: An old joke. Robber: ‘‘Your money or your life.’’ (Pause) ‘‘I said, your money or your life.’’ Jack Benny: ‘‘I know, I’m thinking it over.’’ How droll it is to witness the reprise of this joke by the Western nations at Rio de Janeiro.

The sub-editor, by repeating the phrase in the titles, is granting the status of accepted collocation to green + fundamentalist, albeit unwittingly.8 That something called green fundamentalism exists is being accepted without challenge. And, by the process of semantic prosody, the more frequently green collocates with fundamentalist the more tainted the former is likely to become. Whereas the process of the creation of prosody in green fundamentalist was probably not deliberate, one may speculate that there is often a conscious desire on the part of newspapers to depict certain groups or ideas in a good or bad light, to create or perpetuate particular semantic prosodies. One might perhaps recall the attempt in the eighties on the part of The Sun to create an unfavourable prosody around the political term Left with repeated references to the loony Left, loony Lefties and so on (see Morley 1998). Nor is it confined to one side of the political divide. Suffice it to recall the bad press capitalism received in the seventies and later from the leftwing press. Prosodies—good and bad—are also created by papers around the names of their most or least favoured public figures e.g. Thatcher, the Iron Lady or mad dog Gadafy (for the latter see Fowler 1991: 112–19).

4.6 Conclusion From the observations in this chapter, it is clear that the study of semantic prosody requires the contribution of researchers from a wide variety of disciplines ranging from that of lexicography and corpus linguistics through text and discourse analysis and translatology to cultural and literary studies. As far as the first of these fields, dictionary making, is concerned, Louw’s hope was that prosodies should begin to receive their just attention. The evidence from set in (whose prosody was mentioned in all the modern learners’ dictionaries examined), peddle (prosody indicated in two out of three dictionaries) and dealings (prosody not mentioned in any) suggests that

connotation and semantic prosody

77

progress has been made, at least as far as learners’ dictionaries are concerned, but there is still some room for improvement. Louw’s other concern, the analysis of the unconscious or conscious exploitation of prosody, though still in its infancy, is clearly of very general interest. As Louw points out in the conclusion to his study, the use of semantic prosodies is of great interest to the persuasion industry, and not just the opinion makers who write for newspapers, but to politicians, propagandists and advertisers. To the rest of us, it is important to know when and in what direction our linguistic expectations are being manoeuvred.

4.7 Areas for further research One promising area for analysing semantic prosody is in verb phrase collocation with favourable or unfavourable objects. Sinclair mentioned set in, Louw has examined to be bent on, and here we have looked at peddle. In terms of heightening learners’ awareness of this phenomenon, a beginning could be made by presenting concordances of verbs with well-known prosodies, perhaps commit or perpetrate. An interesting exercise might be to compare the collocational behaviour of pairs such as persist and persevere, which may have very similar cognitive meanings, but widely differing prosodic behaviour. Another promising area is that of adverb-adjective and adjective-noun head phrases in which the first word of the pair is an intensifier. Louw looks at the intensifying adverb utterly, claiming that the Cobuild corpus data shows it to have an unfavourable prosody. Other intensifying adverbs could also be analysed. It would also be interesting to examine utter, along with other intensifying adjectives. Semantic prosody is very frequently hidden from immediate view and therefore difficult to predict out of context—witness the fact that even different forms of a lemma may display different behaviour: ‘‘build up confidence’’ (transitive) is favourable, ‘‘resistance builds up’’ (intransitive) is unfavourable (Louw 1993: 171). It may therefore be worthwhile to begin looking at texts which are likely to exploit prosodic effects—newspapers,political language,advertising etc.—toisolate potential candidates and subsequently follow this up by examining the corpus data. A corpus of advertising texts would be extremely useful for searching out prosody. Finally, semantic prosody is an important area of research for translation studies. From the present author’s research, it seems to be the case that cognate, or ‘‘lookalike’’ words in two related language can have very different semantic prosodies. The adjective impressive in English has a favourable prosody, collocating in the English corpus with items such as achievement, best, talent, dignity and gains etc. The look-alike Italian word impressionante, on the other hand, was found as often

78

alan partington

as not to collocate with neutral or unfavourable items (corpus examples: series of price rises, assassination attempts, amongst others). Conversely, the English verb incite and the related noun incitement almost always relates to unfavourable phenomena (racial hatred, moral hazard, violence among others), whereas the Italian verb incitare (noun incitamento) often collocates favourably, with the meaning ‘to encourage’ (corpus examples including that private banks should be encouraged to adopt more rational policies and that two brothers encouraged each others’ athletic performances). The pitfalls for translators unaware of such prosodic differences are evident.

Notes 1. The terminology can be confusing. Leech (1974) uses ‘‘conceptual’’, whereas ‘‘denotational’’ was first coined by J.S. Mill in 1843 to contrast with ‘‘connotational’’ meaning. The term ‘‘core’’ meaning is sometimes found, but many writers challenge the implication that other types of meaning are less central. I avoid the term ‘‘referential’’ meaning since many authors (including Leech) insist that other kinds of meaning, not just the conceptual sort, refer to something. 2. Judging from the corpus evidence, male old age. 3. There may be a slight overstatement here. That a number of items have negative or positive prosodies was clear before the advent of corpus lexicography, e.g. items like commit or perpetrate are usually described, even in non-corpus based dictionaries as collocating negatively. For example, the OALD (1980 edition) defines commit as follows: ‘‘perform (a crime, foolish act, etc)’’ and perpetrate as follows ‘‘commit (a crime, an error)’’. Nevertheless corpus lexicography seems to be demonstrating that many more items than was previously suspected have such prosodies. 4. Namely the Longman Citation corpus (around 25 million words in 1987) and the Cobuild collection of texts (around 10 million words in 1987). 5. OALD 1995 used data from the British National Corpus (100 million words); CIDE used data from the Cambridge Language Survey (100 million words); Cobuild 1995 used data from the Bank of English corpus (around 200 million); LDOCE 1995 used data from a collection of written and spoken corpora compiled by Longman. 6. Typical is the Shorter Oxford English Dictionary entry: ‘‘(of weather, a condition, etc): to begin (and seem likely to continue)’’. 7. ‘‘Wherever it seemed reasonable, we gave preference to the concrete over the abstract because the concrete meaning is often easier to grasp, and abstract meanings can often be seen as variations on the concrete’’ (Cobuild 1987: xix). 8. It should be stressed that it is the page sub-editor who does this. None of the writers of pro-Green letters use the term fundamentalist to describe the group, which supports the statement above that it is an over-the-fence word.

5 Syntax If in ‘‘Cosi’’ it is the women who are put upon, then in ‘‘L’Italiana in Algeri’’ it is the eponymous heroine who wears the trousers (Corpus).

It is sometimes thought that concordancing is only really useful for the study of single lexical items. To show that this is not the case, in this chapter attention is shifted to a level beyond the word and phrase to focus on the relation between two separate clauses or between phrase and clause, a level which, for the sake of convenience and to distinguish it from the ‘‘lexical’’ level, we can call the syntactic. Halliday’s (1992a) theory of context, as mentioned in Chapter 3 (see section ‘‘The evaluation of context’’), maintains that, in order to discover information about the clause, one must study the rank level above it—which constitutes its context—that is, the grammatical sentence. In this chapter, then, the object of study will be moreor-less whole sentences. Before the advent of corpora, traditional grammar tended to ‘‘assume that the study of grammar has strict limits, and lexis, on the whole lies outside them’’1 (G. Francis 1993: 142). Corpus linguistics of the kind to which the present set of studies belongs, on the other hand, attempts to investigate the interface between lexis and syntax: syntactic structures and lexical items (or strings of lexical items) are co-selected, and [. . .] it is impossible to look at one independently of the other. Particular syntactic structures tend to co-occur with particular lexical items, and—on the other side of the coin—lexical items seem to occur in a limited range of structures. The interdependence of syntax and lexis is such that they are ultimately inseparable.—(G. Francis 1993: 147)

To highlight this, Francis examines concordances of the word possible. This item ‘‘has a wide range of environments which make it unique among adjectives’’ (G. Francis 1993: 147). It appears in the pattern ‘‘the + a superlative adjective + possible + head noun’’, as in the highest possible level, the worst possible outcome as well as in the single unit as soon as possible. It also appears, combined with as, after a wide range of adjectives, adverbs and quantifiers: as early as possible, as often as possible etc. Other possible patterns include where/wherever possible, when/whenever possible, if possible. This range of environments makes possible grammatically unique, but it is by no means unusual in being so:

80

alan partington If we take any one of a huge range of the more frequent words in English, and examine its citations en masse, it will emerge that it too has a unique grammatical profile, which certainly cannot be encapsulated by calling the word in question an adjective or a noun or a preposition.—(G. Francis 1993: 147)

Thus, the behaviour of a lexical item and its environment are interdependent. As an example of how looking at lexis by means of the concordance can provide information at various ‘‘higher’’ language levels, from the syntactic to the discoursal, the main part of this section is dedicated to an analysis of a concordance of if constructions, in which the environment dependent on if is two whole clauses. It begins however with suggestions of how concordancing can be used to analyse the relationship between two phrases, in particular contiguous verb phrases.

5.1 Investigating verb constructions One major source of error in non-native language is the area of verb plus prepositional colligation (defined as the collocation of a lexical and a grammatical item)2 e.g. knowing what prepositions follow rely, accuse, blame, criticise etc. Another frequent difficulty is that of verb plus verb agreement (what the Collins-Cobuild English Grammar (CCEG) calls ‘‘two phase verbs together’’ p. 186) i.e. deciding which verbs require a gerund, which the infinitive (but with or without to?), and which a finite verb clause. Concordance data is particularly helpful for indicating the answers to these problems. Learners’ dictionaries normally also give such indications, at least in their examples. But a concordance has a number of advantages. First of all, the examples are likely to be more abundant, and secondly, rather subtle tendencies of verb constructions can be highlighted. For instance, CIDE gives both in and with as possible prepositions following persevere. The corpus contains seven occurrences of persevere with, and only one of persevere in, an indication of relative frequency. The CIDE examples for persist contain persist in, taking a gerund, and persist with before a normal noun phrase. The 87 occurrences of persist from the corpus paint a more complex picture. Persist in is used with both the gerund and noun phrases e.g. Lautrec persisted in his commitment to scenes from everyday life, whereas persist with is never found before a gerund. In either case, the subject is almost always personal—somebody is persisting in/with a course of action. A whole range of other constructions are found when the subject of the verb is impersonal or abstract such as fears, doubts, the notion, rumours, the feeling. In such cases persist can take prepositions such as about and over (fears persisted over the UK trade deficit), or it can be followed by a ‘‘that plus finite clause’’ construction (the notion has persisted that

syntax

81

Conservatism is not an ideology).3 It is also frequently found with for, until, after, over when introducing a temporal expression (it may persist for many weeks).

5.2 If constructions and language learners The rest of this chapter recounts a fairly detailed concordance study which was carried out with a class of translation students at the SSLMIT (Faculty for Translators and Interpreters, University of Bologna at Forlì). The CCEG makes the point that the foreign learner is usually given very rudimentary information about the kinds of construction this word finds itself in: Foreign learners are often taught that there are three kinds of conditional sentence: • the first conditional, in which the verb in the main clause is ‘‘will’’ or ‘‘shall’’ and the verb in the conditional clause is in the simple present tense; • the second conditional, in which the verb in the main clause is ‘‘would’’ or ‘‘should’’ and the verb in the conditional clause is in the simple past tense; • the third conditional, in which the verb in the main clause is ‘‘would have’’ or ‘‘should have’’ and the verb in the conditional clause is in the past perfect tense. This is largely correct, but does not fully describe the normal patterns of tense in conditional clauses [. . .]—(CCEG 1990: 350)

If is a fairly frequent item, occurring over 8000 times in the combined corpus. It is appreciably more common in academic texts than newspapers, occurring about once every 400 words in the academic corpus and about once every 700 in the newspaper texts. A class of second year Italian mother-tongue translation students were presented with the first 100 corpus examples of the use of if from the Arts section of the newspaper corpus (excluding instances of as if, or where if was synonymous with whether) in the context of a full sentence. This section was chosen because it is generally the part of the newspaper containing the most ‘‘carefully written’’ texts, those produced with the fewest time constraints (book/cinema/ exhibition etc reviews, cultural pieces and so on). Since 100 concordance examples would be rather daunting for students, especially ones with little experience of such material, the class was divided into five groups, so that any one student had 20 sentences to examine.

5.3 Clause combinations Of these 100 sentences, only 17 corresponded to any of the three conditional forms quoted above—7 first conditionals, 6 second conditionals and 4 third. By far the

82

alan partington

most common form was that in which the if-clause had the verb in the simple present and the main clause had the same tense and aspect (28 occurrences), as in the following: (1) (2) (3)

If a great opera such as this has difficulty holding a place in the repertoire, what hope is there for the merely excellent? If all this self-doubt is part of the act, it’s a brilliant performance. If he finds the right character, the actor effectively claims the role for life.

There were a further eight cases where the if-clause verb was in the simple present and the main clause verb had a simple present clause preceded by a modal (must (twice), should, may, might, would, could, have to): (4)

(5)

IF it worries you that the national survey on sexual habits is not going to take place, you could take comfort from the thought that it probably won’t make a vast amount of difference in the long run. The debate is complex, but one thing at least is certain: if British architecture is to develop in a way that expresses professional as well as lay aspirations, the common ground must be rediscovered and enlarged upon.

Four sentences had the if-clause verb in simple present and the main verb in imperative mode, as in the following: (6)

If a composer remembers to keep this audience entertained, think what he can say to them all at the same time.

and on two other occasions the if-clause verb is in the present continuous, as follows: (7)

If you’re rambling round Dublin, you meet politicians. It’s that kind of country.

These constructions are perhaps not very surprising—the present plus present combination is usually mentioned in passing in most didactic grammars—but its relative frequency was most certainly a surprise. There was, however, no lack of other, more exotic, combinations. For example, simple past with simple past (5 examples, of which the following): (8)

(9)

Throughout the afternoon the piano sound was beautiful and even, though the treble lacked body and brilliance [. . .] if either of his hands was sometimes a shade too tactful, it was his left. If Nice Work was the novel with which Lodge gave two of his keenest preoccupations

syntax

83

the slip—the university, partially, and Catholicism, totally—it was also the book which should end the habitual conflict [. . .]

and combinations such as the following, which occurred just once or twice: (10)

(11)

(12)

(13)

(14)

(15)

(16)

[. . .] who, if she was named ‘‘Stella for star’’, now well knows that there is a pragmatic percentage to be paid for adoration. (simple past with simple present) Enesco is a major force in twentieth century music [. . .] and if his little Poemes de Clement Marot show a limited aspect of his art, they still breathed an engaging wit and charm in Miss Cotrubas’s heart-warming performance. (simple present and simple past) The more creative will experiment with less conventional architects if it will help them win planning permission (future simple with future simple) It would help if interested people who spot building work going on in what they know, or suspect to be listed buildings, would phone their local planning office to find out if listed building consent has been sought (would + base verb and would + base verb) The irrational thought almost occurs that Shakespeare couldn’t have been that smart if he didn’t realise that he had here a performer who can turn sows’ ears into silk purses, and silk purses into live butterflies. (simple past and could have + past participle) I sometimes wonder what it would be like if ‘‘Stop the World’’ had had a proper director, who did it for a living. (past perfect and would + base verb) But even if Portobello Road had formerly been grand—which is doubtful—locals do not want artificial Victoriana to replace existing reality. (past perfect and simple present)

In a number of the remaining sentences, one of the two parts of the if construction did not contain a verb or the verb was without tense: (17) (18)

(19)

If a remedy can be found to everyone’s satisfaction—through private enterprise, then well and good. and Jane Leslie MacKenzie’s sad and tense characterisation of the Countess was extremely plausible, if lacking in the real sense of fun which the lady must have had in order to be drawn into the intrigues in the first place. He will also explain the scheme, if asked, in terms of Vedic creation myth, but these reflections are quite incidental to the success of the building.

In (19), if the if phrase were expanded, the verb might be either if he is asked or it might be if he were asked. Because it is frequently dangerous to ‘‘reconstruct’’

84

alan partington

missing parts of phrases in this way, none of these types of sentences were included in the count of conditionals or present–present or past–past combinations, but were counted as separate groups. Although the three ‘‘classic’’ conditional forms may be a useful model, the examples above show how they can be over-ridden at any moment in a discourse by contextual and pragmatic considerations. The tense and aspect of the verbs in the two clauses is dictated by what the text producer wishes to convey at any moment in time. From a didactic point of view it is probably worthwhile exposing learners, at least at a post-beginners phase, to the wealth of possible combinations in if constructions. They can thus see that the three conditionals constitute a model for, rather than a constraint on, natural language production.

5.4 Expanding the investigation: Beyond the use of if as conditional In order to test how far these results were due to the nature of the sample, another 150 concordance lines were generated from other sections of the newspaper corpus (50 from each of the Home news, Business and Sports sections) and 120 sentences from the academic corpus (30 each from the medical, sociology, scientific4 and philosophy subsections). In only one of these sub-groups did the combined three ‘‘classic’’ conditional structures constitute over 50 per cent of cases—the sociology writings (17 out of 30). The lowest percentages were obtained in the scientific (6 from 30) and business sections (13 from 50). Overall, the three conditionals covered 39 per cent of cases (17 per cent first conditional, 17 per cent second and 5 per cent third). The single most frequent construction was the simple present and simple present combination (26 per cent). A further 12 per cent had simple present in the if-clause with a modal in the main clause. 6 per cent had the simple present with an imperative verb in the main clause. 5 per cent contained simple past with simple past and another 5 per cent had present or past in the if-clause and no verb in the other part of the sentence. Although the percentage of the classic three conditionals is higher than in the first 100 sentences examined, many of the same trends seem to be apparent. During the discussion of their findings, the students observed that there is a greater tendency than is normally pointed out in didactic grammars for verb forms in the two clauses to be identical. The usual explanation for this—that an if sentence with two present or two past tense verbs refer to a ‘‘common occurrence’’ (CCEG 1990: 350) or ‘‘a universal truth’’ (e.g. if the sun shines, it is warm)—is inadequate. Universal truths are better expressed by when constructions. The data suggests other explanations. For example, there would appear to be a particular, very common,

syntax

85

construction with if where it is used as a contrastive qualification to the proposition expressed in the main clause: (20)

(21)

(22)

And if Asylnuratova does not possess the imperturbable technique of a Guillem, the expressiveness of her dancing is matched by so mobile and luminous a face that everyone else on stage looks slightly blank. Far from being a harmless slice of programme music, Popes is a remarkably positive statement. If at times the surface gives an eclectic impression, the outline of the whole is marvellously clear. But if it’s taken him 28 years to rediscover the show, the character’s never left him

This contrastive use is especially typical of sentences where there is no verb in the if part of the sentence: (23) (24)

To go further in these directions would also require more boldly gestural language than Philip Vellacott’s plausible, if prosaic translation provides. The singers received buoyant if occasionally untidy support from a classically sized and seated orchestra under that excellent Mozartian Alan Hacker.

These are not, in fact, conditional constructions at all. The meaning of if here is approximately the same as though or even if which grammars usually define as ‘‘concessive’’. Learners, unfortunately, are not always made aware that if can have this non-conditional concessive use. This contrastive use is only one of the many ways that an if-clause or phrase is employed to qualify the main sentence proposition: (25)

(26) (27)

and our own Queen has an almost comparable claim on immortality if the story is to be believed that when asked to attend a performance of Figaro, she replied: ‘‘Is that the one about the pin?’’ What we were offered, if definition is needed, might be called ballet-pantomime: it was certainly not opera. If we’re honest, most of us reviewers prefer to engage in a dialogue with the author and director, addressing concepts of political theatre rather than the actor’s little trick of twitching his left ear

Such sentences rarely have one of the three classic conditional constructions, since the main clause is conditional on the if-clause only in a formal way. Such ifclauses are really acting as modals which evaluate the truth value or the certainty– uncertainty, possibility–impossibility or necessity–non-necessity of the information (or advice, suggestion, speculation and so on) supplied in the main clause.5 Verbless if structures—if necessary, if at all, if successful, if so, if present, if honestly calculated (corpus examples) regularly have this function.

86

alan partington

A similar if construction—when once more the two verbs are in the same tense, usually simple present—is to be found in the following sentences: (28)

(29) (30)

If the first of his houses in Majorca, due for completion next spring, recalls the living quarters of a thirteenth-century Cistercian monastery that should come as no surprise. If the one-man show is as close to ‘‘actors’ theatre’’ as we’ll ever get, perhaps it’s just as well. IF THERE is one thing Joe Jackson particularly dislikes, it is being made a fuss of. And if there is one thing that West London hotels particularly enjoy, it is making a fuss of Joe Jackson.

In each of these cases, there is no real conditional dependency of the main clause on the other. The construction is really being used for rhetorical effect. They could all be reasonably paraphrased with a non-conditional construction—It should come as no surprise that the first. . ., It’s just as well that the one-man show . . . , Being made a fuss of is one thing . . . . Since the propositional content is shared by the two halves, there is usually no need for the verb form to alter. A final use of if which tends to dictate an identity of verb tenses is the ‘‘if p is true, then q is true’’ construction beloved of semanticists: (31) (32) (33)

If in ‘‘Cosi’’ it is the women who are put upon, then in ‘‘L’Italiana in Algeri’’ it is the eponymous heroine who wears the trousers If Peter Greenaway’s camera is anything but a participating human representative, Gilbert’s is nothing else, addressed, ogled, and even winked at by the heroine. If ‘‘On the Town’’ caught the desparate party mood of the Second World War [. . .] then the ironically titled ‘‘Fair Weather’’ captures the disillusioned mood of post-war slump.

There is clearly no real-world conditional dependency of the main clause on the if-clause in the above examples. They bear a formal resemblance to the kind of improbable examples which abound in propositional logic of the type if Sweden is in Africa, then Japan is a republic or if Kennedy was a President, then cabbage is a vegetable,6 in which there is no obvious relationship between the two parts, since the purpose of the exercise is to test the truth value of the whole sentence when true or false values are assigned to each of the two parts.7 In natural language, however, there is almost always some kind of relationship between the two parts of an ‘‘if p is true, then q is true’’ construction. They often have a parallel structure (as example 33) and they are almost invariably linked thematically, by topic. The exception to this rule is ironic phrases of the type if he’s a liberal, then I’m the Queen of Sheba, of which, unfortunately, there was no trace in the corpora.

syntax

87

5.5 Further research The possible avenues of research of this kind are virtually limitless. Similar activities could be performed with other conditional and pseudo-conditional markers provided that, unless, as long as, no matter how/when/what, whenever, supposing and so on) or with other subordinators with a particular syntax (despite, until, in case, etc). Returning to verb plus verb constructions discussed earlier in this chapter, concordance data can throw light on the use of many of the ‘‘two phase verbs together’’ listed in CCEG (p. 186). Two simple examples will suffice. A concordance of the verb afford highlights how, when the context is money (or anything else which someone might ‘permit themselves’), it is well-nigh invariably accompanied by a possibility modal (can, couldn’t, wasn’t able to etc). This verb is really of the form ‘be able to afford’. Without such a modal it is a different verb altogether, meaning ‘provide’. The verb dare is used, in the vast majority of corpus examples, in negative or interrogative contexts. Exceptions are fixed expressions such as I dare say and I dare you (to). In all these cases, concordances can provide teachers and learners with data to help them go beyond, or more clearly understand, the distinctions highlighted in grammars and textbooks.

