Social Stylistics: Syntactic Variation in British Newspapers 9783110851151, 9783110129694


192 68 86MB

English Pages 319 [320] Year 1992

Report DMCA / Copyright

DOWNLOAD PDF FILE

Table of contents :
List of tables
List of figures
Preface
Abbreviations
1. Introduction
2. Variability in sociolinguistics and stylistics
2.1. Introduction
2.2. Three approaches to the concept of style
2.3. The notions “variety” and “style”
2.4. Stylistic variation
2.5. Audience design
2.6. Some recent studies in media language
3. Socio-economic differentiation of the British dailies
3.1. “Qualities” and “populars”
3.2. Readership profiles
3.3. Differences in appearance and coverage
3.4. Conclusion
4. Structure of noun phrases
4.1. Introduction
4.2. Premodification
4.3. Postmodification
4.4. Multiple modifications
5. Noun phrases as style markers
5.1. Introduction
5.2. Noun phrase variables in paradigmatic relationships
5.3. Premodifier variation
5.4. Complexity as style marker
5.5. Conclusions and hypotheses
6. Distribution and complexity
6.1. Introduction
6.2. The density of modifiers
6.3. Multiple modifications and modifier embedding
6.4. Noun phrase complexity in the three newspaper categories
6.5. Noun phrase complexity in the five newspaper sections
6.6. Cluster analysis
7. Premodifications
7.1. Types of premodifiers
7.2. Stylistic significance of premodifier variation
7.3. The individual styles
7.4. Summary
8. Postmodifications
8.1. Types of postmodifiers
8.2. Stylistic significance of postmodifier variation
8.3. The individual styles
8.4. Summary
9. Noun phrase name appositions
9.1. Introduction
9.2. The honorific-title-appositive gradients
9.3. Previous studies
9.4. Linguistic constraints
9.5. Social constraints
9.6. Noun phrase name appositions in the sports section
9.7. The structure of the name in appositions
9.8. Summary
10. Conclusion
Notes
Appendix I Coding scheme
Appendix II Corpus
Appendix III Tables
A.3 Readership profiles
A.6 Noun phrase complexity
A.7 Premodification
A.8 Postmodification
A.9 Apposition formats
References
Index
Recommend Papers

Social Stylistics: Syntactic Variation in British Newspapers
 9783110851151, 9783110129694

  • 0 0 0
  • Like this paper and download? You can publish your own PDF file online for free in a few minutes! Sign Up
File loading please wait...
Citation preview

Social Stylistics

Topics in English Linguistics 6 Editors

Jan Svartvik Herman Wekker

Mouton de Gruyter Berlin · New York

Social Stylistics Syntactic Variation in British Newspapers

Andreas H. Jucker

Mouton de Gruyter Berlin · New York

1992

Mouton de Gruyter (formerly Mouton, The Hague) is a Division of Walter de Gruyter & Co., Berlin.

© Printed on acid-free paper which falls within the guidelines of the ANSI to ensure permanence and durability. Library of Congress Cataloging in Publication Data Jucker, Andreas H. Social stylistics : syntactic variation in British newspapers / Andreas H. Jucker. p. cm. — (Topics in English Linguistics ; 6) Includes bibliographical references and index. ISBN 3-11-012969-8 (alk. paper) 1. British newspapers — Language. 2. Journalism — Great Britain — Language. 3. English language — Great Britain — Style. 4. English language — Great Britain — Syntax. 5. English language — Great Britain — Variation. 6. English language — Social aspects — Great Britain. I. Title. II. Series. PE1474.J8 1992 O72'.014-dc20 91-45888 CIP

Die Deutsche Bibliothek — Cataloging in Publication Data Jucker, Andreas H.:

Social stylistics : syntactic variation in British newspapers / Andreas H. Jucker. — Berlin ; New York : Mouton de Gruyter, 1992 (Topics in English linguistics ; 6) ISBN 3-11-012969-8 NE: GT

© Copyright 1992 by Walter de Gruyter & Co., D-1000 Berlin 30 AH rights reserved, including those of translation into foreign languages. No part of this book may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopy, recording, or any information storage and retrieval system, without permission in writing from the publisher. Typesetting: Satzpunkt, Braunschweig Printing: Arthur Collignon GmbH, Berlin. Binding: Lüderitz & Bauer, Berlin. Printed in Germany.