Notes 1. Owen (1993: 169) cites the following extract from a language teaching textbook to illustrate the traditional view of the relationship between syntax and lexis: ‘‘Our approach to language teaching, then, is structural. The words we choose to present for use in the structures are only of secondary importance, because once the patterns of English are mastered, it is relatively easy to learn new words to fit into the patterns’’ (Broughton 1968: 14). To be fair, the structural approach developed in reaction to previous simplistic ones based largely on the memorisation of word lists. A view of syntax and lexis as interdependent would appear to be a healthy compromise. 2. Benson (1985) instead talks of ‘‘grammatical collocation’’ as the combination of a ‘‘dominant’’ word, typically a noun, verb or adjective with a ‘‘grammatical’’ word, i.e. a preposition or adverb. 3. The nouns used with persist in these examples are, in fact, acting as general nouns (see Chapter 6 below), typically with cataphoric reference, as in: a feeling also persisted that Parke might be able to trouble Jahangir. 4. The texts in this subcorpus are not actual scientific papers, but writings in ‘‘popular’’ science, the history of science, philosophy of science and so on. 5. Brown & Levinson (1987: 145) note the use of if-clauses in conversation as hedging devices. ‘‘A

88

alan partington hedge’’ they say ‘‘is a particle, word or phrase that modifies the degree of membership of a predicate or noun phrase in a set; it says of that membership that it is partial, or true only in certain respects, or that it is more true than perhaps might be expected’’. They clearly see hedging as a major means of expressing modality in conversation. They also note ‘‘that certain usages conveyed hedged performatives—that is, they modify the force of a speech act’’ (see G. Lakoff 1972: 213). Brown & Levinson go on to note that if-clauses are ‘‘a very productive source in English’’ of adverbial clause hedges. They cite Heringer (1972) who claims that certain felicity conditions can be suspended if they are placed in an if-clause. In other words, it is a gesture of politeness on the part of the speaker to not presuppose that the hearer is either able or willing to carry out what the speaker is requesting of him/her. This accounts for the examples quoted: Close the window, if you can/if you want. The corpus contained a similar example—Pity the poor developer, if you can.

6. Both examples from Allwood et al. (1977) Logic in Linguistics. The authors discuss the differences in propositional logic and natural language usage of such if-clauses. 7. Lyons tells us: ‘‘In the implication p ( q (which may be read as ‘p implies q’ or ‘if p then q’) p is the antecedent and q the consequent. The terms ‘imply’ and ‘implication’ are used in a variety of distinct senses both in everyday discourse and in philosophical usage. In the standard interpretation of the propositional calculus, they are used in the material sense, in which there is not necessarily any connexion of meaning between the antecedent and the consequent’’ (1977: 145).

6 Cohesion in Texts1 Jana contended that the Phoenician empire was Arab and therefore Venice was once part of a Pan-Arab civilisation, not to mention the other claim that many Venetians have Arab roots from ninth-century traders. Venetian historians promptly discounted the claim (Corpus).

This chapter attempts to show how corpus data, accessed by the concordancer, can be used to investigate cohesion in texts. The focus of attention in the previous chapter was on units of language larger than the word or phrase, namely the clause and sentence. In this chapter two separate units of language are involved and the relationship between them is the main object of study. The first consists of particular types of phrases—both noun and verb phrases—which have a phoric or referential function i.e. are used to refer to another part of a text. The second unit, corresponding to the part of the text being referred to, can be of varying types and lengths, ranging from the phrase, to the clause, to the sentence and even beyond, and can only be generally defined as ‘‘portion of text’’. In comparison to the large number of studies made by corpus linguists in the areas of lexis and syntax, relatively little work has been done on textual analysis (but see Fligelstone 1992; Stubbs 1996). It may be the case that concordancing techniques have been thought less appropriate to the study of longer stretches of text. However, most if not all modern concordance programs allow the user to study not only single lines of text but also, by means of an ‘‘expanded context’’ or ‘‘View’’ facility, to have access to larger stretches of the texts, perhaps even the whole texts in which they appear (for example, in the case of the newspaper corpus here, the whole article a line originally appeared in). The methodology of these text studies, then, is as follows. The concordance—along with the user’s intuition —makes it possible to isolate items which they suspect of playing a role in textual deixis and provides the immediate phraseological environment of the item. The ‘‘View’’ facility allows one to go on to look at how the item in question fits into a pattern of textual cohesion, in particular to find the ‘‘portion of text’’ which is being referred to (when possible—see the following section). When a number of examples are uncovered, it is possible to build up a picture of the strategies writers typically employ to create cohesion.

90

alan partington

The concordance was here used to look at three particular phenomena relating to text reference: first of all, general nouns, secondly labelling nouns and finally the set of what I have decided to call ‘‘general verbs’’.

6.1 Text reference 1: General noun phrases 6.1.1 The functions of general nouns This section contains a brief overview of general noun phrases and the way they are use to refer within texts. The literature on the class of items known as general nouns is scant; it has undoubtedly not received the attention it deserves. ‘‘These items are often neglected in descriptions of English’’ as Halliday & Hasan point out in Cohesion in English (1976) in a chapter on lexical cohesion, and they have been given relatively little notice since then. They are not mentioned in classic reference works such as Quirk et al. (1985) or Givòn (1993), nor does Hasan herself pay them much attention in her later works on cohesion (see, for example, Hasan 1984; Halliday & Hasan 1989). McCarthy (1991) mentions them in passing, including them among superordinate items which can play a cohesive role. Martin (1992) does much the same, collapsing, as he says, Halliday & Hasan’s categories of hyponymy, superordinate and general items together. However, I would argue that the group of general nouns has a number of distinguishing features, especially their ability to function as vehicle of a writer’s or speaker’s attitude2 and their frequent use as discourse marker (see discussion below) along with their particular role in the tone unit3 which set them apart from other superordinates. The reference statement on general nouns, then, remains that of Halliday & Hasan. They define them as follows: On the borderline between grammatical and lexical cohesion is the cohesive function of the class of general noun. We can speak about a borderline here because a general noun is itself a borderline case between a lexical item (member of an open set) and a grammatical item (member of a closed system). The class of general noun is a small set of nouns having generalised reference within the major noun classes, those such as ‘human noun’, ‘place noun’, ‘fact noun’ and the like. Examples are: people, person, man, woman, child, boy, girl (human) creature (non-human animate) thing, object (inanimate concrete count) stuff (inanimate concrete mass) business, affair, matter (inanimate object) move (action)

cohesion in texts

91

place (place) question, idea (fact) These items [. . .] play a significant part in verbal interaction, and are also an important source of cohesion in the spoken language.—(Halliday & Hasan 1976: 274)

General nouns, or more precisely items when used as general nouns, are part of the system of deixis in English; they function as pro-forms. Some of them tend to have specific referents, which is the case for man, child and the other human general nouns, and also for creature, object and place. But many of the others (including matter, question, thing, stuff) tend to refer to longer stretches of discourse, as in this example (the examples in this section were collected during a previous study using Cobuild corpora (1991 versions)): (1) [. . .] how he reckoned he didn’t think they’d get one million pounds for him because he’s obviously gone down in value since his stay at Forest and stuff like this.—(Cobuild: Spoken corpus)

Often, the reference is so vast or vague that it is not possible to pick out any particular part of the surrounding text as referent: (2) [. . .] in each, three children bickered in the back. In the passenger seat of one, a worrying father dreaded the whole thing, while his placid wife drove on.—(Cobuild: The Times)

In fact, the whole point of the cohesive phrase in the above example is that it should be indefinite. Halliday & Hasan make two points about general nouns that require discussion in the light of what follows. First of all they state: a general noun in cohesive function is almost always accompanied by the reference item the [. . .] The most usual alternative to the is a demonstrative—(Halliday & Hasan 1976: 275)

However, a surprising number of instances of the use of general nouns found in these corpora are in phrases with the indefinite a and such: (3) Those are they who walked through the bitter valley and made it a place of spring. (Cobuild: General) (4) Mabbutt [. . .] must have felt cheated though he is too ‘‘nice’’ a person to say it. (Cobuild: The Times)

92

alan partington

(5) They abandoned the felling of tropical trees to get at the plants established in their branchings, but burned a forest to ensure a monopoly in orchids. Such things hit you in the pit of the stomach.—(Cobuild: General)

On occasion longer determining phrases are found, such as (this or that) sort of: (6) I did a special option on phonetics which was up-to-date sort of stuff.—(Cobuild: Spoken) (7) Those who follow this sort of thing will tell you that yesterday’s recapitalisation of Tees Customer Electronics was only a matter of time.—(Cobuild: The Times)

The second point made by Halliday & Hasan which requires discussion is: this the is anaphoric,4 and the effect is that the whole complex ‘‘the + general noun’’ functions like an anaphoric reference item.—(Halliday & Hasan 1967: 275)

This is perhaps too restrictive. As in example (7) above, a determiner with a general noun (especially of the type thing(s), matter, business, talk etc) can introduce cataphoric reference, as follows: (8) there are also things it is more difficult for parents to get a grip on, such as empathy, and the ability to retain ideas and hold them despite everybody else telling them [. . .]—(Cobuild: General)

However, Halliday & Hasan’s point that it is not only the noun itself but the noun in association with a reference item which is the element creating cohesion is very important. General nouns are found either accompanied by a determiner or some adjunct like similar, or like this (as in example (1) above). The term general noun is even perhaps itself a misnomer. Perhaps it would be better to talk of general noun phrases. 6.1.2 An example: a move as a general noun phrase This section explores the behaviour of one particular general noun phrase, a move. Halliday & Hasan include the item move as an example of the phrase the move functioning as a general noun: We all kept quiet. That seemed the best move. The phrase a move, containing the indeterminate rather than determinate article, was chosen to underline the point that determiners other than the can appear in general noun phrases. This, then, is a concordance of all the occasions when a move is used as a general item in the newspaper corpus:

cohesion in texts

93

1 epay KLM out of operating profits a move analysts said would further endebt 2

creditors' scheme of arrangement a move favoured by the council associatio

3 way to return to mortgage lending a move likely to require the support of a 4

an outright ban on such pursuit, a move ruled out by top police officers w

5 be a cut in personal income taxes a move that many companies have been call 6 o a Las Vegasstyle gambling venue a move that would require a change in Cal 7

extremist religious ideologies _ a move which could prove counterproductiv

8 ference in Perth later this month a move which will not help Labour's claim 9 es by as much as one full point _ a move which could prompt a rise in UK ba 10 he closure of Bosnia's airspace, a move intended to protect mainly Muslim 11 m anywhere in the penalty area', a move designed to encourage attacking pl 12 s in South Africa's urban areas, a move welcomed by Nthato Motlana, chairm 13 publics or to migrate to Russia, a move that could trigger largescale econ 14

was planned in the same office, a move that is said to have catapulted Mr

15 ver next year's legal aid rates, a move that seems likely to fuel industri 16

have to be ratified by Britain, a move that would be deeply unwelcome in

17 f launching the ball into space, a move that would have brought a bellow o 18 in the Retail Price Index (RPI), a move that would lower the apparent infl 19

restraint on GPs' drug budgets, a move regarded by Mr Clarke as a `breakt

20 ow says it will check each case, a move expected to slow repatriation to a 21 ajesty's Sagbutts and Cornetts), a move which is welcome among amateur cho 22 y was barred from the elections, a move which one senior Moscow official s 23 x and spending plans in January, a move which succeeded in stopping Labour 24 hic predominance of the Muslims, a move which their leaders and their Syri 25 First Edition) HONG KONG (AP) In a move highlighting China's power over Ho 26 sitors if BCCI were to close. In a move that appeared to distance it from 27 in the White House this week, in a move that could break the deadlock in t 28 shuffling its management.

In a move to tighten control of a farreachin 29 on Bond stationery group DRG, in a move to rally support against the £697m 30 t> Rushes

IN A move which bids fair to relegate the ph 31 eaf and blind charity, Sense. In a move which they may later regret, Jayne 32

Medical Association backed such a move 18 months ago, but only a handful

33 eographer in the US to make such a move. Among the 13 musicals she choreog 34

Denmark itself to push for such a move, but, of course, it would make our

35 too many editors.' Besides, such a move for Mr and Mrs Evans was not pract 36 on ivory remains in place.' Such a move, he added, would allow the countri 37 must recognise the folly of such a move. Not only would Britain be shootin 38 t monetarists would support such a move. They appear to believe sterling e 39 nst Israel in the Gulf war. Such a move will no doubt alarm the military a 40

and call for a referendum. Such a move would represent a Uturn. Neil Kinn

41 elves. Most observers agree such a move would be an almost certain recipe 42 olonel Muammar Gaddafi, but such a move would be resisted by Libya's main 43

reasonable to devalue. But such a move would be taken after the UK's coun

94 44

alan partington up to 29.9 per cent though such a move would require the approval of Jagu

45 that's happened,' she says, such a move would have to come one step at a t 46

doctors and hospitals.

Such a move would mean a fundamental change in

47 e Tories, the net effect of such a move would be to cut Labour's majority

When the wider context of these lines was examined, it was clear that in most cases a move is anaphoric, reflecting the greater incidence of anaphora compared to cataphora in the varieties of English I have examined. Nevertheless when the phraseology in a move (ll. 25–31) is found, a use probably typical of a journalistic register, a course of action is being signalled cataphorically, that is to say, the move is about to be described. When used in combination with to i.e. in a move to, it is a tactic for information compression, it allows the writer to state very concisely the intention behind the move before giving the actual details of the move itself, as in line 28: –In a move to tighten control of a far-reaching empire and to improve the group’s own image, Maurice and Charles Saatchi, credited with building up the company, have stepped down from the day-to-day running of the group.

In other phraseologies, it permits the writer to give his/her opinion of the possible outcome of the move, in a similar brief fashion: in a move that could (l.27), in a move which bids fair to (l.30), in a move they may later regret (l.31). There would seem to be a high probability of a modal being found in this environment, since future possibilities are being discussed. The most typical realisation in the concordance above is a move in clause initial position, in a kind of illustrative apposition elucidating a previous clause, usually followed by either a relative or a participle, as in line 7: –He effectively cast his own vote against the Islamic trend by calling on his subjects not to vote for extremist religious ideologies—a move that could prove counter-productive.

As with the phrase in a move which/that, where a move is followed by a relative clause, this latter frequently contains would or could or other modal. There is a function of prediction expressed in terms similar to a condition—‘‘if a particular course of action were to be taken, such a move would or could have the following consequences’’. A number of the examples have to do less with the practical result of an action than how it will be received. A move used in this way tends to collocate with opinion or attitude items such as favoured (l.2), welcome (l.21), unwelcome (l.16), disgust (l. 17). This focus is particularly common when a move is followed by a participle: here we find: regarded (l.19), favoured (l.2) and, once more, welcomed (l.12) followed by by. These seem to be typical phraseologies used by journalists to describe

cohesion in texts

95

what someone else thinks of a course of action (or what the journalist claims they think). The limited evidence here also suggests that different phraseologies are employed to express when the course of action has been completed and when it is prospected. Relative clauses with would or could seem more typical of uncompleted moves, while reporting participles seem more frequent when the action is completed. Nevertheless there also seem to be some participles—examples in the above concordance being favoured and ruled out (l.4) which imply that the move has yet to be taken. In fact, line 2 combines both a ‘‘future implying’’ participle (favoured) and a relative clause with would. The concordance evidence also suggests that, on the whole, newspapers talk at least as frequently about moves yet to occur than ones already completed. Examples 26 to 31 all contain in a move to or in a move that/which. In each case with a cataphoric function—the nature of the move has yet to be described. But while in a move to—at least in these examples—usually introduces information about the purpose of the move e.g. to tighten control of a far-reaching empire (l.28) or to rally support (l.29), in a move that/which introduces a clause describing some other kind of effect it might have while the precise nature of the move is relegated to later in the text. In other words, the purpose or result is stated before a description of the move itself, as is clear if we look at the wider co-text of these examples (ll. 30 and 31): –IN A move which bids fair to relegate the phrase ‘taking coals to Newcastle’ to the lexicographical dustbin, the makers of a new dollars 15 million film have decided to build a slum in Calcutta. –In a move which they may later regret, Jayne Torvill and Christopher Dean have lent their name to a challenge cup for the team who raises the most money for charity.

The final group of examples is in the combination such a move which is clearly anaphoric. Eight of the sixteen examples are followed by would, and another by will no doubt. Curiously, there are no examples with any other more tentative expression than would (could, may, seems likely to) which were found with a move in other phraseologies. There also seems to be a tendency for such a move to appear in contexts where either someone has reservations about the move, or where it has limited or negative value for someone: –such a move for Mr and Mrs. Evans was not practical (l.35) –[. . .] must recognise the folly of such a move (l.37) –Such a move will no doubt alarm the military analysts (l.39) –such a move would have to come one step at a time (l.45) –the net effect of such a move would be to cut Labour’s majority [. . .] (l.47).

96

alan partington

The overall association of such a move is that the move is so dramatic that it needs thinking hard about. This is supported by the only example in the corpus where such a move is found with an adjective: –Mansell is unlikely to make such a drastic move and he will be racing until the end of the season.

6.1.3 Suggestions for further research on general noun phrases Halliday & Hasan’s list contains a number of general nouns which might be fruitfully studied. However, it should be borne in mind that general nouns are an open set, with speakers/writers using nouns cohesively all the time, and so this list is only an indication of the possibilities. One particularly interesting area for future study comprises the items which commonly modify general nouns and are used to express a writer’s attitude to the noun referent. For example, a search through the newspaper corpus for ‘‘the old *’’ threw up a good many general noun phrases e.g. the old adage, the old argument/approach/attitudes/notion/fears and others, where old expressed the writer’s distance or disapproval of the referent. In other cases, where it was usually used semi-humorously of people, old often expressed quite the reverse: the old trouper/Bogeyman/codger/warhorse were all found (oddly enough) to communicate closeness or sympathy. Finally, the study of a move highlighted a number of phraseologies involving general nouns which seemed typical of journalistic prose. Corpora consisting of different text types could be compared to discover differences in phoric techniques between genres.

6.2 Text reference 2: Labelling noun phrases 6.2.1 The functions of labelling noun phrases In a dissertation published at the University of Birmingham, G. Francis identifies a group of items she calls ‘‘anaphoric nouns’’ (also called ‘‘A-nouns’’). These are noun phrases which have the function of ‘‘encapsulating’’ a foregoing piece of text or, rather, the concepts and argumentation contained within it. Her first example is the item position: J.R. Lucas, in a famous article published in Philosophy in 1961, argued that the most important consequence in Godel’s work was that the human brain cannot, in principle, be modelled by a computer programme—that minds cannot be explained as machines. For, although computers can be programmed to generate formal systems, they can

cohesion in texts

97

never be programmed to spot the Godelian traps inherent in them. This latter ability, Lucas argued, remains the sole prerogative of the human brain. Surprisingly, perhaps, Hofstadter disagrees with this anthropocentric position [. . .]—(G. Francis 1986: 4)

Such encapsulation is useful since it allows the text-producer both to indicate where the argument has been and to add new information to the argument to allow it to be developed further, as in the example above (where we learn that the argument is ‘‘anthropocentric’’, that Hofstadter disagrees with it, and that this might come as a surprise). In many cases, it also permits the author to add an interpersonal ‘‘value judgement’’ to the preceding proposition, either by the choice of a certain type of A-noun (e.g. compare this account and this concoction) or by qualifying the noun in some way (consider such an illogical notion). A-nouns can be seen as a special category of general nouns. This latter groups together items with a whole range of functions—from those which usually have a very precise referent,through those which can also refer more extensively, to those which can perform text reference. G. Francis uses the term A-nouns only for those items which are functioning in text reference mode, which label text as ‘‘a fact or a report’’, and which have a definite role in moving the discourse on: [A-nouns] refer to and label a stretch of discourse, aligning it with the ongoing argument, which now continues in terms of what has been presented metadiscursively as ‘‘fact’’.—(G. Francis 1986: 29)

Francis returns to this theme in an article entitled Labelling discourse: an aspect of nominal group lexical cohesion (1994). Here she includes noun phrases with a cataphoric function in her analysis, and adopts the terminology advance and retrospective for, respectively, cataphoric and anaphoric noun groups used to label stretches of a text. Advance labels have a predictive as well as an organising function, indicating to the reader how following information is to be interpreted on both an ideational and interpersonal level. 6.2.2 Examples: allegation and claim Surprisingly—given the wealth of her examples—Francis carried out her original 1986 study on a non-electronic corpus, collecting instances ‘‘by hand’’, mostly from the monthly magazine Encounter. As she notes in the later article (1994: 100), with the availability of computer corpora it has now become possible to study text in detail with concordancing tools which allow the analyst to obtain examples of particular labelling noun groups relatively quickly. In this section, we look at the use of three items from the semantic group of ‘allegation’: allegation, claim and asser-

98

alan partington

tion. Only those examples which occurred within a few words of the beginning of a paragraph were collected, since this is the most likely place to find such nouns used as labels (as anaphoric labels at least, which are by far the more common type). Allegation and claim were found to have a much higher frequency in newspaper texts relative to the other genres in the corpus. This is probably because newspapers devote a good deal of space to reporting other people’s words. G. Francis (1986) divides A-nouns into four classes. The use of a ‘‘factive’’ Anoun, such as acknowledgement, disclosure or realisation implies that the proposition referred to is true, or that an event really happened. A ‘‘half-factive’’ A-noun, such as declaration or explanation is one that strongly suggests that something happened or is true, or one that is sometimes used factively and sometimes non-factively. A ‘‘non-factive’’ A-noun such as claim or allegation is one that, superficially, is noncommittal, while a ‘‘counter-factive’’ A-noun, such as concoction, fabrication implies the proposition or event referred to is false/imaginary (Francis 1986: 25–6). The items under examination here belong to the non-factive group. In fact their ‘‘non-factiveness’’ is perhaps the most important aspect of their meaning and function. They are used precisely when it is necessary to imply a distance between the reporter and a reported statement: ‘‘I’m telling you what they said, but I don’t necessarily believe them’’. Perhaps one could go even further. In some circumstances, writers are supposed to be under an obligation to be fair and reasonable—for example in news reporting or legal reports. In such texts, the labelling nouns in this group are frequently used in an apparently neutral fashion, as in the following example from the four-million word Independent corpus: (11)

The allegation that the Royal Ulster Constabulary contains an ‘Inner Circle’ of officers preparing private lists of IRA suspects and working against the Anglo–Irish agreement has lent a new dimension to the security force-loyalist collusion saga. The claim came in yesterday’s Irish News, Belfast’s Catholic morning paper. (The Independent)

In argumentative text, on the other hand, this obligation is relaxed. The choice of a non-factive word rather than a factive or half-factive one (say claim in preference to fact or account) is often a clear message that the foregoing thesis is being dismissed, as follows: (12)

Jana contended that the Phoenician empire was Arab and therefore Venice was once part of a Pan-Arab civilisation, not to mention the other claim that many Venetians have Arab roots from ninth-century traders. Venetian historians promptly discounted the claim.—(The Independent)

cohesion in texts

99

Words from the semantic field of ‘claim’—claim, allege, assert as verbs, along with related nouns claim, allegation, assertion—very frequently appear together in the same text. The cohesive web they weave can be very intricate. The Ulster story cited above continues in the following day’s paper (my bold type and italics): (13)

Headline: Annesley dismisses RUC ‘inner circle’ claim as nonsense By David McKittrick, Ireland Correspondent Hugh Annesley, Chief Constable of the Royal Ulster Constabulary, dismissed as ‘arrant nonsense’ yesterday the idea that a conspiratorial group of pro-loyalist politicallymotivated officers existed within his force. He was responding to Monday’s claim in the Belfast Irish News that at least three dozen officers were members of a secret ‘inner circle’ which had the objects of ‘removing’ republican suspects and bringing down the Anglo–Irish agreement. Mr Annesley’s strongly-worded statement came as the agreement ran into yet another patch of stormy weather [. . .]

The item claim in the headline functions as both new information as befits a headline, but it also creates a cohesive link with the previous day’s article on the story— or rather with the readers’ knowledge of that story. The phrase the idea in the first paragraph is a cataphoric signpost to the statement following that (i.e. a group of pro-loyalist politically-motivated officers existed within his force), but it is also linked to the claim of the headline and therefore is part of the chain connecting this article to the previous one. The same is true of claim in the next paragraph, but the latter item also links this text, exophorically, to another one—that article in the Belfast Irish News which first printed the allegations. Finally, in the third paragraph, we find statement, which is also partly exophoric, relating presumably to a press communication issued by the policeman, containing his dismissal (see headline and first line of the article) of the allegations, of which the only part quoted verbatim in this article is the phrase ‘‘arrant nonsense’’. Another example of a mixture of the anaphoric and exophoric is to be found in this example: (14)

In the book, based on interviews in Europe, the Soviet Union, the United States and Australia, Mr Aarons writes: ‘‘Most of the Nazi war criminals and collaborators who found sanctuary in this country proclaimed themselves as nationalists and antiCommunist ‘freedom fighters’. Ill-informed and blinkered Australian politicians and officials did little to question or probe their credentials and histories.’’ He asserts that the Australian Security Intelligence Organisation (ASIO), Australia’s MI5, later used many former Nazis for anti-Communist counter-intelligence operations. As a result of similar allegations by Mr Aarons three years ago, Bob Hawke’s Labour

100

alan partington government set up an inquiry under Andrew Menzies, a former senior civil servant, who examined thousands of secret files including one provided by Mr Aarons.

Here, a particular allegation is described in some detail, but the phrase similar allegations tells the reader that what provokes the further developments outlined in the second paragraph is not quite the allegation contained in Mr Aarons’ book, but some event which took place in the world outside this text. Items such as similar, same, different, next, latest, further, other and another, as well as the ordinals second, third etc, when functioning in conjunction with a label are called by G. Francis (1994) ‘‘textual modifiers’’, which ‘‘contribute directly to the organisational role of labels: they help to order messages with respect to each other and signal the relationship between them’’. A final example illustrates the dual anaphoric-cataphoric function of many occurrences of labels. Francis stresses how they are a platform for launching into new developments of the topic under discussion, but it is sometimes true that they are themselves in need of further explanation if their reference is to be understood: (15)

Last night the Olympics’ major sport was embarrassingly reminded of them [the scandals which have occurred in athletics during the presidency of a certain Sig. Nebiolo] when Granada Televison’s World in Action chose the ASOIF election week to catalogue them again in a highly critical examination of Nebiolo. The latest damaging assertion came from Sir Arthur Gold, chairman of the British Olympic Association and former president of the European Athletic Association, who claimed that Nebiolo and an associate, when they were Italian athletics federation officials, tried to manipulate the result of a European Cup semi-final in 1973.

The word assertion is undoubtedly linked to other allegations of scandals which have already been mentioned in this text, but the precise nature of this latest assertion is in the text which follows and is therefore cataphoric. Finally, the phrase latest damaging assertion is interesting as an example of a textual modifier— latest—used in conjunction with another modifier—damaging—with strongly interpersonal meaning. As noted above in relation to general nouns, Halliday & Hasan’s statement that ‘‘a general noun in cohesive function is almost always accompanied by the reference item the’’ seemed to be too limiting, and a good number of other reference items were also found with these labels. In the particular case of claim words (and perhaps in general the whole class of illocutionary nouns), the phrase including the noun also frequently includes some indication of the author of the illocution, e.g. The agency’s allegations, Mrs Thatcher’s claim and China’s recent assertion. In one of the above examples, the noun phrase includes mention of the time the claim

cohesion in texts

101

was made, with the author in a post-modifying phrase—Monday’s claim in the Belfast Irish News. 6.2.3 Suggestions for further research on labelling noun groups Other interesting labels which the teacher–researcher might like to investigate via corpus data include the following: argument, notion, hypothesis, boast, controversy, conclusion, conjecture, distortion and version. Many more can be found in the lists given in Francis in both works quoted above. As has been pointed out, if the search for such items is limited to the first few words of a paragraph, there is a high probability of finding them used anaphorically as labels. A similar search could be carried out on the textual modifiers which sometimes precede labels, to see precisely how they function. These are listed in the section above. Another useful study would be to access articles from the corpus relating to the same news story published on different days (the later ones being likely to contain reference back to previous articles) to see what kinds of labels are used in the headlines and the initial paragraphs of such texts.

6.3 General verbs The term ‘‘general verb’’ is not new to text grammar, but in this study it is used to indicate a rather different set of items from that referred to by Martin (1992) among others. Martin uses it to indicate verbs such as make, take, have and so on which take part in lexical verb phrases like take a look, have a shower etc. I prefer to use Sinclair’s (1991a) terminology for these items and call them ‘‘delexicalised verbs’’, thus retaining the term ‘‘general verb’’ to indicate verbal items which can perform a function as text linker, in parallel with the term ‘‘general noun’’.5 In this section we will look more closely at this function, concentrating on examples of the use of the items occur and happen. All the examples in this section are from the newspaper corpus. There seems to be a close parallel between the cohesive activity of general nouns, including labelling noun groups, and this group of verbs. Both are technically lexical items, but perform a referential function more generally associated with grammatical pro-forms. Like many general nouns, these verbs refer to a stretch of preceding or succeeding text, that is, they are commonly devices for carrying out text or extended reference.6 6.3.1 Happen Happen is frequently found to stand for an action/event or series of actions/events

alan partington

102

which are recounted in some other part of the text.7 It is sometimes used as an anaphoric signal to refer to an action/event already recounted, and is often accompanied by this and occasionally by that: (16) (17)

First, consider the situation where weight loss is not as good as expected. This must happen to most dieters at some time or another. The hope was that the morphogen in the mixture would slowly leak out and specify new digits. If that happened, we could then try to purify the morphogen.

Or it can be used cataphorically to introduce some action/event about to be recounted, often in pre-constructed phrases such as it often happens that, as so often happens: (18)

(19)

By analogy, it may sometimes happen that a message is improved by changing, say, an S into a G: my secretary once, misreading my writing, replaced ‘‘Sod’s law’’ by ‘‘God’s law’’. As so often happens, the passage of time showed an increasing number of disadvantages of the treatment.

On occasions, happen is found in rhetorical questions. Authors ask a question of the type what happens when? before going on to give an answer themselves. We can include this device in the cataphoric use of this item: (20)

What happens when you go on holiday for a few weeks, or you are bed-ridden with influenza, or you change job or move house? Ideally, provided you are not ill, you can carry on with some exercise, even if it less than you’re used to.

What happens in an indirect structure has a similar cataphoric function: (21)

We should consider what happens when one moving body comes into contact with another [. . .]

which is followed by a description of ‘‘what happens’’. 6.3.2 Occur Occur, too, can function as a text linker in a wide variety of ways. It is often found following the anaphoric markers this, that, these and those (with this and these being more common): (22)

private property and slavery come into conflict with the community and its commonality; this occurs only when the family begins to move into the political domain.

cohesion in texts (23) (24)

103

think about the circumstances in the past in which you have achieved some modicum of success, and analyse why this occurred. [. . .] either by the stage of the city state or by feudalism. Which of these two occurs seems to depend, for Marx and Engels, on population density.

It is used with a general noun such as event, action, process usually with a cataphoric function: (25) (26) (27)

and how this may create other problems, particularly when a major life event occurs, such as the break-up of an important relationship. The first of these incidents occurred with the attempt to introduce a law on adoption. An example of the strength of this animus occurred in 1982–3 in the celebrated case of the Revd David Armstrong, censured by the elders of his congregation.

On occasion it can be both forward and backward looking, with the word same or similar somewhere nearby: (28)

From their rule of the eastern Frankish kingdom developed Germany. Now the same process occurred all over again.

When it is found with the word such and some general noun the phrase has an anaphoric function: (29) (30)

Some patients have a history of hospitalization and surgery before the age of 5, and in some cases such events have occurred before the age of 18 months. No clearer way of indicating the path to be trod could have been made. It is possible that such authoritarian actions will occur again.

Occur also appears with items such as much the same thing, the first to form a phoric phrase: (31)

(32)

[. . .] the politics of culture itself or the political questions raised by the form as well as, and in relation to, its content. Much the same thing occurs in New Statesman and Society. There are two such fairs in ‘‘Cabbagetown’’; the first occurs early in the novel when Timothy Place has its annual street dance and bazaar [. . .]

Occasionally it even refers to itself: (33)

Although it occurred more slowly than for subjects given non-reinforced preexposure, loss of the OR occurred in control subjects too.

alan partington

104

From the data, it would seem that an important function of occur is to help writers nominalise an action or a process: (34) (35) (36) (37)

In its original form the Pearce-Hall (1980) model applied this analysis only to the loss of associability that occurs with CS-US pairings. Since her return the couple have felt more positive about their relationship and no violent behaviour had occurred. Loss of effectiveness in a shock US is amply demonstrated, for instance by the loss of the CR that occurs with prolonged training. [. . .] those of a promiscuous nature, and particularly those whose first sexual intercourse occurred at an early age.

The italicised sections consist of actions or processes which have been turned into a noun phrase. But when this happens, writers need to accompany the noun phrase with a general verb like occur in order not to lose all sense of action/process. Occur performs this function more frequently than happen thanks to the formality of the former which means it tends to be found in formal and scientific registers where the kind of nominalisation described is very common.

6.4 Conclusion This study had two aims. One of these was to review the cohesive role of general items, using concordancing. As far as general nouns are concerned, evidence was found showing them to be much more flexible in both form and function than has often been suggested. It was seen how general noun phrases can contain a variety of reference items, not just the but also a, demonstratives, such a, sort of and so on. Evidence was also unearthed of how particular general noun phrases such as in a move to or such a move are used to perform particular functions in newspaper registers. The need to re-appraise the cataphoric function of these items was also clear. In the study of labelling noun phrases it was seen how these items frequently fit into a complex network of cohesive items, how they play a part, along with general nouns, other reference items and the use of lexical reiteration, in the construction of lexical chains. Finally, an outline was given of how the concordance data sheds light on the seldom-discussed phenomenon of general verbs, in particular happen and occur. These items enter into ana- and cataphoric relations with the surrounding text in a wide variety of ways. Certain general verb phrases were seen to perform specific functions, for example what happens if/when to introduce rhetorical questions, as so often happens to introduce the discussion of a certain phenomenon, and the special nominalising use of occur.

cohesion in texts

105

The other aim of the study, implicit in the first, was simply to show that concordance technology can usefully be employed in studying linguistic phenomena at a textual level, even using a corpus of fairly limited size. In fact, data from even much smaller corpora, containing generically homogeneous texts, could be compared to discover phoric techniques in particular genres, an area in which little work has yet been done.

Notes 1. A word on terminology is needed here. The term text is used with many different senses in linguistics, often in contradistinction to discourse. For some, text is reserved for written language, discourse for spoken (see, for instance, the Preface to Coulthard (ed.) 1994). For others, discourse means any stretch of naturally-occurring language, which becomes text when subject to the linguist’s analysis (those subscribing to this definition would do well to use them as countable terms—i.e. to talk of a discourse, discourses and a text, texts; unfortunately they are rarely so used). For still others, discourse, as in academic discourse, racist discourse or legal discourse, signifies all the real and potential texts produced in certain professional sectors or social situations (see, for example, Halliday’s (1994) discussion of scientific discourse). In this book, I subscribe to Stubbs’ definition: ‘‘By text, I mean an instance of language in use, either spoken or written: a piece of language behaviour which has occurred naturally, without the intervention of a linguist’’ (1996: 4). This is a rather good definition to use when analysing corpora, which contain authentic data. 2. ‘‘The expression of interpersonal meaning, of a particular attitude on the part of the speaker, is an important function of general nouns […] There are quite a few general nouns which have this interpersonal element as an inherent part of their meaning, especially those referring to human beings, for example idiot, fool, devil, dear. But whether or not it is inherently attitudinal in meaning, a general noun in cohesive function can always be accompanied by an attitudinal Modifier. So we have examples such as the dears, the poor dears; and also the stupid thing, the lucky fellow and so on’’ (Halliday & Hasan 1976: 276). 3. Martin reminds us that we can tell when an item is being used as a general noun because it does not carry a tonic, even when it is the last salient syllable in the tone group. He gives the example: Melbourne’s a terrible place; it actually hailed on me (where place is a general noun) (1992: 287). 4. The terms anaphoric, cataphoric and exophoric are commonly used to describe three ways of referring. Anaphoric reference occurs when the referring item (pronoun, general noun phrase etc) appears later in the text than the expression it relates to. Cataphoric reference occurs when the referring item appears earlier in the text than the expression it relates to. Exophoric reference occurs when the referring item points to something outside the present text, usually something in another text or existing in the real world outside the text. 5. There seems to be even less literature on general verbs than on general nouns. There is no mention of this use of happen words in such general works on lexis as Carter (1987) or Carter & McCarthy (1988) or in Martin’s (1992) grammar of text, and it certainly does not get a mention in any of the

106

alan partington

dictionaries I looked at (Cobuild, LDOCE, CIDE and the Oxford Advanced Learner’s). A footnote in Quirk et al. (1985: 818) mentions in passing ‘‘the generalized event verb happen’’. 6. The analogy between the two is even clearer if we take into account the fact that general verbs, like general nouns (see note 1), can function pragmatically as vehicles for the speaker/writer’s attitude or judgement. They often appear with modals of probability, for instance: may/could/is likely to happen etc, and occasionally with adverbs or adverbial phrases of attitude, for example presumably/allegedly occurs and inexplicably happens. They can express not only the author’s attitude, but also that of other actors who are the object of the discourse (see the second example in note 7). 7. It would be misleading to call happen when used this way a ‘‘pro-verb’’ because in some cases, no particular verb is recoverable in the surrounding text e.g. • he regarded VD as something that happened to other people. • their specifically homosexual perception of the double-edged nature of romantic love, simultaneously wanting it desperately to happen but [. . .]

7 Metaphor It’s a magical-realism love story. A bit like following Everton (Corpus).

This chapter describes an experiment in the study of metaphor.1 In particular, the area of metaphor in business journalism is analysed. Along with concordances, vocabulary frequency lists are used. A number of theoretical implications concerning metaphor itself arise from the study.

7.1 Systematic metaphors and genre-typical metaphors One of the most influential of modern investigations into the nature of metaphor is that by Lakoff & Johnson, entitled Metaphors We Live By (1980). Briefly, in this work the authors note that many everyday metaphors relating to the same topic can be grouped together into a mega-metaphor on a semantic basis. One of their examples is the group of metaphors such as Is that the foundation for your theory?, His argument collapsed, We need to buttress the theory with solid arguments which add up to a mega-metaphor of the form theories are buildings (the use of capitals is their convention). The authors call such sets of metaphors ‘‘systematic metaphors’’. Similarly, metaphors like His ideas have come to fruition, We need to nip that idea in the bud, The seeds of these ideas were planted in her youth are all part of a larger systematic metaphor ideas are plants. They argue the existence of several other systematic metaphors including argument is war (Your claims are indefensible, He shot down all my arguments and so on), time is a valuable commondity (Is that worth your while?, I’ve invested a lot of time in her etc), communication is sending, love is a journey. Metaphors We Live By is a highly theory-driven work (in the area of conceptualpsychological theory), and the authors use invented examples rather than authentic data to support their arguments. I therefore thought it worthwhile to search for evidence for the existence of systematic metaphors in corpus data. Moreover, Lakoff & Johnson unfortunately pay no attention to the question of (using the term loosely) genre. Certain metaphors may well be much more prevalent in one kind of writing than another, in fact, one of the characterising features of a genre is probably the

108

alan partington

kind of metaphor generally to be found therein. Data from the corpus, it will be remembered, can be accessed according to a number of different (mega-)genres (the Arts, Business etc sections of the newspaper, medical, sociological etc texts in the academic corpus), and thus it is possible to collect empirical data on genre-typical systematic metaphors.

7.2 An example search An example investigation in this direction was sparked by the headline of an article in the Sports section of the corpus, which read Guscott’s magic ruins Gloucester’s fairy-tale (Guscott being a rugby player and Gloucester the team he played against). On this (admittedly scant) evidence, the existence of a couple of related systematic metaphors—the sportsman is a magician and sport is a fairy-tale— were hypothesised. By concordancing ‘‘magic*’’ (giving occurrences of magic, magical and magician) from the Sports section of the corpus, a good deal of data was forthcoming in support of the existence of these metaphorical sets. There were thirty occurrences of either magic or magician, of which one fifth are reported below: 1. Even with pools of water to dance around on a saturated pitch, their midfield wove magical patterns. 2. It’s a magical-realism love story.’ A bit like following Everton. 3. Rugby Union: That old Black magic ready to cast its spell. 4. The Yorkshire-based former member of two Chinese world title-winning sides, conjured some new magic for England to beat his aggressive flat-hitting namesake Chen Hong-Yu. 5. With Hurlock and Horne holding the middle, Matthew Le Tissier, Saints’ mercurial magician, was able to display some delightful touches. 6. The whole magic of the game is that small clubs like Ipswich, Swansea, Norwich and Northampton can come up.

Note that the metaphor is sometimes serious, at others self-conscious and tongue-incheek (ll. 2 and 3). The concordance of fairy-tale contained three occurrences in which a sporting career is described as having ‘‘a fairy-tale ending’’, including this example when the career belongs to a horse: –[. . .] while in 1975 his career ended on a fairy-tale note when he became the first Gold Cup winner since Golden Miller to win the Grand National.

Line 4 above contains the word conjured and a concordance of ‘‘conjur*’’ from the Sports section contained fifteen occurrences, of which every third one is reported here:

metaphor

109

1. France conjured a breathtaking try on the counter-attack, by the new flanker Philippe Benetton. 2. Ivanisevic, wearing a protective support to counter a pain in his side, conjured a shot from behind his back during the sixth game of the final set. 3. If Wayne Shelford has been a sorcerer conjuring Northampton’s rise and rise, then Tim Rodber has been his apprentice. 4. At this stage, Davies was three strokes ahead of Connachan, Moon and Italy’s Federica Dassu, who had conjured two eagles in an outward run of 31. 5. oldham athletic’s capacity for conjuring up the outrageous earned an unlikely point last night after Eric Cantona returned to the scene.

However, in order to ascertain whether the magic/magician metaphor is especially prevalent in sports journalism, the items magic*, fairy-tale, conjur*, wizard and bewitch* were also concordanced in the other newspaper corpus sections. Occurrences of these items were generally much rarer in the Home, Foreign and Business sections than in the Sports, but both magic* and conjur* were more frequent in the Arts. This is because, firstly, many of the mentions of magician in this section are not metaphorical but refer to a real-life occupation, and, secondly, the Arts texts are full of references to the ‘‘magic’’ of the theatre and cinema. Similar counts were conducted on the Home, Foreign, Business and Sports sections of The Times CD-ROM material (there is no section explicitly containing Arts texts), and all the items from the semantic field of magic were found to be far more common in the Sports writings (e.g. bewitched occurred 22 times in the Sports section, twice in the Foreign news, once in the Business and not at all in the Home news section). These results seem to confirm the special status of the magic metaphor in sports journalism.

7.3 Theories of metaphor There is a second, more theoretical, aspect to the question of metaphor to be considered. Lakoff & Johnson make some very strong claims about the role of metaphor itself in human cognitive processes. They challenge what they see as the prevailing conventional view of metaphor as simply poetic device, as rhetorical icing on the cake of literalness.2 Metaphor is not just a feature of language, they argue, but the very vehicle of normal thought and action: we have found [. . .] that metaphor is pervasive in everyday life, not just in language but in thought and action. Our ordinary conceptual system, in terms of which we both think and act, is fundamentally metaphorical in nature.—(Lakoff & Johnson 1980: 3)

110

alan partington

Gibbs (1994), primarily a psycholinguist, takes up their mantle and criticises modern pragmatic linguistics, both speech act theory and relevance theory, for misinterpreting and undervaluing the role of metaphor. Speech act theory, argues Gibbs (citing Searle 1979), explains metaphor in terms of anomaly, of a mismatch between sentence meaning and speaker meaning. If a hearer hears an utterance such as Sally is a block of ice they perceive it as anomalous, as logically defective and flouting the Gricean maxims of quality or relevance (Grice 1975, 1978), and they therefore cast around in their knowledge-of-the-world faculty for an alternative, non-literal interpretation. Gibbs rejects this implication that hearers, in order to fathom metaphorical meaning, must employ cognitive processes which are both different from, and in excess of, those used to fathom literal meaning. In his view, there is no priority of one kind of meaning over the other, they are both natural to the process of thought: it is misleading to suppose that one type of meaning (literal) is automatically and immediately prior to another (non-literal).—(Gibbs 1994: 228–9)

He cites his own experimental evidence in support of the argument that hearers expend no more processing effort on one kind of meaning than another. Sperber & Wilson’s relevance theory (Sperber & Wilson 1986, 1995) does not treat metaphor as flouting relevance, but rather as one way of what is called ‘‘speaking loosely’’. Their view is that an utterance is relevant if it has contextual effect, that is, if it strengthens, contradicts or denies an existing assumption, or when it implies something new. An utterance is optimally relevant if it (a) achieves enough effect to be worth the hearer’s attention and (b) puts the hearer to no gratuitous effort in winning that attention. In this framework, speakers are not constrained to saying what they know to be true, and speaking loosely, speaking metaphorically may well be the most effective and economical way of achieving contextual effect, of getting through to the hearer. Nevertheless, according to Gibbs, Sperber & Wilson still imply that the processing of metaphor requires more work than does literal utterance, even though this extra effort may be offset by the extra effect obtained by the force of the metaphor. The metaphor-as-loose-talk view, then, does not assume that metaphor requires special cognitive processes to be understood but it does imply extra processing effort. Gibbs repeats his claim that experimental evidence shows that this is not the case. We thus have three different views of the role of metaphor vis-à-vis the literal: those of the speech act theorists, those of Sperber & Wilson, and those of Lakoff & Johnson and Gibbs. Each of these approaches is highly theory-driven, none of them use authentic data to support their arguments. In this study, therefore, an analysis was carried out into the nature of the metaphors contained in a large body

metaphor

111

of texts to see if an inductive, data-driven study could shed light on this debate. The research was restricted to one genre, business journalism, for the reasons already expounded in the previous section, that is, that much of metaphor is genre-distinctive or specific.