To Stefan and Fabian 's grandparents

Contents List of tables List of figures Preface Abbreviations 1. Introduction 2. Variability in sociolinguistics and stylistics 2.1. Introduction 2.2. Three approaches to the concept of style 2.2.1. Traditional stylistics 2.2.2. Ethnography of speaking 2.2.3. Correlational sociolinguistics 2.2.4. The sociolinguistic triangle 2.3. The notions "variety" and "style" 2.4. Stylistic variation 2.5. Audience design 2.6. Some recent studies in media language 2.6.1. Crystal and Davy - "Journalese" versus "normal" English 2.6.2. Carter and Ghadessy - Newspaper vocabulary 2.6.3. Ferguson - Sports announcer talk 2.6.4. Wallace - Sports versus news section 2.6.5. Floreano - British newspapers and radio news 2.6.6. Ryden and Bell - Determiner deletion 2.6.7. Verschueren - Metapragmatic metaphors 2.6.8. Lüger - Newspaper text types 2.6.9. Conclusion 3. Socio-economic differentiation of the British dailies 3.1. "Qualities" and "populars" 3.2. Readership profiles 3.3. Differences in appearance and coverage 3.4. Conclusion 4. Structure of noun phrases 4.1. Introduction 4.2. Premodification 4.3. Postmodification 4.3.1. Relative clauses 4.3.2. Appositive clauses 4.3.3. Nonfmite clauses

xi xv xvii xxi 1 11 11 12 12 16 18 19 21 26 28 32 32 34 36 37 39 40 41 43 45 47 47 48 56 58 59 59 60 68 69 70 71

viii

5.

6.

7.

8.

Contents

4.3.4. Postmodification by prepositional phrases 4.3.5. Postposed adjectives 4.3.6. Appositions 4.4. Multiple modifications Noun phrases as style markers 5.1. Introduction 5.2. Noun phrase variables in paradigmatic relationships 5.2.1. Relative clauses 5.2.2. Genitive versus o/-construction 5.3. Premodifier variation 5.4. Complexity as style marker 5.5. Conclusions and hypotheses Distribution and complexity 6.1. Introduction 6.2. The density of modifiers 6.3. Multiple modifications and modifier embedding 6.4. Noun phrase complexity in the three newspaper categories 6.5. Noun phrase complexity in the five newspaper sections 6.6. Cluster analysis Premodifications 7.1. Types of premodifiers 7.1.1. Precentral 7.1.2. Central 7.1.3. Postcentral 7.1.4. Prehead 7.2. Stylistic significance of premodifier variation 7.3. The individual styles 7.3.1. The sports section 7.3.2. The home-news and foreign news sections in the down- and mid-markets 7.3.3. The business section in the down-and mid-markets 7.3.4. The arts section in the up-markets 7.3.5. The business, home news and foreign news sections in the up-markets 7.4. Summary Postmodifications 8.1. Types of postmodifiers 8.1.1. Non-verbal postmodifiers 8.1.2. Non-finite verbal postmodifiers 8.1.3. Finite verbal postmodifiers

75 77 77 80 83 83 83 84 89 91 93 102 107 107 108 110 114 125 132 137 138 138 140 142 143 148 152 154 157 161 164 169 175 177 177 179 183 184

Contents

8.2. Stylistic significance of postmodifier variation 8.3. The individual styles 8.3.1. The sports section 8.3.2. The home-news and foreign news sections in the down- and mid-markets 8.3.3. The business section in the down- and mid-markets 8.3.4. The arts section in the up-markets 8.3.5. The business, home news and foreign news sections in the up-markets 8.4. Summary 9. Noun phrase name appositions 9.1. Introduction 9.2. The honorific-title-appositive gradients 9.3. Previous studies 9.4. Linguistic constraints 9.4.1. The semantic constraints 9.4.2. The syntactic constraints 9.5. Social constraints 9.6. Noun phrase name appositions in the sports section 9.7. The structure of the name in appositions 9.8. Summary 10. Conclusion Notes Appendix I Coding scheme Appendix II Corpus Appendix HI Tables A.3 Readership profiles A.6 Noun phrase complexity A.7 Premodification A.8 Postmodification A.9 Apposition formats References Index

ix

184 186 187 190 194 196 200 203 207 207 213 216 223 223 226 230 239 242 249 251 257 259 261 273 273 274 276 277 278 281 293