7.4 Metaphor in business journalism: Data and methodology The database for this study consisted of the financial and business sections of the corpus (containing circa 800,000 words of running text from The Independent (1992) business section), the business sections of The Telegraph newspaper of 1993 and of The Times of 1993, held on CD-ROM. Paper issues of The Economist magazine were also used for collecting or checking data. In terms of the methodology involved, it was not sufficient to refer to the business sections of the above newspapers in isolation. If, for example, it was discovered that a systematic metaphor existed of the type business is x (say war or gambling), this would be significant only if this metaphor was particularly prevalent in this sectorial language. If metaphors were also found of the type that business is x, sport is x and art is x, this would mean that the metaphor was common to journalism as a whole, perhaps even to most genres of the English language. Thus all findings were compared to data from the other sections of the newspaper corpus and CD-ROMs. As has been said, an inductive approach was desirable in order to let the data itself suggest what kinds of metaphor might be found in business journalism, rather than guessing a priori what these might be. It was decided that the first step would be to find what the principal vocabulary differences were between business texts and other newspaper texts, since these differences might give some insight into which metaphors were more frequent in this sector. If, to take an example, gamble/ gambling was particularly common in business texts in comparison to others it would then be worthwhile to look at related words such as betting, casino, chips and so on, to try to discover if a mega-metaphor of the type business is gambling existed. For this purpose a program called WordSmith Tools,3 written by Mike Scott was used,which first of all prepared a list of the most frequent lexical items in the business texts on hard disk and then another of all the other newspaper section texts treated together. By comparing the two lists, the program provided a third list of all words which were significantly more frequent in business texts, and another of those which occurred significantly less frequently. Both lists proved to be useful. A final non-computerised source of data was a short dictionary of business terms entitled Business Buzzwords and subtitled The tough new jargon of modern business (Michael Johnson 1991). Though semi-humorous in style, it was written by the

alan partington

112

editor of International Management, a prestigious European business journal, and offered an interesting reference against which to compare and contrast the newspaper findings.

7.5 Metaphor in business journalism: Results The frequency lists showed one vocabulary set to be highly typical of business texts: words relating to ‘up’ and ‘down’. This was perhaps not surprising, but what was striking were the variations on the theme. ‘Up’ and ‘down’ were expressed in all the following ways: up above high(ly) raise rise grow(th) jump escalate climb spiral soar mount

down below low fall slip drop slump slide tumble crash plunge collapse

Closer investigation of the material and of Business Buzzwords also revealed: upmarket downmarket upstream downstream up and running downsize up-front Lakoff & Johnson discuss up-down metaphors at some length as a central example of what they call ‘‘orientational metaphors’’, which are based on spatial orientation. Other such metaphors include in-out, front-back and so on, and for Lakoff & Johnson they are important because they ‘‘arise from the fact that we have bodies of the sort we have and that function as they do in our physical environment’’ (1980: 14). They are the leading proof for these authors that metaphors are the result of experience and are therefore basic, natural features of thought and action. They point out that the up-down metaphor has a number of extensions, two

metaphor

113

of which are highly relevant to the field of business, namely the metaphors up is more, down is less and up is better, down is worse. There is probably no sectorial language in which the metaphors up-down are more apparent than in business. Concentrating on the up part of the metaphor, there is certainly no lack of evidence to show that up is more: 1. 2. 3. 4. 5. 6. 7. 8. 9.

[. . .] helped the Hang Seng index to climb 39.95 points to 2,786.65 in moderate trade. BM shares rise as profits climb 78 percent. [. . .] warned that interest rates could climb even further to clamp down on rising prices. [. . .] predicted that unless the market can climb above dollars 2,810 [. . .] [. . .] was reflected in a 266.66-point climb by the Nikkei average to 35,374.22. The Cisco index closed at a new high of 185.6, a gain of 2.1 points on the week. It aims to achieve a high level of income for ordinary shareholders. Jaguar shares jumped from 36p to 669p. The half-term dividend is up 15 per cent to £1.95p.

However, it is by no means the case that up is always, or even predominantly better. When, for example, costs, debts, inflation or unemployment are up, then this is far from good. In fact there are even items, notably mount, escalate and spiral which are all used in the business subcorpus specifically to express that up is definitely worse: Edited concordance of ‘‘mount*’’ (excluding the sense ‘mount a bid’, every second occurrence): 1. Mr Lapointe and his secretariat are facing a mounting barrage of accusations [. . .] 2. If tensions mount between and within the former Soviet Union [. . .] 3. It follows mounting concern in the West over the build-up of some $600 million dollars in interest arrears. 4. [. . .] amid mounting controversy over the conduct of [. . .] 5. Mr Seeling abandoned his lawyers and opted to defend himself after growing alarmed at the mounting cost of his legal bills. 6. James Rodwell, 32, a company director facing mounting debts [. . .] 7. [. . .] in 1942, at a time when RAF aircraft losses were mounting disastrously. 8. [. . .] would expose them to the prospect of mounting fines and eventual seizure. 9. [. . .] declined to comment on the effect BR’s mounting losses might have [. . .] 10. Mounting losses at Dan-Air 11. Calls for early return to ERM reinvigorated as alarm mounts over economic policy. 12. the pound plunged on the foreign exchanges with concern mounting over the continued absence of a coherent economic policy. 13. with Pima borrowers defaulting on their loans and losses mounting, Pima went bust. 14. PARIS: mounting political uncertainty sent the CAC–40 index down [. . .]

114

alan partington

Edited concordance of ‘‘escalat*’’ (every second occurrence): 1. but by escalating a guerrilla war on the lira to a frontal assault [. . .] 2. The company also had to contend with the escalating cost of both public and employer’s liability. 3. The escalating costs of the Magnox stations [. . .] 4. hopes of raising up to £1.6bn to meet the escalating costs of the Channel tunnel. 5. Despite the jitters the cost escalation has caused in the City [. . .] 6. The insurers can be accused of having mistakenly relied on the continued escalation of house prices. 7. BA alliance row to escalate. 8. [. . .] demanding immediate repayment that would have escalated the country’s financial crisis. 9. and the present value of the pound look like escalating the slowdown in the economy. 10. [. . .] sharp falls in prices combined with escalating unemployment and repossessions. 11. The LSB also warns in its International Economic Outlook today that an escalating world trade war, instability [. . .]

Edited concordance of ‘‘spiral*’’ (every second occurrence): 1. 2. 3. 4. 5.

[. . .] and political uncertainties that have become a dreadful spiral. the need for drastic cuts in spending to avoid succumbing to the spiralling costs of unification. took fright at indications that the spiralling costs of the Channel tunnel would not [. . .] Spiralling interest rates have tightened the noose around their necks. With public sector borrowing spiralling, Norman Lamont did not in truth have much scope [. . .]

These three items clearly display an unfavourable prosody (see Chapter 4), being reserved for the rise of something bad. Another relatively frequent orientational metaphor in business is aheadbehind, although the ahead part of the metaphor is much more developed than the behind. Share prices, stocks and so on can edge ahead, move ahead, race ahead, forge ahead, surge ahead and sprint ahead (corpus examples). This metaphorical use appears to be part of a systematic metaphor of the type the stock market is a race. More will be said of sports metaphors later. A word must be spent on the particular temporal metaphorical use of ahead: –Looking ahead to economic and political union –difficult trading conditions ahead.

This is the metaphor ahead is the future. But in the following:

metaphor

.

115

–Conditions ahead of the 1992 budget –SG Warburg made positive noises ahead of today’s shareholders’ meeting

the phrase ahead of indicates an event previous to some other event, and so ahead of is the past. This could cause problems for a non-native student of business English. Another common experiential metaphor in the business texts is based on the hard–soft opposition. Taking soft first, we find (either in the corpus or CD-ROM material) soft prices, soft rates, soft credits and soft commissions. Hard is to be found in hard cash, hard transaction and hard bargain. A number of items can be either hard or soft—option, sell, landing and currency. soft prices commissions credits rates

hard cash bargain transactions

soft & hard option sell landing currency

It is not always easy to say precisely what form the metaphors are in. Sometimes it might be soft is weak as in soft currency, or soft is easy as in soft option or soft is less as in soft prices. Sometimes hard is tangible as in hard transactions, at others hard is inflexible as in hard bargain. But on other occasions the metaphorical vehicle of these items appears to be rather indeterminate (as in soft commission), which lends weight to the Lakoff & Johnson and Gibbs claim that metaphor is not secondary to the literal. If it is not possible to translate a metaphor into a literal alternative then the metaphorical meaning must be primary, must be semantically primitive. Business people themselves are notoriously hard-boiled, hard-nosed, hardheaded or hard-line. When they wish to be especially tough they also play hardball, which is rather more gruelling and aggressive than soft-ball—another sporting metaphor, this time from baseball. When their companies are hard-hit, executives become hard-pressed. Returning to the frequency lists, an examination of those items which are relatively uncommon in business journalism reveals a large number of personal pronouns, along with titles like Mr and Mrs and verbs requiring personal subjects like know, try and say. This provides evidence that business texts, relative to other newspaper prose, tend to deal with impersonal trends rather than people’s actions. Among the relatively infrequent items there are also a number of past tenses of common verbs such as was and did. This is probably because business is very much interested in future events. Investigating this hypothesis further, the follow-

116

alan partington

ing vocabulary was found to be relatively very frequent in the business texts: believes, forecast, fear(s), confidence, outlook, doubt(s), expectations, speculation, apprehension(s), estimates, likely From this evidence we might hazard the existence of a systematic metaphor of the sort business is forecasting or guessing, and often enough, business is fear. Other items of below average frequency are war, army and military. Further investigation revealed that, in the CD-ROM and corpus material, items such as war, warfare, battle were less frequent than in other sections. This was in contrast to my original intuition about business writings, and also to the evidence of many of the entries in Buzzwords (we might recall the subtitle—the tough new jargon of modern business). Buzzwords gives a wide variety of military or violence metaphors: war cabinet, frag, take no prisoners, war chest, hired guns, dawn raid, war room, hit squad, hip-shooter, body count, kill a deal (i.e. ‘conclude’), saboteur, killer bee However, in all the authentic material held on CD-ROM, only one of these—war chest (meaning a special fund set up by a company to fight off a take-over bid) was found. There might be two explanations for this disparity. Firstly, Michael Johnson’s book, compiled in the eighties may reflect the more confident and aggressive ethos of business in that period. Secondly, there is always a difference between the language used by the participants of any given professional sector, and the language used about the sector—all studies of sectorial language should keep this distinction in mind. Business people may well talk differently than business commentators and we have already noted the tendency of business journalism to be formal and impersonal. Having said this, however, the conflictual nature of business is probably exaggerated by its practitioners. It is more glamorous to see oneself as a warrior than a salesman, which, when all is said and done, is the principal activity of the business operator. It is interesting to note that agreement, alliances, harmonise, harmonisation, co-operate, co-operation and even helpful are all relatively frequent in business journalism in both the corpus and CD-ROM material. Business people, incidentally, have quite a range of complimentary ways of talking about themselves. A competent financier is a wizard (especially, for the sake of alliteration, if Welsh), anyone who writes a book on business is a guru. A good manager has the V-word (vision) and champions their company. That business is war is unproved on the present evidence. It is most definitely, as has been said, sport, and especially a race. A company which is world class will have an excellent track record and will be found in the fast lane or track (note

metaphor

117

the metaphor fast is good). The word front-runner is more common on the business pages than even the sports. One’s turf is one’s own territory, one’s area of business. Companies emerge as favourite to accomplish some plan: –Allied Lyons is tipped as favourite to buy the brands –BSN, the sprawling French group, has emerged as the new favourite for the Scottish stake (i.e. to buy out the Scottish and Newcastle Breweries Co.).

If, as suggested by these two examples, business is essentially a horse race then there is no surprise in the amount of betting which goes on: –There is something about penny shares that appeals to tipsters –they say Chrysler has essentially ‘‘bet the company’’ on a series of new models –those who recently bet against sterling [. . .] –They are more at home having a flutter than making hard–nosed investments –Gambler Bock raises the stakes at Lonrho.

Other prominent systematic metaphors which emerge from Buzzwords, and for which evidence was found in the corpus and CD-ROM material,include business take-overs as hunting with predators, sharks, shark repellents, stalking horses and vulture capitalists all at work.. Business institutions are seen as buildings with Chinese walls, gatekeepers and glass ceilings. Occasionally the boss carries out some housecleaning. Companies can be seen as many entities. For example, a company is a person with or without deep pockets or teething troubles or a company is a machine with managers working levers, and which might need a kick-start. a company is also an animal which may have a spider organisation or may be a dinosaur. Or more probably it is a plant which the managers grow until it becomes mature.4

7.6 Conclusion: The living and the dead As regards the theory of metaphor, the question remains whether any information has been unearthed to resolve the controversy between the pragmatists’ and the Lakoff & Johnson and Gibbs view of metaphor. From the evidence here, they are both right within the framework of their own arguments, but they each deal with different sorts of metaphor. There is a cline in the originality of metaphorical use, from the unusual, through the well-trodden, to what is usually called the dead metaphor. The Lakoff & Johnson approach, concerned as it is to demonstrate the all-pervasive nature of metaphor

118

alan partington

in language and thought, tends to concentrate exclusively on the well-trodden to dead extremity of the scale—argument is war, ideas are objects, the mind is a machine and so on. If one recalls Gibbs’ argument that hearers take no longer to process a metaphor than any eventual literal equivalent, this would seem predictable if the metaphor in question is one they are already familiar with. However, this study has been concerned with the language of a sector to which the author and most of the readers will be outsiders and in which many of the metaphors are not only unusual but also opaque until explained. At times processing difficulties will almost certainly have been experienced. This would seem to lend weight to Sperber & Wilson’s argument. The Lakoff & Johnson and Gibbs school, on the other hand, was principally making claims about English in general. Nevertheless, perhaps, more attention to sectorial languages would have made them more sensitive to the pragmatists’ ideas about processing difficulties and strategies. A final thought which arises from this study concerns the nature of welltrodden and dead metaphors. It should be remembered that what is a dead metaphor and totally transparent to people working in a sector may be quite opaque to outsiders. How does someone who is not a member of a particular discourse community5 know if a given metaphor is alive or dead for that community? The answer is to examine the way the metaphor collocates. Henderson (1982 and 1986) notes the frequency of what are usually called ‘‘mixed metaphors’’ in economics texts, and I found several examples in business (both from The Economist, August 26th 1995): –And newly liberated South African companies are already compiling shopping lists of potential targets abroad. –inflation [. . .] has fallen recently primarily because of falling food prices. It could easily pick up again.

An object falls downwards, but pick up refers to speed in a forward direction. In these examples neither targets, nor falling nor speed picking up is being used with any metaphorical weight at all. These are not really mixed metaphors, only technical jargon. See also line 9 in the concordance above of escalate (. . .and the present value of the pound look like escalating the slowdown in the economy). Another clue to a dead metaphor is in the conventionality of its collocations. A good example is the health metaphor. Businesses or currencies are often described as healthy, robust or strong, but never as fit or sturdy. Conversely, they are described as weak or ailing, but never as ill or frail or sick or feeble. Only a part of the possible vocabulary set of the vehicle of the metaphor is used to describe the topic. This means that expressions like a strong mark, an ailing business have ceased to be parts of a metaphor and have become fossilised collocations. New

metaphor

119

expressions based on the original metaphor—the pound is ill today—would probably be heard as jokes. Noun–verb or noun–adjective collocation in the phrase may also give evidence of metaphorical intent. An example of the word flow retaining a metaphorical force was found in the international news section of The Independent (my emphasis): –His purge has been West Germany’s much-needed transfusion, as its 26 million new citizens bring with them precious skills. Although East Germany managed to staunch the flow on the eve of its 40th birthday, observers said it was still in danger of bleeding slowly to death.

By contrast the business texts contain no such extended metaphors with flow. The concept of money as a liquid whose movement therefore flows is a completely dead metaphor as can be seen from the way flow and especially cash-flow collocates. It is generated or helped, it can be positive or negative, one can run into cash flow difficulties. It can even be under pressure, which, if it were a liquid would result in greater speed of flow, the opposite of what the writer intends. The truth is, of course, that such metaphors have become genre–specific technical language. They have no figurative content, and to all intents and purposes are no longer metaphors at all. It might be possible to posit a general rule of metaphor, which states that ‘‘a metaphor ceases to be a metaphor when it has no simple literal alternative, or when a metaphor is much more common than its literal alternative in its genre’’.6 This conclusion gives strong support to Gibbs’ contention of the hazy boundary between the literal and the non-literal. However, the evidence of this study suggests that the principal Lakoff & Johnson and Gibbs claim—that not only language but also thought is fundamentally metaphorical in nature—remains unproved. Language, as we have seen is heavily metaphorical, but the abundance of dead and dying metaphor—that is, metaphor which has become literal—could well mean that the cognitive processes behind the linguistic expression rely less on metaphor than might at first appear.

7.7 Further research Researchers in stylistics and authorship studies have always been interested in ways of distinguishing lexical features in two bodies of texts (as, famously, between the works of Shakespeare and those of Bacon). As far as genre studies are concerned, the methodology described here, based on the use of a program like WordSmith Tools, for studying some of the particular

120

alan partington

features of a genre is of general application. Here attention is concentrated on business journalism, but the same approach can be used to study the language of any other genre or sectorial language. In a newspaper, other sections could be investigated—the letters page, sports, features, reviews. Here we focus on metaphor, but very similar methodology could be employed for many other kinds of research into idiom or phraseology, and even into particular grammatical features of a genre (as here a relative paucity of common past tenses such as was and did was found to be a feature of the business journalism examined).

Notes 1. For definitions of metaphor see the Introduction to Bollettieri-Bosinelli (1988) and Malmkjaer (1991: 308–12). 2. An overview of the different schools of thought on the existence or otherwise of the literal vs. figurative divide is to be found in Ponterotto (1993). 3. At the time, only available in prototype version. Now available over the Internet via anonymous ftp from: ftp.oup.co.uk/pub/outgoing/elt-epd/demos. 4. Henderson (1982) notes the presence of very similar metaphors in economics, in which the economy is a person (usually ailing), or a machine (which can overheat) or a plant (hopefully growing or blossoming). 5. See Swales (1990: 23–9) for the definition of ‘‘discourse community’’. 6. The transformation of metaphor into a literal descriptive item is, of course, a common diachronic process. Few speakers think about the original meaning of, say, stereotype (a template once used in printing) or sniper (one who shoots snipe, a game-bird), but their modern meanings result from the demetaphorisation of an original metaphor.

8 ‘‘Unusuality’’ Play up and play the word game (Corpus).

8.1 Definition and methodology The case studies outlined so far in this book have used corpus technology to highlight patterns and so provide a more general or more accurate description of normal language use. However, there are a number of circumstances in which the norms or regularities of collocation and phraseological patterning are upset. Partington (1995) investigates how, in ‘‘creative’’ language use, authors exploit the framework of habit that collocation imposes on language—i.e. the expectation that collocation will be normal—for literary and for humorous effects. A number of points were made in this article. First, what is known as ‘‘wordplay’’ is, in general, not simply playing with words, but draws its effect from some alteration at the lexico-semantic level of the phrase, is really play on the senses of a whole expression or idiom. Thus: That night in Southern Australia brought its first snuffle of tidings of great horror [. . .]—(A. Burgess: The End of the World News)

depends for its effect on the knowledge of the expression/quotation tidings of great joy. Similarly in the following (from Redfern 1984): Said the circus manager to the human cannonball who wanted to leave: ‘‘You can’t quit! Where will I find another man of your calibre?’’

the joke depends on the idiomatic expression man of my/your/his calibre. Carroll’s celebrated visual pun in Alice about the mouse’s long tale (tail), whose story takes on a curving tail-like shape, is a play not just on the words tale and tail but on the two collocations a long tale and a long tail. Second, much of what is seen as creative use of language is not invented ‘‘out of the blue’’, but is an imaginative reworking of the usual, as in the extract from Burgess above, and the following fragment of poetry:

alan partington

122

And God still sits aloft in the array / That we have wrought him, stone deaf and stone blind.—(Edward Thomas: February Morning)

By applying stone in an unusual collocational context, the poet achieves a particularly strong intensification of blind since the novelty of an intensifier–adjective collocation almost invariably reinforces it. However, the interpretation of the meaning of stone blind depends on the knowledge of that of stone deaf. This reworking of the normal for rhetorical effect can be called ‘‘unusuality’’. Finally, ‘‘unusuality’’ is more likely to be found in some text types than others. Many technical genres, for example, rely on well-worn and recognisable language patterns. Although many speakers in daily conversation attempt to ‘‘colour’’ their speech for this kind of effect, there are probably practical limitations both on the time available to work out ‘‘novel’’ collocations and the on-line processing effort required of listeners to interpret them. On the other hand, poetry, humorous writing and advertising are all sub-genres particularly rich in word- (or phrase-) play, in the exploitation of normal collocation and phraseology, as are newspaper headlines on which the study described in this chapter is based. Approximately 2500 headlines from the CD-ROM of The Independent newspaper (see Introduction)—500 from each of the Home news, International news, Arts, Business and Sports sections of the paper—were downloaded onto hard disk to form a separate headline corpus. With the help of colleagues and recourse to reference books, these were scrutinised for any occurrence of preconstructed word strings in the original or adapted form. Preconstructed word strings include proverbs, quotations, sayings, idioms, expressions and collocations, in other words, any stretch of language which is normally used and recognised as a unit. It was apparent that the body of headlines from the Arts section was the richest in the kinds of effect under analysis, followed by the Sports section, whilst the News sections were the poorest.

8.2 Quotations A substantial number of the headlines in this mini-corpus contained quotations of some description which had been ‘‘manipulated’’ in some way: (1)

(2) (3)

ART/Prints charming: This week, you can visit the Royal Academy and take an Old Master home with you—if you have the odd few thousand pounds to spare. Iain Gale reports THEATRE/Another brick at the Wall: Della Couling reports on the attempts of German playwrights to cross the East–West divide TELEVISION/The naked and the well-read: Giles Smith on Melvyn Bragg, Fay Weldon, fleas and skimpy underwear

‘‘unusuality’’ (4)

(5) (6) (7) (8) (9) (10) (11) (12) (13)

(14) (15) (16)

(17) (18) (19)

(20) (21) (22) (23) (24) (25)

(26)

123

ART/Those that can, those that do: Three art critics, William Feaver of the Observer, Giles Auty of the Spectator, and William Packer of the Financial Times, are exhibiting paintings this month. Iain Gale gave three artists the chance to put the palette knife in DANCE/A fridge too far THEATRE/Songs from an age of innocence TELEVISION/Sex and the high-street banker TELEVISION/Murder of the cathedral POETRY/Bonfire of the sanities: Words, words, words: John Whitworth, a judge of this year’s National Poetry Competition, puts the experience to verse MUSIC/None shall sleep, nor reach the loo RADIO/Sail of the quincentenary FILM/It all glitters, but it’s just not Goldie FOOD & DRINK/Give us this day our daily starch: To be healthy—and slim—fill up with potatoes, bread and pasta. Michael Bateman looks at the benefits of complex carbohydrates FILM/Are the first cuts the deepest? Defiant Saddam looks back in anger and pride The world won’t end with a bang nor with a whimper: Nicholas Schoon argues that predictions of environmental doom are misplaced, but we may face global hardship and uncertainty The US Presidential Elections/Election rivals with Georgia on their mind American Football: Elway does it his way Rugby Union/Pilkington Cup: Russ responds to the pressure: England’s biggest rugby union club are prepared to conquer at The Stoop in Saturday’s semi-final: Steve Bale reports on the responsibilities that go with coaching Leicester, a club with the greatest of expectations Cricket: Gong with the wind for the toiling bowlers: Scyld Berry reflects on injustices in the man-of-the-match business Play up and play the word game: Jamie McKendrick on two rich and racy new collections of verse from Simon Armitage Racing: Lessons of another lost Arc for Britain Rugby Union: A tale of two teams Football: Rovers reaping return on investment Accountancy & Management: For what we are about to receive. Simon Pincombe finds that company administrators and receivers are among those destined to do well this year We’re all going to the coup tomorrow: Scalphunters are on the warpath at London Zoo —a ‘Gang of Six’ wants to sack half the council and make managers redundant

It is interesting to note that, in the material examined, the kind of ‘‘adapted’’ quotation is rather more common than the unadapted or ‘‘canonical’’ sort. Moreover, those that are found in the original form tend to be applied in contexts so different from their original one as to make their use highly ironic. We will see in the

124

alan partington

next section that this is just as true of popular sayings and proverbs. This phenomenon is probably not reserved to the kind of material being examined here. In a personal communication, Ramesh Krishnamurthy, senior lexicographer at the Cobuild research group, remarked that it is extremely difficult to find examples in the Cobuild corpus of native speakers and writers producing the full, unaltered version of a proverb of any length. They tend to either state only a part e.g. Many a slip . . . , Well, silver linings and all that . . .(see also (25)), or substitute part of the phrase or rephrase the saying entirely (see also Moon 1994: 117 and Arnaud and Moon 1993: passim). To test this, I looked through the five-million-word corpus for examples of the use of canonical or proverbial-type similes (as happy as a lark or as poor as a church-mouse and so on). The entire corpus contained but two examples—as brown as a berry and as sharp as a razor, whereas it contained, among others, the following ‘‘non-canonical’’ similes: (27) (28) (29) (30) (31)

On the other hand, I’ve been as happy as a turkey in January, and all because of a story I spotted in a medical magazine. All four racks are crammed with fish as oily as a politician’s smile. Mankind as sickly as a parrot: Douglas Adams leaves his Apple Macs to tell Steve Homer how technology can save the human race. And where names are concerned, let’s face it, a myth is as good as a mile. Brentford, whose form has been as steady as a drunkard’s roll in the past month, returned to promotion form last night.

Note how, in this last example, the simile is used to express the opposite of steady. There were several other examples of this ironic use. The fact that these kinds of phrases are so often modified would suggest that native speakers/authors use quotation and sayings less to express any ‘‘universal truths’’ that lie within them and more to make highly personalised and context– dependent comments. Nor is the illocutionary intention simply to adapt the information content of the original quotation or saying to the present communicative situation. More often the information content itself is drastically altered when the original ‘‘integral’’ quotation or saying is disrupted.1 Even minimal changes to the original have this effect. A good example is found in (8). The items Murder and Cathedral (from the original title of the play Murder in the Cathedral by T.S. Eliot) have both been accorded drastically new senses, and communicate the topic of the TV programme i.e. the neglect and physical decay of churches and cathedrals in Britain. When the quotation or saying is simply abbreviated, on the other hand, it may be that the original information content is being related to the present situation, as in example (60) listed below: Once a Catholic (part of the saying Once a Catholic, always a Catholic). Presumably the fact that they are shortened to a form which is

‘‘unusuality’’

125

just long enough to make them interpretable by the reader/hearer is a good example of the Gricean illocutionary maxim of Manner ‘‘Be brief’’ and perhaps also of Quantity—‘‘Do not make your contribution more informative than is required’’. These facts have important pedagogical consequences. Non-native speakers learning English, if they are unaware of the ways natives treat this kind of phrase, may think they are impressing their interlocutors with their command of the language, but by exceeding Grice’s maxims, they may be boring them. There is a variety of sources of the reformulated quotations found in the headlines listed above. A good number are titles of literary works, some of which are classical, for example, (6) which blends Blake’s Songs of Innocence with Edith Wharton’s The Age of Innocence; (19)—first Goldsmith: She Stoops to Conquer, then Dickens: Great Expectations and (23) Dickens on his own: A Tale of Two Cities. Others are more modern: (3)—Norman Mailer: The Naked and the Dead; (7)—Helen Gurley Brown: Sex and the Single Girl; (8)—T.S. Eliot: Murder in the Cathedral; (9)—Tom Wolfe: Bonfire of the Vanities;2 (15)—Osborne: Look Back in Anger. Reference to poetry is also found: e.g. (16) from Eliot’s The Four Quartets and (21) Newbolt: Play up and play the game from the poem Vitaï Lampada. Others are from popular music: (2), (14), (17), (18), (26). Two are from prayer: (13), (25). Some are film titles: (5), (20), (22), whilst a couple have their origins in television: (11)—Sale of the Century, (24)—The Rover’s Return: the name of the public house in the British TV programme Coronation Street. There is reference to a quotation from G.B. Shaw: (4)—originally He who can does, he who cannot teaches, whilst Shakespeare, perhaps surprisingly, is quoted just once: (12)—All that glitters is not gold. Of the remaining two, one is from the world of fairy-tale: (1)—Prince Charming, the other: (10)—None Shall Sleep is the English translation of the title of ‘‘the world’s favourite aria’’—Nessun Dorma, which became more widely known than ever in Britain when the BBC decided to use it to introduce the 1990 World Cup football matches broadcast from Italy.

8.3 The four mechanisms of change A number of the modifications performed on the quotation phrases in the collection above are minimal. Example (1) Prints Charming is a homophone phrase of the original expression Prince Charming, that is to say, in their spoken forms there would be no difference between the two (at least in some accents of English, notably R.P.): only the written form is altered. In other quotations a single phoneme has been substituted in one of the constituent items in the phrase: (5): A fridge too far; original: A Bridge Too Far, (9): Bonfire of the Sanities; original: Bonfire of the Vanities, (20): Gong with the Wind; original: Gone with the Wind and (26): We’re

126

alan partington

all going to the coup tomorrow; original: We’re all going to the zoo tomorrow. These are, of course, traditional-style puns. These homophone-to-heterograph or single phoneme alterations can be considered as the minimal form of substitution, which is the first of four categories of phraseological alteration. One of the items of the original collocation is replaced by another, but the replacement must not change the phrase so drastically as to make the original unrecognisable to the text receiver. An example of whole word substitution has already been discussed—(8), in which a grammatical word of replaced another grammatical word in. Other grammatical substitutions occur in (2): at for in, (15): -s of singular verb for zero morph, and (17): their for my. Example (18): Elway does it his way is an extreme case of grammatical substitution. Of the five lexical items which make up the original quotation (the Sinatra song—I did it my way), as many as three have been replaced, the only ones which remain unaltered are it and way. This begs the question of how the text receiver is expected to recognise the original. The solution must be along the following lines. Each of the words in the new version is related to the one in the corresponding position in the original—thus I and Elway are both personal phrase subjects, does and did are parts of the same verb and his stands in the same relation to Elway as my to I. There seem to be two important deductions to be made. Firstly, it is clearly very much the phrase pattern, the phraseology, which is being recognised. Secondly, in order to recognise the original phrase, the receiver is expected to think on the level of word lemma rather than word form—particularly in order to make the link between the verbs does and did. The question of phrase recognition will be discussed below. There are also a number of examples of non-grammatical or lexical substitution in the quotation examples above, namely (3), (7), (11), (13) and (23). Comparing the latter two (Give us this day our daily starch and A tale of two teams), where in both cases the final word is substituted for another, the former seems to be somewhat more successful in engaging the reader than the latter. This is because the substituting word in (13)—starch—has a connection with the substituted word bread, they belong to similar semantic fields, and the humorous effect is achieved by the contrast in register between the Lord’s Prayer and the semi-technical item starch. In (23), the substituting item teams seems to have no semantic relation to the original item cities and so the alteration seems pointless, it is an empty pun. In (3), the substitution of well-read for the dead of the original book title may also appear rather empty, but there is a phonic link between the rhyming old and new items. The pun is also justified in the rest of the sub-headline. Indeed, the majority of such word-phrase plays are opaque in sense if no context is apparent. In the case of (11): Sail of the quincentenary (original: Sale of the Century), we are dealing with a double substitution: the first is the replacement of an item with

‘‘unusuality’’

127

a perfect homophone—sail for sale, the second is quincentenary for century. The links between the original and the modified are therefore principally phonic, though there is also a semantic connection between quincentenary and century. The humour, the ‘‘cleverness’’, is achieved by integrating two substitutions in a single phrase and linking the title of a popular TV programme with the five-hundredth anniversary celebrations of Columbus’ discovery of America an effect very similar to the contrast of register noted above in (13): Give us this day our daily starch. The second mechanism of change is to be found in (6), (10) and (21), where an element has been added to the original quotation. This we can call insertion or expansion. The simplest of these examples is the last (21), where the item word has been inserted into the original Play up and play the game. Of course, word game is itself a recognisable collocation. Although it is not a quotation, there is another example of expansion in the sub-headline to (4): Iain Gale gave three artists the chance to put the palette knife in, where the popular metaphor put the knife in, meaning ‘criticise viciously’ is expanded by the insertion of palette to describe what kind of knife (an artist’s knife). It is a particularly effective comic introduction to a light-hearted article in which artists get the opportunity to criticise the critics. In (10): None shall sleep, nor reach the loo, on the other hand, the expansion is at the end of the quoted phrase. It is a comic expansion whose effect derives once more from the contrast in register and topic between the opera title in somewhat archaic English and the genteel slang expression reach the loo. From the sublime to the functional. A special case of expansion is what might be termed ‘‘blending’’, when two word strings are run together—as in (6): Songs from the Age of Innocence (Blake’s Songs of Innocence and Wharton’s The Age of Innocence). The third mechanism—which has already been mentioned above is—that of truncation or abbreviation. The only ‘‘pure’’ example of abbreviation in the quotation phrases above is to be found in (22): Lessons of another lost Arc for Britain. The two words lost Arc recall the film title The Raiders of the Lost Ark, and the journalist is reinterpreting (and relexicalising) them to refer to the French horse race Le prix de l’Arc de Triomphe, which British riders have failed to win again. It is interesting that a two-word collocation is deemed sufficient to recall the original phrase. The fourth and final mechanism is rephrasing or reformulation, which, along with substitution, is the most common.3 Considerable rephrasing of an original quotation has taken place in (12), (14), (16), (19), and (24). When discussing (18): Elway does it his way, it was argued that the phraseology was what the text receiver recognised, but when the rephrasing of a quotation or saying is particularly drastic, this can no longer be the case. What is it in (14): Are the first cuts the deepest? which enables the reader to recognise The first cut is the deepest, or in (19) to conquer at The Stoop which recalls She stoops to conquer? The answer must be that the reader simply

128

alan partington

recognises the collocation of two or three items in the new version—first, cut(s) and deepest in (14) and stoop(s) and conquer in (19). In discussing the Elway example, we hypothesised that the collocation recognition works at the level of the lemma, since the reader was expected to pick up the relationship between does and did. However, in (19), the alteration is more severe. The new Stoop has changed, not morphologically, but is no longer of the same word class. The clue to recognising the original verb form would seem to be entirely graphic or phonological. In fact, readers get no semantic clues to help them recognise the original since items in the new version tend to be used in entirely new senses, as was mentioned above. In the examples here, the Stoop of the headline is no longer the verb meaning ‘to lower/debase oneself’, but the name of Harlequins rugby club’s stadium. Similarly cuts in the modified version of (14) refers to extracts of film. Something similar has happened in (24), where the original collocation Rover’s Return is modified so that the first item is no longer a genitive but a proper name (of a football club—Blackburn Rovers) and the second item has an entirely different meaning (from ‘return’ as in return home to ‘return’ as in return from investment). Once again it is the graphic/phonological form which enables the reader to pick up the original collocation. It is noticeable how the cooccurrence of just two words in the new version are enough to recall the original, even when one or both the words have changed their form: (32) (33)

FILM/It all glitters, but it’s just not Goldie. (orig: All that glitters (or glisters) is not gold) ROCK/The kooky that didn’t crumble. (orig: That’s the way the cookie crumbles)

These are evidence that the brain’s ability to store and recall collocations is extremely powerful. Thus there seem to be two distinct cognitive mechanisms of which hearers avail themselves to link the new version to the original: the recognition of a phrase structure and collocational recall.

8.4 Productive phraseology The phrase in (7): Sex and the high-street banker merits particular attention because it represents a special and interesting case. The original quotation is the title of the book Sex and the Single Girl, but the phraseology ‘‘Sex and’’ + ‘‘the single X’’ has become very productive, spawning all sorts of versions in modern prose. The CD-ROM of The Independent 1992 contained Sex and the single man and Sex and the single Philippina. The CD-ROM of The Times 1993 had Sex and the single superstar.

‘‘unusuality’’

129

A number of other quotations have had a similar destiny. The Iceman Cometh gives rise to The Aceman Cometh and The Taxman Cometh. Chips with Everything is the father of a model ‘‘X with Everything’’. These are all examples of what can be called ‘‘productive quotation phraseology’’. Here is a list of versions of the line Cometh the hour, cometh the man from the CD-ROM of The Independent 1992: (34)

(35)

(36)

(37)

(38)

(39)

(40)

His surprise diminished when he thought back to the Cup game at Goodison. ‘I was sitting near the dug-out when Everton scored their fourth goal and I remember thinking how ill Dalglish looked. Heysel and Hillsborough took a lot out of him. They say ‘Cometh the hour, cometh the man’, and he took on a lot after Hillsborough.’ Cometh the hour, cometh the woman. Lloyds Bank’s threat to elevate the P45 to cult status should it bid for Midland Bank will more than tax the mind of Edwina Hart. Yesterday she became President of Bifu, the finance union, the first woman president in the union’s 74-year history. Never, before this season, had Farnborough progressed beyond the first round of the FA Cup. West Ham in the third? The whole town was buzzing, according to Ted Pearce, their long-serving manager. Not on training night, it wasn’t—nor on Wednesday, judging by the crowd (975) for the 1–1 draw at home to Welling, but cometh the hour, cometh the fan, and 1,500 are expected to trek up the M3 tomorrow. Cometh the hour, cometh David Gower. Having come to earth at Lord’s, where they paid a predictable price for over-extending themselves in attack, England have recalled the man grounded since the tour of Australia. Scyld Berry on Gower’s recall for the third Test. The penalty, safely dispatched by Ari Hjelm, conjured visions of the dismal draw with Tunisia which preceded the 1990 World Cup, but cometh the hour, cometh the £7m man. Platt was guilty of a second bad miss, shooting wide from left to right, but in injury time at the end of the first half he finally looked the part as Britain’s most expensive footballer when he accepted Steven’s through pass before driving the ball in low, from 15 yards. When asked to select the highpoint of his career, Simpson says, with less modesty than usual, that he has probably not yet achieved it. Given his sense of timing and knack for being in the right place, that may not be a bad assumption. Cometh the boardroom coup, cometh the man. IND 03 FEB 92/Cometh the election, cometh the smear: Well, well, ‘Labour’s Moscow links’. Anthony Bevins, once of the Daily Mail and The Sun, gives his account of how a story is spun and damage done.

In (34) we find the canonical quotation—the only example in the whole collection. In (35) man is substituted by its complementary opposite woman, which also (almost) rhymes with the original. Another rhyme is found in (36) where the minimal (single phoneme) change to fan occurs. A different kind of rhyming effect is achieved in (37), where the name of the cricketer Gower rhymes with hour. In (38)

130

alan partington

we have an expansion with the insertion of £7 million before man. Example (39) contains a substitution in the first part of the quotation, whereas in (40) a double substitution has taken place. This last example is of special interest because there is also a drastic change in the sense communicated by the new version. In the others, despite the changes, the expressions still depend on the meaning of ‘in a crisis, a hero (or heroine) steps forward to resolve it’, roughly the same as Joseph P. Kennedy’s ‘‘when the going gets tough, the tough get going’’. But cometh the election, cometh the smear seems to mean ‘here we are again, as soon as an election looms, politicians will start to make groundless accusations against each other’. Election and smear seems to be a not uncommon collocation in newspaper texts. The following altered versions of the title of Attenborough’s famous film A Bridge Too Far were collected by Aston (1996: see also (5)): (41)

(42)

(43)

(44) (45) (46) (47)

No bridge too far SYDNEY (AFP)—A woman who climbed Sydney Harbour Bridge to watch the official opening 60 years ago abseiled down it yesterday at the age of 80. ‘You get nowhere if you’re scared,’ said Dorothy Butler after completing the feat to publicise Senior Citizens’ Week in New South Wales. Ms Butler was 20 when she and her boyfriend slipped past police to climb the bridge on 19 March 1932 to watch the opening. A bridge too heavy A bridge too far: engineers trying to take away the 2,045 tonne Ingst Road Bridge over the M4 ran in trouble yesterday when two of the 328 wheels on a giant transporter sank into mud as it took the strain of the heaviest load ever to be carried on a motorway. The bridge, condemned as unsafe, is being moved one mile down the road to be demolished. A bridge too far-fetched?: Stephen Weeks wants to turn a few carriages on an unused railway viaduct into holiday cottages, but he can’t get the people of Monmouth on board, says Peter Dunn. STEPHEN WEEKS, film director, castle owner and man of vision, has run into a few technical hitches with his latest production, The Bridge on the River Wye. Media/Talk of the trade: An outburst too far Leading article: A dam too far HANGUPS/A joke too far: Gallery owner Bernard Jacobson on why he’s tired of Duchamp City: Bull run in ‘93, or a case of mad cow? . . . with recovery well under way in the US, all is sweetness and light. And a merry Christmas to you too, but can Mr Knight really be taken seriously? He has had some good calls over the past couple of years and last week’s surge in shares, taking many to record highs, seems to auger well. Let’s hope this one doesn’t prove a call too far.

A call in this sense is an economic prediction.

‘‘unusuality’’

131

These are interesting for a number of reasons. The phrase A Bridge Too Far is in itself metaphorical. In Cruse’s terminology (1986), it is also a ‘‘xenonym’’, in the sense that there is a semantic dislocation between the parts of the phrase. In the terms used in this study, it is a creative exploitation of more usual collocations such as a step too far, a mile too far and so on. Some of the uses in the list above also have the metaphorical sense of ‘embarking or having embarked on an overambitious project’, or ‘having overdone, exaggerated in some way and now having to pay the consequences’ (meanings consistent with the original film title). If a word substitutes the bridge of the original, it gives some indication of where or how the exaggeration has taken place. In other examples however, there is no real metaphorical sense in the use of the phrase, the bridge in question being real and concrete. In others there is a punning play between the two—the literal and metaphorical senses. These last two techniques—the use of a metaphorical expression in a literal way, and the punning combination of literal and metaphorical—are the basis of very many word-plays in headlines (see section below Double sense and relexicalisation).

8.5 Proverbs and sayings The following selection of headlines from the corpus contains reworkings of proverbs and sayings, all of which are easily recognisable as such to native speakers: (48) (49) (50) (51) (52) (53) (54) (55) (56) (57) (58) (59)

DOUBLE PLAY/Another catch for the early birds: David Fanning and Stephen Johnson on recordings of Mozart’s piano music and Schubert’s Symphony No 9 OPERA/Hear no evil, see no evil?: Eric Griffiths finds appearances can be deceptive in the Paris premiere of John Adams’s Nixon in China Golf: Faldo makes many a slip ‘twixt cup and lip DANCE/From little acorns: Judith Mackrell reviews Mikhail Baryshnikov and the White Oak Dance Project at Sadler’s Wells THEATRE/An insincere form of flattery: The Guardsman—Theatr Clwyd, Mold TELEVISION/Any airport in a storm Port stormed Political Commentary: Never mind policy, feel the charisma No bones about it: ART/Andrew Graham-Dixon comes face to face with death at the Victoria and Albert’s new exhibition of funereal art MUSIC/The ears have it: Robert Maycock reviews the BBC’s Berg Festival BOOK REVIEW/All Greek to Latin lovers: Games of Venus: An Anthology of Greek and Roman Erotic Verse from Sappho to Ovid ed Peter Bing and Rip Cohen MUSIC/An open or shut case?: Is the piano finished? Bayan Northcott asks why composers go on writing for an overworked instrument

132 (60) (61) (62) (63) (64) (65) (66)

alan partington THEATRE/Once a Catholic—Tricycle, Kilburn, NW6 RADIO/For queen and countryside: Robert Hanks charts the common ground between Postman Pat and the Poet Laureate RECORDS/You could hear a superlative drop: Edward Seckerson and Stephen Johnson compare notes on recordings of Ravel and Stravinsky ROCK/The kooky that didn’t crumble. Football: Scots savour happy returns: Celtic and Hearts lead the way as clubs north of the border make striking progress in Europe. Bunhill: Russian sailors left all at sea for America’s Cup City: Closing a deal is not his Forte

The mechanisms of change are similar to those discussed in the section on quotation headlines. In (59): An open or shut case? there is an example of grammatical substitution. The saying an open and shut case means an uncontroversial question, but the substitution of or for and in the headline has practically turned the sense on its head—the role of the piano (which, of course, is housed in a case) is being indicated as controversial. Lexical substitution is to be found in (55), (57), (61) and (62). The first of these (Never mind policy, feel the charisma) derives from the original Never mind the width, feel the quality, which is reasonably transparent in meaning—‘it’s quality not quantity which counts’. However, the phrase in (55) derives from an already transformed version of this which is Never mind the quality, feel the width, the title of a TV comedy, in which the protagonists were a couple of tailors whose work was not always up to scratch. The headline recalls this comic transformation and introduces an article which comments upon politicians who rely on the abundance of their charm rather than the quality of their ideas. There are two other reworkings of this phrase in the five-million-word corpus, which show that the phraseology is productive: (67)

(68)

The message appears to be that nowadays, in Bath at least, you can more or less build what you like as long as it’s got a pediment, a colonnade and windows with spacing bars. Never mind the quality, feel the neo-Palladianism. you wonder at the paradox of constructing inordinately expensive and sophisticated cars, then handing them to folk patently incapable of driving them properly. Never mind the quality, feel the cheque.

The mechanism of expansion is employed only once in (53): Any airport in a storm, which derives from the proverb any port in a storm meaning ‘when someone is in trouble, they accept help from anyone, even from people they would normally avoid’ (Cobuild 1987). The modified version is the headline to an article which reviews a TV program about common phobias, including fears of storms and of flying. The program itself is entitled Plane scared, yet another pun based on

‘‘unusuality’’

133

homophony. The same proverb is referred to in (54) Port stormed which, this time, is an example of rephrasing. The new meaning is that a port has been captured in an act of war. Abbreviations to sayings are found in (49), (51), and (60). As is frequently the case when a saying is abbreviated, it is the segment that is missing, which is the part expressing the relevant message. In (51) the full proverb is From little acorn–great oak trees grow, meaning that important happenings can sometimes have quite insignificant origins, and the implication is that the dance project is potentially important. The choice of this proverb becomes clear when the name of the project (White Oak Dance Project) is stated, making explicit the semantic connection acorn-oak. In (60) the unreported part of Once a Catholic, always a Catholic is also relevant since the headline introduces the review of a play about life in a Catholic convent school. ‘‘Once a (noun), always a (noun)’’ is another example of productive phraseology. The five-million-word corpus contains the expression Once a housewife, always a housewife. In (48), (52), (54), (58), and (63), the original sayings have been subjected to rephrasing. Example (58) contains another example of blending, in which the saying It’s all Greek to me/him etc, usually used to mean ‘it’s beyond my understanding’ (LDOCE 1987), has been blended with the common collocation Latin lover. The words Greek and Latin commonly co-occur in the language, indeed, they have status of a binomial.4 The use of the phrase Latin lovers in a headline reviewing classical Roman poetry produces a comic effect due to the contrast of register-topic similar to that we have met elsewhere (examples (10), (11), (13)).