List of tables Table 1.1: Table 3.1: Table 3.2: Table 3.3: Table 3.4: Table 3.5: Table 4.1: Table 5.1: Table 5.2: Table 5.3: Table 5.4: Table 5.5: Table 5.6: Table 6.1: Table 6.2: Table 6.3: Table 6.4: Table 6.5: Table 6.6: Table 6.7: Table 6.8: Table 6.9:

43 cells analysed across five sections of eleven British national daily newspapers Circulation of British national daily newspapers Categories of British daily newspapers Social grading according to the National Readership Survey Sex profile of readership, 1987 Newspaper circulations in Britain in 1965 and 1987 Examples of multiple premodification Implicational scale for relative deletion in Middle Scots Distribution of "light" and "heavy" items after Aarts Noun phrase structure and distribution after Quirk et al. Distribution of NP types in engineering English as compared with a "norm" Noun phrase structure and distribution in headlines of The Times and the Daily Mirror Noun phrase structure in the writings of Sir Thomas More (1478-1535) Number of modifiers per 1000 NPs per section and newspaper Percentage of modifiers that are premodifiers Average number of noun phrases with one, two, three or four and more concatenated modifiers Average number of noun phrases embedded within modifiers Noun phrase structure and distribution in the British daily newspapers Noun phrase structure and distribution in the three down-market papers Noun phrase structure and distribution in the three mid-market papers Noun phrase structure and distribution in the five up-market papers Noun phrase structure and distribution in two sections of the down-market papers

4 47 48 50 55 58 66 87 93 95 97 99 101 108 110 111 113 115 118 120 122 126

xii

List of tables

Table 6.10: Noun phrase structure and distribution in the four sections of the mid-market papers Table 6.11: Noun phrase structure and distribution in the five sections of the up-market papers Table 6.12: Five styles according to the cluster analysis Table 7.1: Frequency of open class premodifiers in nominal headlines of The Times and the Daily Mirror Table 7.2: Premodifiers in precentral position in per cent of all premodifiers Table 7.3: Premodifiers in central position in per cent of all premodifiers Table 7.4: Premodifiers in postcentral position in per cent of all premodifiers Table 7.5: Premodifiers in prehead position in per cent of all premodifiers Table 7.6: Prehead adjectives in per cent of all preheads Table 7.7: Prehead nouns in per cent of all preheads Table 7.8: Prehead names in per cent of all preheads Table 7.9: Average number of types of premodifiers in six styles Table 8.1: Frequency of different types of postmodification in nominal headlines of The Times and the Daily Mirror Table 8.2: Non-verbal postmodifiers in per cent of postmodifiers Table 8.3: Nominal postmodifiers (adjectives, nouns and names) in per cent of postmodifiers Table 8.4: Prepositional postmodifiers in per cent of postmodifiers Table 8.5: Non-finite verbal postmodifiers (past and present participle phrases and infinitive clauses) in per cent of postmodifiers Table 8.6: Finite verbal postmodifiers (appositional and relative clauses) in per cent of postmodifiers Table 8.7: Average number of types of postmodifiers in six styles Table 9.1: Patterns of noun phrase name appositions Table 9.2: Appositive-title-honorific gradients Table 9.3: Overall distribution of noun-name collocations, in per cent (1971) Table 9.4: Types of noun phrase name appositions in percentage Table 9.5: Average number of main elements in preposed and postposed descriptive appositives Table 9.6: Density of complex appositions

128 129 135 138 139 140 142 143 145 146 147 153 179 180 181 182

183 184 186 211 215 217 226 228 229

List of tables

Table 9.7: Table 9.8: Table A3.1: Table A3.2: Table A6.1: Table A6.2: Table A7.1: Table A8.1: Table A9.1: Table A9.2: Table A9.3: Table A9.4:

Structure of names in appositions, in the home news sections Structure of names in appositions, in the sports sections Social profiles of the British daily newspapers, 1987 Age profiles of the British daily newspapers, 1987 Number of noun phrases with one, two, three or four and more modifiers Number of noun phrases embedded within modifiers Premodifications Postmodifications Noun-name appositions in the home news sections Noun-name appositions in the sports sections Structure of names in noun-name appositions in the home news section Structure of names in noun-name appositions in the sports section

xiii

243 244 273 273 274 275 276 277 279 279 280 280

List of figures Figure 2.1: Figure 2.2: Figure 2.3:

Sociolinguistic triangle Persons and roles in the speech situation Percentage of intervocalic I\J voicing by four newscasters on two New Zealand radio stations Figure 3.1: 1987 class profile of the British national daily newspapers Figure 3.2: Relationship between percentage of ABC1 readership and average circulation Figure 3.3: Net revenue per copy sold in 1986 (pence) Figure 3.4: 1987 age profiles of the British daily newspapers Figure 6.1: Dendrogram of the 43 samples, based on squared Euclidean distance and average linkage between groups Figure 7.1: Ambiguous noun phrase structure Figure 8.1: Schematic representation of a noun phrase with two embedded postmodifiers Figure 8.2: Schematic representation of a noun phrase with three concatenated postmodifiers, one of which is further modified Figure 9.1: Percentage of determiner deletion on 4 British and 4 United States prestige media, 1980/82 Figure 9.2: Percentage of determiner deletion in 7 British daily papers, 1980 Figure 9.3: Determiner deletion on six New Zealand radio stations, 1974-1980 Figure 9.4: Preposed versus postposed descriptive appositives Figure 9.5: Determiner deletion in eleven British dailies Figure 9.6: Frequency of the four noun phrase name apposition patterns in per cent Figure 9.7: Determiner deletion in postposed position Figure 9.8: Noun phrase name appositions in the home news and sports sections of the down-market papers Figure 9.9: Noun phrase name appositions in the home news and sports sections of the mid-market papers Figure 9.10: Noun phrase name appositions in the home news and sports sections of the up-market papers Figure 9.11: Structure of names in appositions, in the home news sections

20 28 31 50 52 53 54 134 174 204

205 216 218 219 220 231 232 233 239 240 241 244

xvi

List of figures

Figure 9.12: Figure 9.13: Figure 9.14: Figure 9.15: Figure 9.16: Figure 9.17:

Structure of names in appositions, in the sports sections Down-market papers, home news vs sports section Mid-market papers, home news vs sports section Up-market papers, home news vs sports section Home news in the three categories of papers Sports section in the three categories of papers

245 246 246 247 248 248

Preface In the twenty-five years or so of its existence, sociolinguistics has achieved quite remarkable results. Institutionally it has established itself as a legitimate and important branch of linguistics, which is reflected in the number of academic posts that are held by sociolinguists and by the still growing number of journals devoted to this field. Academically it has developed an important and fairly coherent methodological paradigm in a Kuhnian sense, in which a lot of important work on specific language situations all over the world is being carried out. If there is a limit to this success story, it is the fact that the most successful applications of the sociolinguistic research paradigm are still largely restricted to the area of phonological - and possibly morphological - variation. This is perhaps not very surprising. It is after all a commonplace that we judge interlocutors according to their accents, that is to say according to the way in which their pronunciation may differ from our own. Many people - and not just linguists - have a remarkable ability to locate interlocutors regionally and socially simply on the basis of their pronunciation. Hence it seems natural that scientific investigations into language variation across regional or social spectra should take phonological differences as their starting point. Moreover, phonological, and to some extent also morphological, variation is particularly suitable for this kind of approach because the pronunciation of a word, for instance, may vary considerably from one speaker to the next without changing the basic meaning of this particular word. And similarly different morphological realisations of a given construction may not affect its meaning even if the different realisations indicate the regional and/or social origin of the speaker. Syntactic phenomena, on the other hand, do not easily lend themselves to this kind of analysis. In what follows, I shall argue that the main reason for the limited success of most of the extant studies in syntactic variation lies in the fact that they rely too much on the methodological tools that have been developed for phonological variation. They usually rely on variables that can be argued to stand in a paradigmatic relationship, as for instance the contrast between who, that and 0 as different realisations of the relative pronoun. Such an analysis tries to set out all the linguistic and nonlinguistic criteria that affect the choice of one variant rather than the other, in the same way that sociolinguistic studies of phonological variation try to correlate systematically different phonological realisations of the same word with their different contexts of use.