8.6 Double sense and relexicalisation It may have been noted that some of the examples given in the two lists of headlines above contain a word string where the form of the original phrasing has not, in fact been altered. These include: (69) (70)

(71) (72)

Bunhill: Russian sailors left all at sea for America’s Cup. Idiom: To be all at sea = to be thoroughly confused. Golf: Faldo makes many a slip ‘twixt cup and lip. Proverb: There’s many a slip ‘twixt cup and lip = not everything goes according to plan. Football: Scots savour happy returns: Celtic and Hearts lead the way as clubs north of the border make striking progress in Europe. City: Closing a deal is not his Forte.

134

alan partington

To which can be added the following from the next section: (73)

(74)

Hoping to enter a Euro lottery? Don’t bet on it: Maria Scott describes how German lotteries are being promoted here in defiance of the law. Idiom: don’t bet on it = don’t be too sure about s.th. DESIGN/Some day you might be sitting on a small fortune: Fashionable young furniture makers of recent years have proved to be worth investing in. Idiom: sitting on a fortune = to be in possession (esp. unawares) of a source of wealth.

These are clearly classic puns. What the writer has done in each of these cases is to imply a second meaning to the phrase—one which is dependent on the context of the article in question. In literary criticism or stylistics, the process of replacing or coupling an idiomatic sense with a concrete one is often called demetaphorisation.5 For linguistic purposes, we might use the term relexicalisation. Most preconstructed phrases have some kind of idiomatic quality. The Cobuild 1987 dictionary’s definition of idiom is as follows: An idiom is a group of words which, when they are used together in a particular combination, have a different meaning from the one they would have if you took the meaning of all the individual words in the group.

The authors of the headlines (69)—(74) have relexicalised them, reinstating ‘‘the meaning of all the individual words in the group’’, thus ‘‘freeing them up’’ for a second interpretation. In the most successful cases, the effect is to make the saying creatively ambiguous i.e. it keeps its old value as a saying, including whatever idiomatic quality it had, but also acquires a new sense in the current context.6 The simplest of the four is (69). The expression (to be) all at sea has a metaphorical sense of ‘to be thoroughly confused’, which the Russian sailors, subject of the headline, no doubt are. But they are also quite literally at sea. The proverb in (70): There’s many a slip ’twixt cup and lip expresses the popular wisdom that making a plan and carrying it out successfully are not always the same thing. And so, one of the meanings of (70) is that Faldo may not find winning the tournament as easy as expected. But the items cup and lip have literal meanings too—the cup is the tournament trophy, and the lip is the edge of the golf hole which the player was having difficulty in getting beyond. Don’t bet in (73) has the double sense of ‘don’t rely on’ and ‘don’t wager your money’, whilst the sitting in (74) may be done (metaphorically) on a source of wealth or (literally) on a piece of furniture. This co-existence of two senses holds in all four headlines above, and this type is generally more common in the corpus. On other occasions, the new sense supplants the original one, this latter being no longer applicable in the new situation, as in (71): Football: Scots savour happy returns, where the original reference to

‘‘unusuality’’

135

birthday greetings is lost in favour of that to return matches in two–legged sporting encounters. In this case we might talk about ‘‘empty word-play’’ as opposed to the ‘‘motivated word-play’’ of the previous examples. Empty word-play may at first sight appear banal but it draws its effect from the sheer contrast between the old sense and the novel context in which the phrase is being applied. Although the expression ‘‘word-play’’ is used here for convenience, we should remember that the effect is achieved by playing with meaning at the phrase level. Closing a deal is not his Forte contains an expression not my/your etc forte, where forte means ‘strong point’, ‘speciality’. Other examples of the use of this item in the five-million-word corpus mentioned above include I guess that interviews are not his forte and humane warmth not being Flannery’s forte. Forte here, however, is also the name of a businessman, who, evidently, is having difficulty closing a deal. Plays on proper names (both empty and motivated) are fairly common in headlines: other examples in the mini-corpus are Rhyme and Reagan (rhyme and reason, usually there’s no rhyme and reason to something) , Pilate errors (pilot errors), Palmer fells Forest (Nottingham Forest F.C.) and Forgot to remember Forget (Guy Forget, the French tennis player). The term demetaphorisation is perhaps misleading since the second sense given to a phrase is not always more literal than the first but is often just as metaphorical: (75) (76)

BETWEEN THE LINES/Hitting the Sack: The legendary roaring girl Moll Cutpurse defends her sex against the slurs of Sir John Falstaff Football: Aerial warfare gives armchair fans a game of six halves: Giles Smith reflects on a busy afternoon of channel-hopping and goal-hunting

In the first of these, hitting the sack, an expression which usually means ‘going to bed’ but here is reinterpreted, equally metaphorically, as hitting (a colloquial term meaning ‘consuming a large quantity of’) the wine, popular in Shakespeare’s and Falstaff’s day, called ‘‘sack’’. In the second, Aerial warfare usually means ‘war in the air’, but here means competition between TV stations, a more metaphorical use. Channel hopping usually refers to the popular practice of crossing the English Channel to bring back cheap drink, but here it means changing TV channels. It is hard to say which sense is the more metaphorical.7

8.7 Expressions and collocations This last section will deal with headlines which contain a play on expressions or collocations which do not have the status of proverb or saying.

136

alan partington

(77) Pretty vacant: Hang-ups/David Coke on why the Mona Lisa fails to bring a smile to his face (78) BETWEEN THE LINES/Hitting the Sack: The legendary roaring girl Moll Cutpurse defends her sex against the slurs of Sir John Falstaff (79) Hoping to enter a Euro lottery? Don’t bet on it: Maria Scott describes how German lotteries are being promoted here in defiance of the law (80) TELEVISION/First class male (81) ART/Art of stone: Andrew Graham-Dixon finds new evidence of Mantegna’s ascetism at the Royal Academy’s latest exhibition (82) ART/Industrial resolution: Edna Lumb has made an artistic career out of engineering. Susan Watts talked to her at her show (83) DESIGN/Some day you might be sitting on a small fortune: Fashionable young furniture makers of recent years have proved to be worth investing in, says Jonathan Glancey (84) Stage set for gay interpretation of Shakespeare (85) Brushes with fiction: Artists and paintings are often the subject of movies, from Van Gogh in Lust for Life to Gauguin in the upcoming The Wolf at the Door. They also feature frequently in fiction. How successfully do they make the transition from gallery to film or book? (86) RECORDS/Old year’s honours: Full list of this year’s winners of the coveted Andy Gill awards (87) When a gamble is fair game for the EC: The bureaucrats are looking at the future of gambling. Are they set to turn the tables on the industry, Elizabeth Heathcote asks? (88) BOOKS/Novel voting system: Literature at the polls: New and retiring candidates choose their favourite (89) THEATRE/Catholic distaste: Sex, drugs, religion—in The Pope And The Witch, Dario Fo returns to subjects he knows well. Sid Smith reports (90) Radio tapes from ‘Desert Island Discs’ cast away (91) Women track athletes gaining ground on men (92) Columbus paraded as an all-American villain (93) Food features large on summit menu (94) Blue blood boils as aristocrats mix talk and cheese (95) Football: Palmer fells Forest (96) Football: Midfield engine fuels Anfield drive: Ray Houghton’s form can give encouragement to Liverpool in their FA Cup semi-final at Highbury tomorrow (97) Rugby Union: Pilkington Cup semi-finals: Quins maul sleeping Tigers while Guscott’s magic ruins Gloucester’s fairy-tale (98) Ice Hockey: Wasps draw Steelers’ sting (99) Rugby League: Pride of Lions transcends all (100) Tennis: Wimbledon ‘92/Grunt and Graf in way of Seles dream: The determination of Monica Seles came over loud and clear as she beat Martina Navratilova yesterday (101) Cricket: Oxford quick to lose the initiative (102) Tennis/Wimbledon ‘92: Sabatini left in dark on brink of victory

‘‘unusuality’’

137

(103) Rugby Union: Wales set for trying time (104) Football: Aerial warfare gives armchair fans a game of six halves: Giles Smith reflects on a busy afternoon of channel-hopping and goal-hunting (105) Rugby League: Gregory hunts Fox in Great Britain chase (106) Football: Limpar a walk-on star

A word needs to be said about the difference between these phrases, labelled as ‘‘expressions’’, and those of the previous sections. Quotations and proverbs are typically quite lengthy and have a canonical form (though this may allow a certain variation—see Further research below). Quotations, in addition, tend to be attributable to a known source. It is more difficult to define a distinction between ‘‘sayings’’ and other kinds of expressions. Sayings too are usually fairly fixed in form, though grammatical variation can occur e.g. I’m/You’re/They’re etc all at sea. They are institutionalised in the sense of being known to most if not all members of particular linguistic community, and have been around in that community for a long time (which differentiates them from slang expressions and catch-phrases). Many sayings can be traced back to an origin in some particular situation, as is the case with The ayes have it (see the discussion above of (57): The ears have it). And sayings are usually rather opaque in their meaning (e.g. all at sea or many happy returns). Nevertheless, it would be impossible to find a hard and fast distinction between some of the phrases I have classified as ‘‘sayings’’ in the above section and those classified as ‘‘expressions’’ in this. Readers may in fact disagree with the contents of these lists. There is, in any case, a cline from one to another with the perfect saying at one end—institutionalised, with a precise situational origin, fixed in form and very opaque—and ‘‘non-saying’’ type expressions—less dependent on a particular situation, perhaps less fixed and more transparent—at the other. Not all the mechanisms of phraseological exploitation are present in this list. There are no examples of abbreviation—perhaps because most of the expressions already consist of only two or three words. More surprisingly, however, is the fact that there are no examples of expansion either, and the only candidate for rephrasing is (90): Radio tapes from ‘Desert Island Discs’ cast away in which the noun castaway has become a phrasal verb cast away. (The reference is to the radio programme Desert Island Discs in which a guest chooses songs they would like to hear if shipwrecked on an island. The article recounts that the BBC has thrown out —cast away—recordings of these). There is, on the other hand, no lack of cases of substitution. We find minimal substitution in (80): First class male is homophonous with first class mail, (81): art (heart) of stone and (82): industrial resolution (revolution). In the latter two cases the pun is justified in the sub-headline. There are no examples of grammatical substitution but plenty of the lexical kind. In (86): Old (New) Year’s

138

alan partington

Honours, (89): Catholic distaste (taste), (92): all-American villain (hero) and probably (101): quick to lose (seize/gain) the initiative we have substitution by opposite. In (104), the sporting cliché a game of two halves has become . . . of six halves, suggesting how the columnist was able to watch three whole matches in the space of a single afternoon. A large proportion of these headlines depend on relexicalisation. In (91), gaining ground usually means catching up in a figurative sense, but when applied as here in the context of athletics, has a second, fully physical sense. Similarly, in (84) the expression set the stage for/the stage is set for is usually used metaphorically, but on this occasion, talking of Shakespeare, the stage is also a real one.

8.8 Plays on frequently co-occurring items Many of the headlines in the list in this section depend for their effect, not on the exploitation of any specific expression, but on the use of words and phrases which commonly co-occur in texts generally, that is, items which belong to some particular lexico-semantic class. A lexico-semantic class may be defined as a set of lexical items which have a high probability of being found together in a text or collection of texts of the same genre. For example, the items table, chair and furniture may be said to belong to the same lexico-semantic class because, in texts which contain one of these items, there is a good chance of finding the others. The word astronaut is not of the same class because there is a low probability of finding it in the same texts. This, of course, is a statistical definition of lexical class, but the co-occurrence of items obviously felt to belong together in many of the headlines above is evidence that the concept of lexico-semantic class also has a psychological reality for language users, because these classes are a reflection of the way our brains organise the world. The sporting headlines are particularly rich in the use of this technique: (96): Midfield engine fuels Anfield drive, (97): Guscott’s magic ruins Gloucester’s fairytale, Quins maul sleeping Tigers, and (98): Wasps draw Steeler’s sting . In the last two examples, the headlines may well be upsetting collocational phrases, respectively tiger mauls . . . (man/tourist etc) and draw a wasp sting. The phenomena of collocation and lexical class membership are closely related. Common collocates are items frequently found very close together in a text (usually within the space of very few words), whereas lexical class members are items often found in the same texts. The two phenomena shade into each other.8 On occasion, journalists create highly complex semantic networks, as in (87) which exploits the lexical class ‘gambling’: When a gamble is fair game for the EC: The bureaucrats are looking at the future of gambling. Are they set to turn the tables

‘‘unusuality’’

139

on the industry? or (88) which uses the class ‘voting’: Novel voting system: Literature at the polls: New and retiring candidates choose their favourite. These function rather like extended metaphors and it is the techniques of collocation and selection from lexico-semantic classes which do the extending.

8.9 Collocation and the brain In the first chapter of the book, it was noted how collocation is not only a textual phenomenon but also a psychological one. The awareness of what is normal collocation is clearly an important part of a native speaker’s communicative competence (see Chapter 1, and also Bastiaensen 1994). This psychological aspect to collocational theory is clearly highly relevant to the analysis of unusuality, since, when it occurs, habitually collocating word strings do not in general actually physically appear in the utterance but are evoked in the mind of the receiver. Even in the case of those examples of relexicalisation where the original string is intact, the usual meaning of the string is either missing or not fully appropriate to the new context. The receiver expects that normality will prevail, that one of the many preconstructed phrases that he/she feels is being called into play will have its usual form and be related to its usual meaning. Unusuality functions by upsetting these expectations. Nattinger & DeCarrico make the following comment: Whatever the description of patterns, it is generally agreed that the sequence of words in phrases with little variation is more predictable than in phrases with a lot. This is an extremely important fact in communication, one that Oller exploits in his ‘‘grammar of expectancy’’ (Oller and Richards 1973; Oller and Strieff 1975), and accounts for much of the way we process language. The degree to which words constrain those around them, and the assurance we have that certain words are going to follow certain others, are the facts we use to make sense of language and to create all sorts of subtle variations and surprises. The cliché, a good time was had by all, is a relatively frozen pattern, yet a bad time was had by all, a glorious time was had by all and a good time was had by none, are all possible variations on this basic pattern and each would have its proper effect. The effect comes about because we expect something else, with varying degrees of certainty [. . .]—(Nattinger & DeCarrico 1992: 34)

Normal communication, then, is heavily dependent on predictability and therefore on preconstructed language. The use of such language makes communication, in de Beaugrande & Dressler’s (1981) terms, more efficient. The employment of unusuality, on the other hand, arrests readers, makes them think, smile and (especially in headlines) want to read on, making communication more effective.

140

alan partington

This ‘‘surprise effect’’ helps to account for most of the examples of unusuality in the material, but not all. It may shed light on the motivation behind An allAmerican villain or A Myth is as Good as Mile, but it does not explain set to conquer at the Stoop or less exalted, less ‘‘literary’’ examples such as It’s carnival time again down in Rio, when Brazilians go nuts. In these cases, the reader’s expectations are not upset, because the original strings (She Stoops to Conquer, Brazil nuts) are not presented explicitly. Instead they work by allusion. The readers, having taken in the new version, are challenged to reconstruct the original. And, of course, they are able to do so by accessing their vast mental lexicon of preconstructed phrases. It is not difficult to understand why this effect is so sought after in headlines. Its perlocutionary intent is obviously to engage or amuse the reader, and often enough, to impress him/her with the journalist’s cleverness. In a still wider view, they help to create a sense of ‘‘collusion’’ between the paper and the readers. What we might call a ‘‘smugness effect’’ is sought after, whereby the readers feel flattered by the paper’s treatment of them as intelligent enough to get the allusions and congenial enough to appreciate the jokes.

8.10 Other possible research methodology The research methodology used in this study differs somewhat from that used in previous chapters in that, in its earlier stages, it required native-speaker or nativelike awareness to spot the headlines containing an expression. There are other methodologies more suited to non-natives, as proved by Lama (1996). An Italian mother tongue speaker at the Faculty for Interpreters and Translators in Forlì, she used concordance technology to analyse the use of proverbs in English newspapers,—within text and headlines, not just headlines as here—partly to test the findings of the current study, partly to discover something about the pragmatic functions of proverbs in modern day English. The study was restricted to proverbs having to do with the semantic field of ‘money’, ‘thrift’. She referred, among other reference works, to The Oxford Dictionary of Proverbs (1970) and The Penguin Dictionary of Proverbs (1983) to prepare a list of proverbs pertaining to this semantic field, including the following: Business before pleasure Where there’s muck there’s brass Buy in the cheapest market and sell in the dearest Let the buyer beware To buy a pig in a poke A moneyless man goes fast through the market

‘‘unusuality’’

141

Why keep a dog and bark yourself? Pay beforehand was never well served Easy come, easy go Bad money drives out good A fool and his money are soon parted In for a penny in for a pound Money has no smell Money doesn’t grow on trees Money makes the world go round He who pays the piper calls the tune By keyword search, the use of these proverbs in the CD-ROMs of The Telegraph (1993), The Times (1993) and The Guardian (1994) was examined. She was faced with the problem that, since many proverb references were reworkings, the keyword used for the search had often altered or disappeared. Nevertheless a good number of examples were collected. Her findings tended to confirm the present study in that a large proportion of proverb uses uncovered were not in their canonical form. However, during the course of the study, it was found that it was at times difficult to decide what the precise canonical form of a proverb was (there appeared to be more than one version of some) as well as their precise extent i.e. what was an integral part and what was optional (Sinclair 1991a: 111; and see discussion in Chapter 1: Types of Collocation). This degree of indeterminacy was not evident from the way information was presented in the proverb dictionaries. Lama also noted that reference to proverbs could have particular functions in the organisation of a text. They could be used cohesively, as in (my italics): –[. . .] as many economists did, in the days when they believed that a touch of inflation made the world go round, so that two touches would make it rotate twice as fast

She found a disproportionate number of references to proverbs appearing in the headline or opening paragraph and also in the closing paragraph of an article. This she feels may be linked to ‘‘the need to gain the favour of the public by appealing to concepts or truths that belong to a community’s shared knowledge’’. Non-native researchers can use this technology for investigations in the field of idiom by defining and limiting, if possible, the semantic area of research and by using reference works as a source of initial data. However, these works should be treated as guidelines rather than as a straitjacket. In fact, one of the interesting aspects of such studies—as in the others contained in this book—is the comparison between what reference works state and what the corpus data says.

142

alan partington

8.11 Conclusion Perhaps the most interesting conclusions to be drawn from the investigations conducted in this chapter are psychological. How is the reader expected to recognise the original quotations, expressions etc when, so often, so much has changed? First of all, we can observe as a general point that the longer the original stretch of language, the more changes can be made. Thus quotations and proverbs can be reworked a great deal and recognition can still be expected. This explains why all the four reworking mechanisms were found in the quotation examples, whereas only substitution was really active in the expressions/collocations examples. Secondly, it seems clear that a large number of recognition strategies are called into play. Some of the alterations are graphological, some phonological, some grammatical, others upset the collocational and phraseological patterns of the original. Many work by combining the above. To recover the original, readers are calling into play quite an array of cognitive abilities which are subsumed into the general human mental ability of pattern recognition. This is also further evidence for the existence of a mental lexicon containing not just discrete lexical items but chunks of language i.e. the very patterns being recovered. As a final example, we might look at the following: (106) Blue blood boils as aristocrats mix talk and cheese

This is a complex headline in which many of the techniques outlined in this chapter are called into play. Blue blood and blood plus boils are clearly common collocations, but so is blue and the final word cheese (the controversy the article discusses surrounds the blue-veined Roquefort cheese). In talk and cheese we find the minimal substitution of talk for chalk, exploiting the proverb as like as chalk and cheese (that is, ‘very different’). And finally the items blue blood and aristocrats belong to the same lexico-semantic class. If, as was said, one of the principal intentions of this kind of word-play is to impress the reader with the author’s ingenuity, in this case it is highly successful.

Notes 1. The process of relexicalisation – see the section in this chapter entitled ‘‘Relexicalisation and double sense’’. 2. Although most newspaper readers would probably associate the quotation with the novel (or its film), Wolfe is in turn quoting the translation of a phrase by Savanarola.

‘‘unusuality’’

143

3. It should perhaps come as no surprise that the four mechanisms for altering a phrase listed here— substitution, abbreviation, expansion and reformulation—are all found in the material. These mechanisms are, after all, the only four basic operations it is possible to carry out on a word string, or indeed on any kind of information string. 4. They appear as a binomial collocation five times in the five-million-word corpus, and 43 times in The Times 1993 CD-ROM. They seem to occur in the order Greek and Latin when the topic is classical history, culture etc. but Latin and Greek when they refer to the languages, especially as taught in school. 5. Bollettieri-Bosinelli (1988) uses this term in her study of the exploitation of metaphor in advertising. 6. Fernando & Flavell (1981: 7) recognise ‘‘the potential ambiguity of all idioms […] arising from the possibility of literal interpretation’’ and their consequent ability to ‘‘function with different meanings in more than one environment’’. 7. Makkai stresses what he calls the ‘‘disinformation potential’’ of idiom, in that there is always the possibility that the hearer/reader ‘‘will decode the idiom in a logical yet sememically erroneous way’’ (1972: 118). What is happening in the examples reported here, however, is that the logical reading is no longer ‘‘erroneous’’ but is expected in order to produce the rhetorical effect. What I am claiming here is that the ‘‘logical’’ (in the sense of ‘reasonable’) second reading is not necessarily entirely literal but can itself be a new idiomatic reading. 8. There are, of course, special cases in which two members of the same lexico-semantic group could never enter into a collocational relationship with each other. This happens when they are in a paradigmatic relation to each other as, for example, the items yours truly and yours sincerely. Nevertheless, there is a high probability of finding both items in a corpus of texts of similar genre—in this case formal letters.

9 General Conclusion

9.1 Some criticisms and limitations of corpus study At the very beginning of this book mention was made of an early criticism of corpus study, that it could never supplant the grammarian’s intuition. Although this particular objection has been proved irrelevant, a number of other criticisms have been moved which need to be addressed. It is argued that the attention currently being paid by linguists to collocation, phraseological patterning and the collocational principle in general is disproportionate and is the result of the technology of the tool i.e. the corpus and concordancer, rather than any intrinsic importance of the phenomenon itself. In other words, collocation is studied because the concordance is so well-suited to highlighting it. The real object of study of grammar, the argument runs, is not the idiosyncrasies in behaviour of individual lexical items, but the broader, higher-level rules which can make it possible to propose useful generalisations about a language in order both to understand it better and to teach it more effectively. We should not lose sight of the grammatical wood for the lexical trees. There is much truth in the idea that collocational studies are the direct result of the availability of a new research tool, but this fact should not embarrass us unduly. The paradox of the observer – that we can only perceive physical reality by means of (and some would say distorted by) the tools we use to observe it – is common to all sciences which rely on data. Think, for example of how the object of study of, say, astronomy is very much defined by the tools available to it – the telescope and radio-telescope. When new tools become available in linguistics, new phenomena will be accessible to study. In the meantime we make the fullest use of what we have. On the other hand, there is no conflict between the study of the behaviour of lexis and higher-level grammar. As Halliday explains ‘‘grammar and vocabulary are not two different things’’, they are simply different ways of regarding the same phenomenon, ‘‘they are the same thing seen by different observers’’ (Halliday 1992b: 63). Grammarians take a ‘‘top-down’’ approach to language, thinking in terms of wider rules but forever in search of ways to refine the details of the information they

general conclusions

145

dhave. The lexical approach, conversely is ‘‘bottom-up’’, in which researchers investigate the behaviour of individual words but are constantly on the look-out for similarities of collocational behaviour among items which would enable them to be grouped together. The first group burrow down into language whilst the second burrow up until eventually they will meet in the middle. A good example of the symbiosis of the two approaches is the way in which Altenberg (1991) was able, by studying the collocational behaviour of individual items to refine the contents of the two sub-groups of amplifying intensifier adverbs proposed in Quirk et al. (1985), showing, for example, how extremely belonged among the booster group rather than the maximisers, with all this entails for its use with gradable adjectives. The study of grammar and lexis are different principally because ‘‘they lend themselves to different techniques of analysis’’ (Halliday 1992b: 63) but essentially they complement each other. Another criticism levelled at corpus research of the type treated in this book is that language is studied divorced from its context of communication. Any information derived from the type of corpus which contains texts from a variety of sources and authors, and the concordances arising from such a corpus, can have little validity since we tend to know nothing about the author of the message of a concordance line and their illocutionary intentions, maybe very little about the intended audience and the circumstances in which the message was produced. Concordance data is as decontextualised as any linguistic information could possible be and therefore cannot count as communication. The argument has philosophical implications. Corpus linguists have tended to justify what they do by claiming that corpora enable us to retain a record of and therefore analyse linguistic performance, to use the Chomskian term, as opposed to competence, which is the private domain of the introspective, usually generativist, grammarian (see W. N. Francis 1982: 8). But this clearly will not do, at least not in these terms. The justification for studying performance is that it enables us to discover how language interplays with the non-linguistic features of the communicative environment, i.e. speaker intention, message planning, distractions, pressures of time and a host of other circumstantial phenomena. But the circumstantial features of the communicative event are precisely what is lost when a corpus is interrogated. Most corpora contain only the linguistic record and other features are either of necessity or even deliberately excluded. The real answer to this objection is that the corpus is neither performance nor competence but supersedes the distinction between the two. Many linguists have never been happy with the duality in any case. Competence was a rather ill-defined Platonic concept—the ideal speaker’s knowledge of the language—with the grave defect that whatever is claimed about ‘‘ideal knowledge’’ is not objectively disprovable. On the other hand, many researchers feel the need to go beyond the pure indi-

146

alan partington

viduality of performance to be able to attempt more generalisable statements about what language users do. In fact information about particular communicative events by itself is of limited, we might even say purely anecdotal value. By and large, we are not methodologically justified in interpreting the significance of a particular linguistic event unless we can compare it with other similar events. The corpus can provide ‘‘background information’’ against which particular events can be seen. A number of examples of this are to be found in this book, perhaps most notably in Chapter 4, where it is shown how particular instances of irony, veiled criticism, or hidden attitudes can only be grasped by an awareness of the semantic prosody that the linguistic items employed usually evince. In terms of the kind of text reference dealt with in Chapter 6, it is only possible to evaluate the particular reference strategies adopted by a particular author in an individual text if we have some information about what the normal, most frequently employed reference strategies are in such text types. In Chapter 7 the point was made that any analysis of which metaphors were typical of any particular text type had to be based on a comparison with the nature and frequency of metaphors in other text types. Corpus study can help provide such information. There are however a number of dangers inherent in corpus-based study which are sometimes underestimated by enthusiastic researchers. The first is the problem of representativity and the facile over-generalisation of findings. It is self-evidently true that whatever discoveries are made about the behaviour of linguistic items in a corpus hold true only for the portion of language contained in that corpus. It is an empirical matter whether the same facts will hold in other corpora let alone in the ‘‘language as a whole’’ which is, in any case, a mythical beast (Biber 1988, Biber et al. 1994). A corpus, no matter how large and varied, is only ever representative of itself, and claims made about the behaviour of linguistic items after studying corpus data should bear this in mind. Moreover, it is important not to overestimate what is possible with corpora – they are neither infallible nor omnipotent. For instance, the fact that a structure or a vocabulary item is not found in a corpus is no proof that it does not exist, simply that it is not frequent in the text types present in that corpus. It is not possible to prove the non-existence of items using corpora, just as, to cite Popper’s famous analogy, the non-existence of black swans cannot be proved no matter how many white or pink or whatever swans are collected. More controversial is the question of what status should be assigned to the so-called hapax legomena, that is, items which appear only once in a corpus, especially if this is a large one. A vocabulary item which occurs just once might be a rare word, a foreign borrowing or a proper name. Or it might be an invention, a neologism. Or it might be an error of typing or transcription. A structure occurring once may be a very rare construction or a grammati-

general conclusions

147

cal mistake (and may well as such be interesting as evidence of the kind of error made in on-line production by speakers). The analyst has to make decisions on the evidence in each individual case. We must also be aware of the strict limits to what it is possible to look for in corpora. Whilst lexical items can be searched automatically, as Halliday points out, grammarians’ data are much less accessible: ‘‘I cannot even today ask the system to retrieve for me all clauses of mental process or marked circumstantial theme or high obligation modality’’ (1992b: 64).1 Things can be worse still in other fields. For example, as yet the present author has not found a way to trawl the corpus for examples of irony (apart for looking for obvious expressions like the irony is/was. . .which in any case seems to produce uninteresting examples), or of hyperbole or litotes or of similar kinds of rhetorical devices. In some cases it is possible to devise heuristic or approximative research methodologies, mixtures of automatic and ‘‘manual’’ searches, such as the one outlined in Chapter 6 in looking for labelling noun phrases where the search concentrated on paragraph initial sentences as the most likely place to find such phrases. But the phrases themselves had to be localised by eye. Finally we must beware of how the linguistic findings emanating for corpus research are applied, particularly in the pedagogical world. As Widdowson points out ‘‘the relationship between descriptive fact and pedagogic prescription cannot be one of determinacy’’ (1990: 36), and the finding that, say, one sense of a word is more frequent than another does not mean ipso facto that it should be given priority in syllabus design or dictionary construction. Widdowson discusses the example of bet which, in the Cobuild corpus is used more frequently as a modal marker in phrases like I bet he’ll turn up tomorrow than as a fully lexicalised verb meaning to lay a wager. However, the first of these senses is logically derived from the second (whereas the second could not be derived from the first) and the former sense is restricted to particular grammatical environments (first person and phrase initial, without a direct object). Thus the fully lexicalised form should be given priority for teaching purposes and, in fact, the Cobuild dictionary lists them in this order. Widdowson suggests the adoption of a criterion he calls valency, defined as ‘‘the potential of an item to generate further learning’’, thus teachers would wish to teach a certain structure or word meaning ‘‘because its acquisition provides a basis for the learner to understand and learn other structures or meanings by extension’’ (1990: 36). Nevertheless, the frequency of a structure or word sense remains an important component of valency. These limitations and caveats notwithstanding, the field of corpus studies continues to expand, as shown in the overview given in the Introduction to the present work of the number and variety of different types of research being conducted using

148

alan partington

corpora. It is now nearly 40 years since the appearance of the Brown corpus and in this time corpus research can be said to have come of age in the sense of having been accepted into the mainstream of linguistic practice.

9.2 The role of this book In the course of this work, I have tried to stress the variety of different language areas which may be looked at through corpus analysis. It will have become clear that different techniques and procedures are needed for different investigations, and the type of procedure tends to be dictated by the nature of the area to be studied. Concordance lines, which reveal collocations and phraseological patterns in the immediate cotext of an item, are often sufficient for lexical analyses, whereas they are only the starting point for text studies, indicating what to look for in wider stretches of text. The comparison of data from different subcorpora is generally necessary for translation studies and genre analysis. The investigation into ‘‘unusuality’’ concentrated on only a part of the newspaper corpus (headlines). More than any other, this study also highlighted how the data collection is simply a ‘‘prestep’’ to analysis in corpus research, for there is no automatic, technological means to pick out occurrences of unusuality. That is the domain of the mind. The book is an attempt to show, as Sinclair (1991a: 137) has said, that ‘‘there is a lot more to learn about the English language than it was possible to imagine a few years ago’’ (here we take ‘‘the English language’’ to mean the set of its varieties). Much of what was hidden from view or taken for granted, either for convenience or by necessity, is now open to investigation. If one aims to design dictionaries or more refined grammars of the language, then, by and large, the bigger and more varied the corpus at one’s disposition, the better. However, many of the large research or commercial corpora are either unavailable, or difficult or expensive to access. Along with other studies using similar resources, this book shows how a reasonably sized corpus, resident on the user’s computer, can allow researchers, teachers and learners to make a contribution to enlarging the boundaries of knowledge of the language for their own sakes, that of the linguistics community and that of the learning process. Finally, as mentioned in the previous section, it should be stressed that the conclusions reached in these case studies are fully valid only for the kind of text types included in the corpus. This, of course, is true for all data observation study. Any corpus of data is only truly representative of itself and not of the entire universe of study. As pointed out in the first chapter, corpus studies in language description are observations of language production to be used to predict how it may behave. This is fully congruous with standard scientific methodology of what Ellis (1985, 1994),

general conclusions

149

following Long (1985), calls the ‘‘research-then-theory’’ type, in which the researcher collects data and looks for patterns of recurring events, since what is repeated is likely to be meaningful. If some patterning is found, one collects further data and looks for the same patterns in that. These patterns are, in fact, the equivalent of inductively achieved hypotheses. If one does not find the same patterns, the hypothesis is disproved. If one does, it is necessary to collect still more data and analyse it for similar patterns. And so on, theoretically ad infinitum. Thus, case studies such as those described here have no definitive ‘‘findings’’, only predictions which can either be tested against data from corpora of similar text types, or, for genre study, contrasted with data from different text types. This book is essentially a set of suggestions for future research.

Notes 1. If a corpus has been tagged then some kinds of grammatical features can be searched automatically. However, the interrogator can only put their questions using the same grammatical system and categories as were employed by the taggers, and only a fairly broad degree of delicacy is catered for, e.g. it might be possible to retrieve modal auxiliaries but not, to repeat Halliday’s example, ‘‘high obligation modality’’.

References Aarts, J. and W. Meijs (eds). 1984. Corpus Linguistics. Amsterdam: Rodopi. Allerton, D. 1987. ‘‘English intensifiers and their idiosyncrasies’’. In R. Steele and T. Threadgold (eds), 15–21. Allwood, J., L. Anderson and Ö. Dahl. 1977. Logic in Linguistics. Cambridge: Cambridge University Press. Aitchison, J. 1997. The Language Web: (The Reith Lectures). Cambridge: Cambridge University Press. Altenberg, B. 1990. ‘‘Spoken English and the dictionary’’. In Svartvik (ed.), 177–92. ——. 1991. ‘‘Amplifier collocations in spoken English’’. In Johansson and Stenstrom (eds), 127–47. Anderson, R. 1977. ‘‘The notion of schemata and the educational enterprise’’. In R. Anderson, R. Spiro and W. Montague (eds), Schooling and the Acquisition of Knowledge. Hillsdale, NJ: Lawrence Erlbaum. Aresca, N. and B. Schmid. 1989. Chi cerca trova: Guida all’uso del dizionario inglese. Firenze: Sansoni. Arnoud, P. and R. Moon. 1993. ‘‘Fréquence et emploi des proverbes anglais et francais’’. In C. Plantin (ed.), Lieux communs, topoi, stéreotypes, clichés. Paris: Kimé, 323–41. Aston, G. 1996. ‘‘Electronic newspaper corpora of newspapers as a language learning resource’’. In E. Siciliani, A. Cecere, V. Intonti and A.-M. Sportelli (eds), Le Trasformazioni del Narrare. Fasano: Schena, 663–87. ——. 1997. ‘‘Involving learners in developing learning methods: exploiting text corpora in self-access.’’. In P. Benson and P. Voller (eds), Autonomy and the Independence in Language Learning. London: Longman. Austin, J. 1962. How to Do Things with Words. Oxford: Oxford University Press. Backhouse, A. 1992. ‘‘Connotation’’. In W. Bright (ed.). The International Encyclopedia of Linguistics. New York: Oxford University Press, 297–8. Backlund, U. 1973. The Collocation of Adverbs of Degree in English. Studia Anglistica Upsaliensia 13. Stockholm: Almqvist & Wiksell. Baker, M. 1992. In Other Words: A Coursebook on Translation. London: Routledge. ——. 1993. ‘‘Corpus linguistics and translation studies’’. In Baker et al. (eds), 233–50. Baker, M. 1995. ‘‘Corpora in translation studies: An overview and some suggestions for future research’’. Target 7. —— and M. McCarthy. 1990. ‘‘Multi-word units and things like that’’. Unpublished research paper. Birmingham: University of Birmingham. ——, G. Francis and E. Tognini-Bonelli (eds). 1993. Text and Technology. In honour of John Sinclair. Amsterdam and Philadelphia: John Benjamins.

152

references

Barlow, M. 1996. ‘‘Corpora for theory and practice’’. International Journal of Corpus Linguistics 1/1: 1–37. —— and S. Kemmer. 1994 ‘‘A schema-based approach to grammatical description’’. In S. Lima, R. Corrigan, and G. Iverson (eds), The Reality of Linguistic Rules. Amsterdam and Philadelphia: John Benjamins, 19–42. Bastiaensen, M. 1994. ‘‘Littérature et titres de presse: le cas italien’’. Meta XXXIX / 1: 229–40. Beaugrande, R. de and W. Dressler. 1981. Introduction to Text Linguistics. London: Longman. Bell, A. 1991. The Language of the News Media. Oxford: Blackwell. Benson, M. 1985. ‘‘Collocations and idioms’’. In R. Ilson (ed.), Dictionaries, Lexicography and Language Learning. Oxford: Pergamon/The British Council, 61–8. Biber, D. 1988. Variation across Speech and Writing. Cambridge: Cambridge University Press. ——. 1989. ‘‘A typology of English texts’’. Linguistics 27: 3–43. ——, E. Finegan and D. Atkinson. 1994. ‘‘ARCHER and its challenges: Compiling and exploring a representative corpus of historical English registers’’. In U. Fries, G. Tottie and P. Schneider (eds) Creating and Using English Language Corpora: Papers from the Fourteenth International Conference on English Language Research on Computerized Corpora, Zürich 1993. Amsterdam: Rodopi, 1–13. ——, S. Conrad and R. Reppen. 1994. ‘‘Corpus-based approaches to issues in applied linguistics’’. Applied Linguistics 15/2: 169–89. Bolinger, D. 1975. Aspects of Language (2nd ed.). New York: Harcourt Brace Jovanovich. ——. 1976. ‘‘Meaning and memory’’. Forum Linguisticum 1/1: 1–14. ——. 1982. Degree Words. The Hague: Mouton. Bollettieri-Bosinelli, R.-M. 1988. Quando parlano le metafore. Bologna: CLUEB. Broughton, G. 1968. Success with English. Teacher’s Handbook. Harmondsworth: Penguin. Brown, G. and G. Yule. 1983. Discourse Analysis. Cambridge: Cambridge University Press. Brown, P. and S. Levinson. 1987. Politeness: Some Universals in Language Use (2nd ed.). Cambridge: Cambridge University Press. Burgess, A. 1983. The End of the World News. Harmondsworth: Penguin. Carpenter, E. 1993. Confusable Words. London: HarperCollins. Carter, R. 1987. Vocabulary. London: Allen & Unwin. —— and M. McCarthy. 1988. Vocabulary and Language Teaching. London: Longman. Chomsky, N. 1965. Aspects of the Theory of Syntax. Cambridge (Ma): MIT Press. Cook, G. 1992. The Discourse of Advertising. London: Routledge. Coulthard, M. 1993. ‘‘On beginning the study of forensic texts: Corpus, concordance, collocation’’. In Hoey (ed.), 86–97. ——. 1994. ‘‘On the use of corpora in the analysis of forensic texts’’. Forensic Linguistics Vol 1/1: 27–43. —— (ed.) 1994. Advances in Written Text Analysis. London: Routledge.

references

153

Cowie, A. 1992. ‘‘Multiword lexical units and communicative language teaching’’. In P. Arnaud and H. Bèjoint (eds), Vocabulary and Applied Linguistics. London: Macmillan, 1–13. Crick, F. 1979. ‘‘Thinking about the brain’’. Scientific American 9: 218–32. Cruse, D. 1986. Lexical Semantics. Cambridge: Cambridge University Press. Ellis, R. 1985. Understanding Second Language Acquisition. Oxford: Oxford University Press. ——. 1994. The Study of Second Language Acquisition. Oxford: Oxford University Press. Fernando, C. and R. Flavell. 1981. On Idiom: Critical Views and Perspectives. Exeter: University of Exeter. Firth, J. 1957. Papers in Linguistics. London: Oxford University Press. Fligelstone, S. 1992. ‘‘Developing a scheme for annotating text to show anaphoric relations’’. In G. Leitner (ed.), New Directions in English Language Corpora. Methodology, results, software developments. Berlin: Mouton de Gruyter, 153–70. Flowerdew, J. 1993. ‘‘Concordancing as a tool in course design’’. System 21: 231–44. Francis, G. 1986. Anaphoric Nouns, Discourse Monograph No. 11, University of Birmingham: English Language Research. ——. 1993. ‘‘A corpus-driven approach to grammar—principles, methods and examples’’. In Baker et al. (eds), 137–56. ——. 1994. ‘‘Labelling discourse: an aspect of nominal-group lexical cohesion’’. In Coulthard (ed.), 83–101. —— and J. Sinclair. 1994. ‘‘‘I bet he drinks Carling Black Label’: A risposte to Owen on corpus grammar’’. Applied Linguistics 15/2: 190–200. Francis, W. N. 1982. ‘‘Problems of assembling and computerizing large corpora’’. In Johansson (ed.), 7–24. Gass, S. and C. Madden (eds). (1985). Input in Second Language Acquisition. Rowley, MA: Newbury House. Gavioli, L. and G. Mansfield (eds). 1990. The PIXI Corpora. Bologna: CLUEB. Gentzler, E. 1993. Contemporary Translation Theories. London: Routledge. Gibbs, R. 1994. The Poetics of the Mind. Cambridge: Cambridge University Press. Givón, T. 1993. English Grammar: A Function-Based Introduction. Amsterdam: Benjamins. Greenbaum, S. 1970. Verb-Intensifier Collocations in English: An Experimental Approach. The Hague: Mouton. Greenbaum, S. 1984. ‘‘Corpus analysis and elicitation tests’’. In Aarts and Meijs (eds), 193–201. Grice, H. 1975. ‘‘Logic and conversation’’. In P. Cole and J. Morgan (eds), Syntax and Semantics 3: Speech Acts. New York: Academic Press, 41–58. ——. 1978. ‘‘Further notes on logic and conversation’’. In P. Cole (ed.), Syntax and Semantics 9: Pragmatics. New York: Academic Press, 113–28. Halliday, M. 1985. An Introduction to Functional Grammar. London: Edward Arnold. ——. 1992a. ‘‘Language theory and translation practice’’. Rivista internazionale di tecnica della traduzione 0: 15–25. Udine: Campanotto. ——. 1992b. ‘‘Language as system and language as instance: the corpus as a theoretical

154

references

construct’’. In J. Svartvik (ed.), Directions in Corpus Linguistics. Berlin: Mouton, 61–77. Halliday, M. 1994. ‘‘The construction of knowledge and value in the grammar of scientific discourse, with reference to Charles Darwin’s The Origin of Species’’. In Coulthard (ed.), 136–56. —— and R. Hasan. 1967. Cohesion in English. London: Longman. —— and R. Hasan. 1989. Language, Context and Text: aspects of language in a socialsemiotic perspective. Oxford: Oxford University Press. Hartmann, R. 1987. ‘‘Dictionaries of English: The user’s perspective’’. In R. Bailey (ed.), Dictionaries of English. Cambridge: Cambridge University Press, 121–35. Harvey, K. and D. Yuill. 1994. ‘‘The COBUILD testing initiative: The introspective, encoding component’’. Unpublished research paper. Cobuild / University of Birmingham. Hasan, R. 1984. ‘‘Coherence and cohesive harmony’’. In J. Flood (ed.), Understanding Reading Comprehension. Newark, Delaware: International Reading Association, 181– 219. Hatim, B. and I. Mason. 1990. Discourse and the Translator. London: Longman. Henderson, W. 1982. ‘‘Metaphor in economics’’. Economics 18/4: 147–57 (Reprinted in M. Coulthard (ed.). 1986. Talking about Text, Discourse Monograph No. 13, University of Birmingham: English Language Research). Heringer, J. 1972. Some Grammatical Correlates of Felicity Conditions and Presuppositions. Working Papers in Linguistics (Ohio State University) 11: 1–110. Higgins, J. 1988. Language, Learners and Computers. London: Longman. ——. 1991. ‘‘Looking for patterns’’. In T. Johns and P. King (eds), Classroom Concordancing. ELR Journal Vol. 4. Birmingham: Birmingham University Press, 63–70. Hockett. C. 1958. A Course in Modern Linguistics. New York: Macmillan. Hockey, S. 1980. A Guide to Computer Applications in the Humanities. London: Duckworth. Hoey, M. 1991. Patterns of Lexis in Text. Oxford: Oxford University Press. ——. (ed.). 1993. Data, Description, Discourse. London: HarperCollins. Holmes, J. 1973–4. ‘‘On matching and making maps: From a translator’s notebook’’. Delta 16/4: 67–82. Hymes, D. 1971. On communicative competence. Philadelphia: University of Pennsylvania Press. Ilson, R. 1991. ‘‘Lexicography’’. In Malmkjaer (ed.), 291–8. Johansson, S. (ed.). 1982. Computer Corpora in English Language Research. Bergen: Norwegian Computing Centre for the Humanities. ——. 1993. ‘‘ ‘Sweetly oblivious’: Some aspects of adverb–adjective combinations in present–day English’’. In Hoey (ed.), 39–49. —— and A. Stenstrom (eds). 1991. English Computer Corpora: Selected Papers and Research Guide. Berlin: Mouton-de Gruyter. Johns, T. 1986. ‘‘The concordancer: A language learner’s research tool’’. System 14/2: 151– 62. ——. 1991a. ‘‘Should you be persuaded: Two examples of data–driven learning’’. In T. Johns and P. King (eds), Classroom Concordancing. ELR Journal Vol. 4. Birmingham: Birmingham University Press, 1–16.

references

155

——. 1991b. ‘‘From printout to handout: Grammar and vocabulary teaching in the context of data–driven learning’’. In T. Johns and P. King (eds), Classroom Concordancing. ELR Journal Vol. 4. Birmingham: Birmingham University Press, 27–46. —— and P. King (eds). 1991. Classroom Concordancing. ELR Journal Vol. 4. Birmingham: Birmingham University Press. Johnson, M. 1991. Business Buzzwords: The tough new jargon of modern business. Oxford: Blackwell. Jordan, G. 1993. Concordances: Research Findings and Learner Processes. MA dissertation: London Institute. Kempson, R. 1977. Semantic Theory. Cambridge: Cambridge University Press. Krashen, S. and R. Scarcella. 1978. ‘‘On routines and patterns in language acquisition and performance’’. Language Learning 28: 283–300. Kytö, M. and M. Rissanen. 1990. ‘‘The Helsinki Corpus of English texts: Diachronic and dialectal’’. Medieval English Studies Newsletter 23: 11–14. Ladefoged, P. 1972. ‘‘Phonetic prerequisites for a distinctive feature theory’’. In A. Valdman (ed.), Papers in Linguistics in Memory of Pierre Delattre. The Hague: Mouton. Lakoff, G. 1972. ‘‘Hedges: a study in meaning criteria and the logic of fuzzy concepts’’. In Papers from the eighth regional meeting of the Chicago Linguistics Society. Chicago, 183–228. —— and J. Johnson. 1980. Metaphors We Live By. Chicago: University of Chicago. Leech, G. 1974. Semantics. Harmondsworth: Penguin. ——. 1991. ‘‘Corpora’’. In Malmkjaer (ed.), 73–80. Long, M. 1985. ‘‘Input and second language acquisition theory’’. In Gass and Madden (eds). Louw, B. 1993. ‘‘Irony in the text or insincerity in the writer?—The diagnostic potential of semantic prosodies’’. In Baker et al. (eds), 157–76. Lyons, J. 1977. Semantics (2 vols). Cambridge: Cambridge University Press. ——. 1981a. Language and Linguistics. Cambridge: Cambridge University Press. ——. 1981b. Language, Meaning and Context. London: Fontana. McCarthy, M. 1991. Discourse Analysis for Language Teachers. Cambridge: Cambridge University Press. McIntosh, A. 1961. ‘‘Patterns and ranges’’. Language 37: 325–37. Ma, K.C. 1993. ‘‘Small-corpora concordancing in ESL teaching and learning’’. Hong Kong Papers in Linguistics and Language Teaching 16: 11–30. Mair, C. 1993. ‘‘A corpus–based study of grammaticalization in present-day English: The case of help’’. Talk given at ESSE/2. Bordeaux: Université de Bordeaux II. Makkai, A. 1972. Idiom Structure in English. The Hague: Mouton. Malmkjaer, K. (ed.). 1991. The Linguistics Dictionary. London: Routledge. ——. 1993. ‘‘Who can make nice a better word than pretty?: Collocation, translation and psycholinguistics’’. In Baker et al. (eds), 177–92. Martin, J. 1992. English Text: System and Structure. Amsterdam: Benjamins. Martin, M. 1984. ‘‘Advanced vocabulary teaching: the problem of synonyms’’. In Modern Language Journal 68: 130–7. Moon, R. 1994. ‘‘The analysis of fixed expressions in text’’. In Coulthard (ed.), 117–35.