xviii

Preface

I shall argue that we should free ourselves from such a restricted view of syntactic variation and consider a broader range of possible variables. Traditional stylistics is an older branch of linguistics, which, like sociolinguistics, tries to correlate the realisations of linguistic elements with their contexts of use. But traditional stylistics has never been concerned with paradigmatic relationships between two or more variants of some linguistic variable. Its main methodological tool is the density of specific features, that is, the number of occurrences of a particular linguistic item, whether phonological, morphological, lexical or syntactic, within a given stretch of text. However, traditional stylistics, unlike sociolinguistics, has never had a coherent and generally accepted methodological framework. Most notably it has tended to be fairly indiscriminate in its categorisation of contexts, whereas sociolinguists have always been meticulous in differentiating the axes of social and regional variation for instance. I shall use the structure and the complexity of the noun phrase as a particularly pertinent example. The noun phrase can vary from a simple personal pronoun to very complex constructions with several modifiers attached to the same head noun and with further modifiers embedded in these modifiers and thus can vary enormously in the informational load it is made to carry. I have chosen the language of all eleven British national daily newspapers as a corpus because they provide data which is fairly coherent - the language is produced on a day-to-day basis and is addressed to a mass audience. Even the paper with the smallest circulation, the Financial Times, sells over a quarter of a million copies every day. At the same time there are clear internal stratifications in that the various papers target very different segments of the newspaper reading public. This study, therefore, tries to achieve several things simultaneously. It provides an account of the structure and the variability of noun phrases in British English newspapers, it assesses the limitations of syntactic variability studies that are essentially based on the methodological tools developed for phonological variation, and it outlines a possible theoretical framework for studies in syntactic variation by combining research tools from sociolinguistics and traditional stylistics. An earlier version of this book was accepted by the University of Zurich as my Habilitationsschrift. I would like to express my gratitude to all the people who have helped to make it possible. In particular, my thanks go to Udo Fries and Peter Matthews, who have given me much advice and encouragement and who have read several draft versions of it. Allan Bell, Silvia Kubier, Tom Lundskaer-Nielsen, Terence Moore, Mats Ryden, and Susan Wright read various parts of earlier drafts and provided very useful comments. Andreas

Preface

xix

Fischer and Helena Raumolin-Brunberg gave me detailed and helpful comments on the entire manuscript. They should of course not be held responsible for any remaining errors and infelicities. I am also grateful to audiences at the Universities of Cambridge, London and Zurich, where I had the chance to present earlier versions of some of the material in this book. Their stimulating discussions have given rise to several improvements. Much of this book was written with the financial help of the Swiss National Science Foundation, which allowed me to pursue my studies from October 1987 to September 1989 as a visiting scholar at the University of Cambridge. I thank the Swiss National Science Foundation for their support, and Peter Matthews, the head of the Linguistics Department in Cambridge, for his hospitality, and all my friends in Cambridge for providing such a congenial atmosphere for my work. And finally, my most deeply felt gratitude goes to my wife, Ursula JuckerKälin, without whose constant support, encouragement and love this book would not have been possible. Zug, July 15th, 1991

Abbreviations

The following abbreviations are used in references and tables for the eleven newspapers of the main corpus. DMi St S DE DMa To DT FT G I T

Daily Mirror The Star The Sun Daily Express Daily Mail Today The Daily Telegraph Financial Times The Guardian The Independent The Times

In the text, the papers are referred to by their full titles. The definite article is included as part of the title, as in the above list, only if it forms an integral part of the masthead of a particular paper. In tables and occasionally in references, the newspaper sections are abbreviated as follows: ar bu fn hn sp

arts business and finance foreign news (overseas news, European news, American news, etc) home news (UK news) sports news

For attested examples, references are given. The absence of a reference can be taken as an indication that the example is a modified version of an attested example or an example invented for illustrative purposes.

1. Introduction That's not style. But it's what gets into newspapers Waterhouse 1981: 111 The concept of style has a wide - not to say confusing - currency in many seemingly disparate linguistic frameworks. It is one of those terms that has an even wider currency outside the confines of linguistics, as for instance in the history of art and in literary criticism, quite apart from the fact that it also exists as an everyday term with a rather imprecise meaning and many fuzzy edges. In linguistics, in spite of the existing differences, it is generally agreed that style is a comparative concept in that it describes some relevant differences between a text or a discourse and some other texts or discourses; or, in some methodological frameworks, with some kind of explicit or implicit norm. It generally applies to instances of real language, language that has been produced by speakers with their beliefs, aims and goals in specific situations, and in particular physical, social and temporal environments. In other words, the concept of style applies to what de Saussure called parole and Chomsky (1965), with a slight shift of meaning, performance. More recently Chomsky (1986) has coined the term Ε-language (i.e. externalised language, as opposed to I-language for internalised language) for instances of actual language. He rejects the study of Ε-language as completely uninteresting. It "appears to have no significance" and it "has no status in linguistic theory" (1986: 31 and 151). However, it is one thing to rely on native speaker intuition about language (and even the most die-hard empiricists depend to some extent on intuition, if only in choosing some data as normal and rejecting other data, such as pathological speech for instance, as less than normal); but it is quite another thing to rely exclusively on intuition as linguistic data without any recourse to actual instances of language, that is to say without checking whether the linguist's intuition about the native speaker's knowledge of his or her language (i.e. about Chomsky's I-language) bears any resemblance to its end product in the form of utterances in a spoken discourse or sentences in a written text. Elanguage, in spite of the seeming entropy, is highly organised, and there is system in the variation. A stylistic investigation is addressed first and foremost to the problem of uncovering some of the regularity underlying the variation and seeming unorderliness of E-language.

2

Introduction

The concept of style is of course not a recent invention in linguistics, but in the last twenty years or so, it has gained a new prominence through its use in a new branch of linguistics, i.e. sociolinguistics. Sociolinguists have had some spectacular successes in the field of phonological variation, but their methodology has not carried over easily to other areas of the grammar. Syntactic variation in particular has repeatedly defied researchers trying to use the new sociolinguistic methodologies. Romaine (1982) gives a very full account of why these sociolinguistic tools are less than optimal outside the confines of phonological variation, for which they were initially developed. One of the main problems is the concept of free variation, which is fundamental to the inventory of sociolinguistic tools. It applies in a very plausible way to phonological variables, the variants of which can be said to differ only in social or indeed stylistic meaning but not in referential meaning, but it is much more difficult, if at all possible, to apply it to syntactic variables. Traditional stylistics, on the other hand, has not been restricted by methodological tools that were developed for one area of the grammar only. Any formal feature which differs in its density from one text or discourse to the corpus of comparison may be used as a style marker irrespective of the meaning of this formal feature. Traditional stylistics, however, often fails to distinguish with sufficient methodological rigour between different types of varieties. Dialects, sociolects and registers, for instance, can all be called styles. The present study arose from a feeling that the two branches of traditional stylistics and sociolinguistics should join forces in order to enhance our understanding of the regularity that exists in the seemingly bewildering variation of syntactic features across a range of different types of texts. As a corpus for this investigation I have made use of extensive extracts from all the national daily newspapers of Great Britain. The limitation of my corpus to British national dailies is intended to ensure a maximum of coherence within the corpus, and it should guarantee comparability of different texts within the corpus. As all the newspapers have a national circulation and all of them are produced in London, no regional differences will have to be reckoned with. Moreover, in spite of the obvious differences between such newspapers as for instance The Sun and The Guardian, all newspapers share a large number of non-linguistic discourse features. The language is transmitted in printed form, and it is public in that it is intended for a very large audience. Furthermore, it seems reasonable to suggest that all newspapers have fairly similar communicative intentions. They aim both to inform and to entertain their audience, even if the different papers assign widely differing priorities to these two intentions. All the papers of my corpus have circulations higher than a quarter of a