156

references

Morley, J. 1998. Truth to tell: Form and function in newspaper headlines. Bologna: CLUEB. Mparutsa, C., A. Love, and A. Morrison. 1991. ‘‘Bringing concord into the ESP classroom’’. In T. Johns and P. King (eds), Classroom Concordancing. ELR Journal Vol. 4. Birmingham: Birmingham University Press, 115–34. Murison-Bowie, S. 1993. MicroConcord: Manual. Oxford: Oxford University Press. Nakamura, J. 1993. ‘‘A new tool for describing text types’’. In Baker et al. (eds), 293– 312. —— and J. Sinclair. 1995. ‘‘The world of woman in the Bank of English: Internal criteria for the classification of corpora’’. Literary and Linguistic Computing 10.2: 99–110. Nattinger, J. and J. DeCarrico. 1989. ‘‘Lexical phrases, speech acts and teaching conversation’’. AILA Review 6, Vocabulary Acquisition: 118–39. —— and ——. 1992. Lexical Phrases and Language Teaching. Oxford: Oxford University Press. Oller, J. and V. Streiff. 1975. ‘‘Dictation: a test of grammar-based expectancies’’. In R. Jones and B. Spolsky (eds), Testing Language Proficiency. Arlington, Va. Oller, J. and J. Richards (eds). 1973. Focus on the Learner: Pragmatic Perspectives for the Language Teacher. Rowley (Mass): Newbury House. Owen, C. 1993. ‘‘Corpus–based grammar and the Heineken effect: Lexico–grammatical description for language learners’’. Applied Linguistics 14/2: 167–87. Palmer, F. (ed.). 1968. Selected Papers of J.R. Firth 1952–59. London: Longman. Partington, A. 1992. ‘‘Evidence for attitude: What the corpus says’’. In R. Rutelli and A. Johnson (eds), I Linguaggi della Passione: Atti del XIII Congresso dell’Associazione Italiana di Anglistica. Udine: Campanotto, 291–302. ——. 1993. ‘‘Corpus evidence of language change: The case of the intensifier’’. In Baker et al. (eds), 177–92. ——. 1995. ‘‘Kicking the habit: The exploitation of collocation in literature and humour’’. In J. Payne (ed.), Linguistic Approaches to Literature, English Language Research Journal 17. Birmingham: University of Birmingham, 25–44. Pawley, A. and H. Syder. 1983. ‘‘Two puzzles for linguistic theory: Native–like selection and native–like fluency’’. In J. Richards and R. Schmidt (eds), Language and Communication. London: Longman, 191–226. Peters, A. 1983. The Units of Language Acquisition. Cambridge: Cambridge University Press. Ponterotto, D. 1993. ‘‘Figurative language: a cognitive linguistic viewpoint’’. In Castorina G. and V. Villa (eds), La fortuna della retorica: Atti del XII congresso nazionale dell’ associazione italiana di anglistica. Chieti: Métis, 309–20. Quine, W. 1960. Word and Object. Cambridge (MA): The MIT Press. Quirk, R., S. Greenbaum, G. Leech and J. Svartvik. 1985. A Comprehensive Grammar of the English Language. London: Longman. Redfern, W. 1984. Puns. Oxford: Blackwell. Rudska, B., J. Channell, P. Ostyn and T. Putseys. 1982. The Words You Need. London: Macmillan. ——, ——, —— and ——. 1985. More Words You Need. London: Macmillan.

references

157

Rutherford, W. 1987. Second Language Grammar: Learning and Teaching. London: Longman. Sager, J. 1994. Language Engineering and Translation. Amsterdam and Philadelphia: John Benjamins [Benjamins Translation Library 1]. Scott, M. and T. Johns. 1993. Microconcord. Oxford: Oxford University Press. Searle, J. 1979. ‘‘Metaphor’’. In A. Ortony (ed.), Metaphor and Thought. Cambridge: Cambridge University Press, 92–123. Sinclair, J. 1966. ‘‘Beginning the study of lexis’’. In C. Bazell et al. (eds), In Memory of J. R. Firth. London: Longman, 410–30. ——. 1987a. ‘‘Collocation: A progress report’’. In R. Steele and T. Threadgold (eds), 319–31. —— (ed.) 1987b. Looking Up: An Account of the COBUILD Project in Lexical Computing. London: Collins. ——. 1991a. Corpus, Concordance, Collocation. Oxford: Oxford University Press. ——. 1991b. Council of Europe Multilingual Lexicography Project. Report submitted to the Council of Europe under contract no. 57/89. ——. 1992. ‘‘Trust the text: The implications are daunting’’. In M. Davies and L. Ravelli (eds), Advances in Systemic Linguistics: Recent theory and practice. London: Planter, 5–19. Sperber, D. and D. Wilson. 1986. ‘‘Loose talk’’. In Proceedings of the Aristotelian Society 1985/86. Oxford: Blackwell, 153–72. Sperber, D. and D. Wilson. 1995. Relevance: Communication and Cognition (2nd edition). Oxford: Blackwell. Steele, R. and T. Threadgold (eds). 1987. Language Topics: Essays in Honour of Michael Halliday. Amsterdam and Philadelphia: John Benjamins. Stenström, A.-B. 1990. ‘‘Pauses in monologue and dialogue’’. In Svartvik (ed.), 211–52. —— and J. Svartvik. 1993. ‘‘Imparsable speech: Repeats and other nonfluencies in spoken English’’. In N. Oostdijk and P. de Haan (eds), Corpus-Based Research into Language. In honour of Jan Aarts. Amsterdam and Atlanta: Rodopi, 241–4. Stubbs, M. 1995. ‘‘Collocations and semantic profiles: on the cause of the trouble with quantitative methods’’. Functions of Language 2/1: 1–33. ——. 1996. Text and Corpus Linguistics. Oxford: Blackwell. —— and A. Gerbig. 1993. ‘‘Human and inhuman geography: On the computer-assisted analysis of long texts’’. In Hoey (ed.), 64–85. Svartvik, J. (ed.). 1990. The London–Lund Corpus of Spoken English: Description and research. Lund: Lund University Press. —— (ed.). 1992. Directions in Corpus Linguistics. Berlin: Mouton. Swales, J. 1990. Genre Analysis. Cambridge: Cambridge University Press. Thomas, E. 1981. The Collected Poems of Edward Thomas. Oxford: Oxford University Press. Tognini-Bonelli, E. 1993a. ‘‘From a reliable source: Uses and functions of the adjective real’’. Dialogue Analysis IV, Papers from the 4th Conference, Basel 1992. Tübingen: Max Niemeyer Verlag, 429–36.

158

references

Tribble, C. and G. Jones. 1990. Concordances in the Classroom. London: Longman. Ulrych, M. [forthcoming]. ‘‘Reconstructing context and translation. The role of geographical and cultural factors’’. Atti del XVII Convegno dell’Associazione Italiana di Anglistica. Bologna: CLUEB. van Dijk, T. 1981.‘‘Review of R.O. Freedle (ed.) 1979’’. Journal of Linguistics 17: 140–8. Zanettin, F. 1994. ‘‘Parallel words: Designing a bilingual database for translation activities’’. In Corpora in Language Education and Research, Talc94 (Vol 4). Lancaster: University of Lancaster, 99–111. Zgusta, L. 1967.‘‘Multi-word lexical units’’. Word 23: 578–87. Reference works cited: Cambridge International Dictionary of English. (1st edition) 1995. Cambridge: Cambridge University Press. Collins-Cobuild English Language Dictionary. (1st edition) 1987. London: Collins. Collins-Cobuild English Language Dictionary. (2nd edition) 1995. London: HarperCollins. Collins-Cobuild English Grammar. 1990. London: HaperCollins. Longman Dictionary of Contemporary English. (2nd edition) 1987. London: Longman. Longman Dictionary of Contemporary English. (3rd edition) 1995. London: Longman. Oxford Advanced Learner’s Dictionary. (4th edition) 1989. Oxford: Oxford University Press. Oxford Advanced Learner’s Dictionary. (5th edition) 1995. Oxford: Oxford University Press. Oxford Dictionary of Proverbs (4th edition) 1970. Oxford: Oxford University Press. Paravia’s (Dizionario Bilingue Italiano–Inglese). 1989. Torino: Paravia. Penguin Dictionary of Proverbs. 1983. London: Penguin Books. Webster’s Unabridged Encyclopedic Dictionary of English. 1989. New York: Portland House.

Index A Bridge Too Far–125, 130–1 a move–92–6, 104 abbreviation see mechanisms of change absence, semantic field of–43, 58, 62 advertising–28, 77, 122, 143 n. 5 absolutely/assolutamente–48, 56, 57–62 ahead-behind metaphor–114–5 allegation–8, 73, 97–100 Altenberg, B.–3, 58, 64, 145 anaphora see reference anaphoric nouns–8, 96, 97 appropriacy–8, 27–8 n. 1, 39 assertion–99–100 Aston, G.–7, 130 attitude of writer/speaker–8, 57, 65ff., 74, 90, 94, 96, 105, 106 authentic examples/data–4, 6, 14, 105 n. 2, 106 n. 6, 110 authorship studies–2, 119 Backhouse, A.–65, 66 Baker, M.–3, 21, 52, 63 Barlow, M.–6, 7, 21, 22 bent on–68, 77 Biber, D.–2, 4, 13, 146 blending–127, 133 binomials–26, 28 n. 3 Bolinger, D.–19, 21, 24 bones–10–11, 131 build up–77 business texts–111–2, 115, 116, 119 canonical form pp. (see also non-canonical form)–25, 123–4, 129, 137, 141 Carroll, L.–16, 121 Carter, R.–17, 18–9, 26, 29, 72, 105 cataphora see reference causation, semantic field of–5, 36, 42, 46 cause–37, 68 CD-ROMs Il Sole 24 Ore–13 The Guardian–5, 141 The Independent–5, 12, 13, 69, 70, 72, 75, 111, 119, 122, 128, 129

The Telegraph–5, 13, 69, 70, 111, 141 The Times–5, 13, 109, 111, 128, 141, 143 certainty, semantic field of–58, 59, 62, 139 change, semantic field of–5, 43, 58, 62 Chips with Everything–129 Chomsky, N.–18 claim–82, 95, 97–101 cohesion–89–92, 97 colligation–80 collocation adverb-adjective–11, 56–62, 77 and language acquisition–23–5, 29, 147 definition of–15–17, 25 grammatical–80, 87 noun–adjective–10, 11, 33, 44, 46, 57, 74, 77, 80, 87, 96, 119 noun–verb–36, 53, 70, 77, 78, 119, 128 psychological explanation of–16, 17, 138, 139 collocational overlap–34, 54–5 patterns–9, 10, 16, 32, 68 principle–17, 19, 24, 144 recall–128 restriction–25, 26 collocative meaning see meaning combinatorics–17 cometh the hour, cometh the man–129 commit–66–7, 77, 78 n. 3 communicative competence–18, 65, 139 competence (and performance)–18, 27, 38, 45, 145 completely/completamente–48, 57 Computer-Assisted Language Learning–5 concordance–9, 10–13, 32–4, 39, 46–7, 48, 54, 65, 77, 80 ff., 140, 144, 145, 148 KWIC–9 conjure–108–9 connotation–1, 24, 65 ff. and language learners–29 ff. conversation see spoken language correct/corretto–48, 53–6, 61 co-text–9, 16, 32, 73, 95

160

index

corpus/corpora academic–4, 12, 13, 81, 84, 105, 108 ARCHER–2 Bank of English–4, 14, 78 BNC–4, 14 Brown–2, 4, 14, 148 Cancode–4 Cobuild–4, 5, 14, 33, 67, 68, 70, 77, 91, 124, 147 definition of–1 ff., 5, 32 equivalent–3, 8 Helsinki–2 Italian–13, 48, 53 ff. LOB–4, 14 London-Lund–4, 14 mini-corpora–3 ff., 13, 122, 135 newspaper–13, 25, 75, 81, 84, 89, 92, 96, 101, 109, 111, 148 Coulthard, M.–2, 105 creative ambiguity–134 language use–1, 8, 17, 19, 24, 75, 121, 131 Cruse, D.–30, 31, 32, 131 data-driven learning–6 studies–111 dealings–8, 72–6 deixis–89, 91 demetaphorisation–120, 134–5 diachronic linguistics–2, 120 dictionaries learners’–4, 16, 33, 34, 40, 69, 70, 71, 72, 76, 77, 80 non-learners’–4, 69, 70, 72 discourse community–118, 120 Economist, The–5, 111, 118 Ellis, R.–24, 148 entailment–32 entirely/interamente–48, 57, 61 equivalence see translation, equivalence equivalent corpora see corpus/corpora escalate–112, 113–4, 118 expansion see mechanisms of change factive nouns–98 fairy-tale–108, 109, 125 false friends–8, 48 ff. Firth, J.–15–17, 50, 68 form-meaning pairing–22 Francis, G.–3, 14, 79, 80, 96–8, 100, 101 Francis, W.N.–2, 14, 145

frequency–2, 9, 42, 43, 80, 147 lists–107, 112, 115 fundamentalism (-ist)–74–6, 78 ‘‘general English’’–13 general nouns–8, 87, 90–2, 96, 97, 101, 103–6 general verbs–8, 90, 101, 104–6 Gibbs, R.–110, 115 ff. grammar–3, 11, 16, 47, 79ff., 144, 145, 148 grammar of expectancy–139 grammatical collocation see collocation grammatical word(s)–33, 47, 80, 90, 126 grammaticality–27 n. 1 Grice, H.–110, 125 Halliday, M.–3, 50, 51, 79, 144, 145, 147, 149 Halliday, M. and R. Hasan–90, 91, 92, 96, 100, 105 hapax legomena–146 happen–8, 67, 72, 101, 102, 104–6 hard-soft metaphor–115 Harvey, K. and D. Yuill–29 Hatim, B. and I. Mason–50 health metaphor–118 hedging–87–8 n. 5 Henderson, W.–118, 120 Higgins, J.–1, 5 historical linguistics–2, 120 Hoey, M.–16 humour–127 (see also word play) Hymes, D.–8, 18, 28, 65 hyperbolic adjectives–57, 58 nouns–44–6, 57 hyponymy–32, 40, 46, 90 ICAME–14 idiom–21, 22, 25, 26, 38, 120–2, 133, 134, 141, 143 idiom principle–19–21 if-constructions–8 and the language learner–81 ff. clause combinations–84 for hedging–87–8 n. 5 for rhetorical effects–86 non-conditional–85 ff. verbless–83 impressive/impressionante–77 incite/incitare–78 intensifiers–24–5, 34, 43, 56ff., 64, 77, 145 adjectives–44, 58, 61, 122 adverbs–24, 25, 58, 64, 145 Internet–14, 120

index

161

intuition–14, 18, 69, 72, 116, 144 irony–69, 72, 146, 147 item and environment–15–6, 33, 40, 79–80, 89

money, thrift, semantic field of see proverbs mount–112, 113 Mparutsa, C.–5, 63

Johansson, S.–57, 64 Johns, T. (see also Scott, M. and T. Johns)–5–7 Johnson, M.–111

native-speaker disagreement–18 Nattinger, J. and J. De Carrico–19, 20, 25, 139 natural language(s)–21, 63, 84, 86, 88 necessity, semantic field of–58, 85 newspapers Il Sole 24 Ore–57 The Guardian–141 The Independent–5, 12, 13, 69, 70, 72, 75, 111, 119, 122, 128, 129 The Telegraph–5, 13, 69, 70, 111, 141 The Times–5, 13, 109, 111, 128, 141, 143 nominalisation of processes–104 non-canonical form–25, 26, 124

labelling nouns–90, 98 Lakoff, G. and J. Johnson–8, 88, 107, 109, 110, 112, 115 ff. Leech, G.–2, 15–6, 65, 66, 78 lexical word(s)–33, 57, 126 lexico-grammar–3, 7, 79–80 lexico-semantic class see semantic field lexicography–4, 68, 76, 78 logic–29 ff., 86, 88 n. 6, n. 7 look-alike words–27, 48, 51, 56, 57, 60, 63, 77 Louw, B.–8, 68–9, 72, 76, 77 Lyons, J.–29, 30, 31, 32, 65, 88 magic(ian)–108, 109, 138 Malmkjaer, K.–31, 120 Martin, J.–90, 101, 105 meaning collocative–15–16 connotational–27, 65–6, 68, 78 n. 1 denotational–33, 66 interpersonal–97, 100, 105 n. 2 metaphorical–21, 27, 70, 109–10, 115, 117–9, 131, 134, 135 pragmatic–3, 49, 51, 71, 106, 140 processing of–20, 110, 118, 122 mechanisms of change (see also phrase, preconstructed) abbreviation–127, 133, 137, 143 blending–127, 133 expansion–22, 127, 130, 132, 143 n. 3 rephrasing (reformulation)–124, 133, 137 substitution–39, 126, 127, 130, 132, 137, 138, 142, 143 n. 3 mental lexicon–see phrase(s) metaphor dead–21, 117–19 experiential–115 genre-typical–107–8 orientational–112, 114 role in human cognition–109 ff. systematic–107, 108, 111, 114, 116, 117 theories of–109 ff. vs literal–109, 115, 119 MicroConcord (OUP)–10, 12, 13 mini-corpora–3–5, 13

occur–101, 102–4, 106 Owen, C.–14, 87 Partington, A.–2, 34, 64, 121 pattern(s)/patterning (see also phraseology) and language acquisition–23 ff. analysis–24 imitation–24 memorisation–24 recognition–142 Pawley, A. and H. Syder–19, 21, 22 peddle–70–3, 76, 77 performance (and competence)–18, 20, 27 n. 1, 145, 146 persevere–77, 80 persist–77, 80–1, 87 n. 3 Peters, A.–24 phrase(s) indeterminacy, of–141 mental lexicon of–21, 140, 142 preconstructed–19–21, 23, 122, 134, 139, 140 prefab–19, 20, 24, 28 variability of–22, 26, 141 phraseology (see also pattern/patterning)–1, 36, 120, 126, 127 productive–25, 61, 128 ff. typical–35, 122 poetry–16, 122, 125, 133 possibilty, semantic field of–85, 87 possible–79 pragmatics see meaning predictability–9, 139 probability–9, 16, 73, 106 n. 6 pro-forms–91, 101

162 prosody see semantic prosody proverbs–25, 122, 124, 131–5, 137, 140–2 of money, thrift, etc.–140–1 provide–68, 87 pun(s) see word-play pure–27, 33, 34, 39–44 Quirk, R.–33, 43, 90, 106, 145 quotations–25, 122–130, 132, 137, 142 rather–33 reference anaphoric–8, 92, 94–7, 99–103, 105 n. 4 cataphoric–87 n. 3, 92ff., 102–4, 105 n. 4 exophoric–99, 105 n. 4 register–3, 4, 9, 13, 17, 19, 20, 33, 47 n. 2, 52, 94, 104, 126, 127, 133 relevance theory–110 relexicalisation–131, 133ff, 138, 139, 142 n. 1 rephrasing (reformulation) see mechanisms of change representativity–2, 13, 14–6 rhetorical effect–86, 109, 122, 143 rife–67 routines–23 ff. Sager, J.–49, 50 sanction/sanzione–53 sayings–25, 122, 124, 131 ff. schemas/schemata–21–4 Scott, M.–111 Scott, M. and T. Johns–10 scientific texts–20, 84, 87 n. 4, 104, 105 n. 1 semantic field–48, 55, 58, 62, 68, 75, 99, 109, 140 semantic prosody–66 ff. and corpus linguistics–68, 76 and dictionaries–69–72, 76, 77, 78 n. 3 and insincerity–68, 69 and irony–68, 69, 72 and persuasion–72, 77 and translation studies–77 creation of–76 definition of–67 semi-grammatical words–33–4 set in, 67, 70, 72, 76, 77 Sex and the Single Girl–125, 128 sheer–27, 32–46 similes–124 Sinclair, J.–3, 4, 14 n. 2, 15, 17, 19, 25, 26, 62, 67, 70, 72, 77, 101, 141, 148 smugness effect–140 somewhat–33, 47 n. 2

index speaking loosely–110 speech act theory–31, 88, 110 Sperber, D. and D. Wilson–110, 118 spiral–112–14 spoken language–3, 4, 14, 44, 91 sports texts–13, 17, 56, 108–9, 114, 122, 138 stark–9–10 stone (intensifier)–122 Stubbs, M.–3, 5, 67, 89, 105 style–2, 17 stylistic(s)–17, 56, 119, 134 variation–29, 32, 47 substitution see mechanisms of change synonymy–8 absolute–29, 30–2 and language acquisition, translation–29 ff. cognitive–30, 31 complete–30, 31 definitions of–29 descriptive–30 illocutionary–31 symptomatic–68 syntax–2, 15, 21, 22, 26, 79ff. interface with lexis–89 text–15, 19, 20, 21, 49, 51, 89, 102, 141 definitions of–3, 105 n. 1 text types–2, 4, 13, 17 ff., 49, 62, 96, 122, 146 The Iceman Cometh–129 the old *–96 Thomas, E.–122 translation adequacy–49, 50 and context–50, 51, 54, 55, 62, 63 equivalence–49, 51 and purpose–50 true friends–54, 56, 63 truth conditions–30, 47 n. 1 value–32, 85, 86 unusuality definition of–122–3 psychological explanation of–138, 139–40, 142 up-down metaphor varieties see register, genre, text types regional–2 verb constructions–80, 87 word-play–121, 125–6, 131, 134–5, 138–9, 142 motivated-empty–135 WordSmith Tools–111, 119