Introduction

3

million.1 The printed media enjoy great popularity in Great Britain. Over 30 million adults regularly read one of the national morning newspapers, which have a total circulation of 14.8 million. Beside the national dailies there are some 200 dailies with a locally restricted circulation and some 1,300 weekly newspapers (including business, sporting and religious newspapers). Thus the daily output of language in the print media is enormous and would deserve close linguistic attention for this reason alone. Furthermore, as Wallace (1977: 49) points out: The restricted language of newspapers, journalese, is an excellent subject for empirical research into register variation, because it forms a large convenient corpus, contains several registers, all associated by certain shared features, and is recognized as such by those who use it. Thus we can examine not only the variation in features, but also how the users of this language view what is appropriate to it. It cannot be expected that the language even of one single newspaper is without internal variation. On the contrary, it has often been pointed out that newspapers employ a variety of text types or genres (e.g. Burger 1984: 132; Kniffka 1980: 29-39; Lüger 1983: 18-22; Schmilz 1987: 822). For my corpus I use articles from the following five sections: "home news", "foreign news", "business and finance", "sports" and "arts", which are the most common newspaper sections that appear, at least in the broadsheet papers, on a daily basis. Not all the papers use exactly these names, but the correspondence is always clear. Most papers also carry more specialised sections, such as "science and technology", "media", "motoring", "holiday and travel", etc. However, I have decided to restrict my analysis to the sections that are common to as many papers as possible, in order to ensure the greatest possible comparability. The restriction to these sections furthermore increases the coherence of the analysed language. Lüger (1983: 18), investigating German newspapers, suggests that newspapers have three main intentions, which are best served by articles in different sections. Newspapers want to inform through hard news, features, reports and interviews, they want to express an opinion by means of leaders and commentaries, and they want to entertain with reviews and stories. There is, of course, no hard and fast separation between these intentions, and one single article may serve all three of them, but the sections I have chosen for my analysis, with the exception of the arts section, can be assumed to have a more informative intention than for instance a leader or a gossip column. Moreover these sections, again with the possible exception of the arts section, are more likely to have been written by staff journalists rather than by occasional or regular guest writers who may be granted more linguistic freedom by the subeditors in the feature pages of the various newspapers. Whenever it

4

Introduction

was obvious that one particular article had been written by an outside contributor, it was ignored for my analysis. Table 1.1 gives a summary of all papers and sections that were analysed. For each cell 1000 NPs were collected and analysed. Table l.l. 43 cells analysed across five sections of eleven British national daily newspapers British national newspapers Up-market papers The Daily Telegraph Financial Times The Guardian The Independent The Times Mid-market papers; Daily Express Daily Mail Today Down-market papers; Daily Mirror The Star The Sun

arts

business

χ χ χ

χ χ χ

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

foreign news

home news

sports

X

X

X

X

X X

X

X

X

X

X

X

The down-market papers do not carry regular news on arts, foreign affairs, or business and finance, the only exception being The Sun, which carries a regular "Sun Money" page. The mid-market papers include foreign news as well as business and finance, but do not have a regular arts page. Today usually has only a very small foreign news section, called "WORLD Today", which made it necessary to use all the articles in this section of several issues in order to get a one thousand noun phrase sample. The up-market papers carry the most comprehensive range of sections, with the only exception of the Financial Times, which does not have a sports section. In all instances, I accepted the categorisations made by the papers, even if they are doubtful in several cases. Many of the articles in the Financial Times' home news section, which is called "UK News: General", for instance, deal more or less exclusively with financial matters and would, in all probability, appear in the business and finance section of the other up-market papers. Every paper is represented by between two and five issues bought more or less at random between October 1987 and February 1988. The articles to be

Introduction

5

analysed within each paper were chosen with the help of a table of random digits (to select issue, page and article). The one thousand noun phrase samples from each section should therefore be truly random samples. There is no simple solution to the question of how big a corpus has to be in order to be representative. There is, of course, a point after which an increase in corpus size does not significantly affect the results anymore, but this freezing point is difficult to ascertain, and there can be no generally valid guidelines (cf. Varantola 1984: 57-60, Raumolin-Brunberg 1991: 52-60). Oostdijk (1988: 20) believes to be able to put an exact figure on the size of corpus that is required for syntactic problems in English: Experience with a small subset of English have led us to believe that samples of 20,000 words each are sufficiently large in order to yield reliable information about the frequency of occurrence of most syntactic structures. Such a claim is very difficult to assess, particularly since Oostdijk fails to exemplify the syntactic structures whose frequency could be reliably established by a set of the stipulated size. But again, this depends on the delicacy of analysis. Varantola (1984) investigates the structure of noun phrases in a corpus of engineering English drawn from texts from relevant professional journals. Her corpus is about 20,000 words in size and contains some 2,000 noun phrases. The size of the corpus will, above all, be determined by the frequency of the element under investigation. The inventory of phonological variables, in spite of the possible complexities, is far smaller than the inventory of syntactic constructions. Therefore a syntactic analysis will require a far larger corpus. The ratio between full noun phrases, pronouns and names can confidently be established on the basis of a small corpus, but this is not necessarily true for low frequency constructions, as for instance special types of noun phrase name appositions. Initially I assumed that samples of 1000 noun phrases per cell would be sufficient to account for the frequency of all important types of noun modification. It turned out, however, that one of the most stratifying constructions, noun phrase name appositions, which will be dealt with in detail in chapter 9, do not occur in sufficient numbers within 1000 noun phrases. There are only about 10 to 20 instances in every 1000 noun phrases. This is clearly not enough if several subtypes are to be distinguished, whose relative frequency is the crucial stratifying feature. In this instance more data had to be gathered. As it was not feasible to increase the entire corpus to such an extent that enough noun phrase name appositions would result, between 120 and 150 instances were collected for each

6

Introduction

cell. This does not give any information about the frequency of noun phrase name appositions as such, and even less does it say anything about all those instances that could have occurred but did not. It does, however, provide fairly reliable data on the frequency of the possible subtypes. In all the papers some patterns of noun phrase name appositions were overwhelmingly favoured over some others. This ensures that the general pattern can confidently be established by the number of instances that were collected. However, if relationships between only some of the patterns were considered, it happened that for individual papers too few of these subpatterns were recorded within the collection of individual instances. Newspaper language has often been the subject of linguistic investigations, either using just one newspaper (e.g. Verschueren 1985, Carter 1988, Ghadessy 1988) or comparing two different newspapers usually a broadsheet paper and a tabloid (e.g. Crystal and Davy 1969; Wallace 1977; Märdh 1980). Some studies, however, have tried to adopt a more comprehensive view. Kniffka (1980), for instance, uses a wide range of American newspapers, and Floreano (1986) compares several British newspapers with several British radio stations. The present study, too, adopts a more comprehensive view. It is not restricted to one or two newspapers but covers the entire range of papers within the given geographical limits, and thus it is not restricted to pointing out a range of more or less obvious differences. The differences are correlated in a systematic way with the socio-economic profile of the readership of the individual papers and with the newspaper section in which specific articles appear. Thus it will be possible to distinguish the idiosyncratic differences from the really relevant ones. It is hoped that this approach will also help to bridge the often noted gap between studies dealing with the social function of media products and studies dealing with their linguistic structure (cf. Schmilz 1987:822). One of the most obvious differences between different types of newspapers is the choice of lexical items used. Examples (1) and (2), on the one hand, and (3) and (4), on the other, are extremely unlikely to appear in the same newspaper. (1)

(2)

(3)

Home Secretary Douglas Hurd was jeered by the Tory hang 'em and flog 'em brigade yesterday as he turned down their plea for a referendum on bringing back the rope. (Sun, 8.10.87, 2.2) Just 24 hours after firmly refusing to cough-up any more cash, she sent Health Minister, Tony Newton, to the Commons with an extra £ 100 million hand-out. (Star, 17.12.87, 2.6) It also excluded from debt some A $ 690 m of high yielding preference shares issued during the acquisition of the Fox television stations in the US in 1986. (FT, 22.01.88, 18.3)

Introduction

(4)

Ί

In a £ 600 million anti-pollution programme, the CEGB will also install flue-gas desulphurization (FGD) equipment at the 2,000 megawatt Fiddler's Ferry power station in Cheshire. (T, 8.10.87, 2.4)

These examples appear to be fairly normal and unmarked in the context of the particular newspaper from which they were taken. The enormous difference in tone is to a large extent created by lexical items. The first two examples are very informal in tone, mainly because of the modifier hang 'em and flog 'em and the noun phrase the rope meaning "capital punishment" in (1), as well as the verb phrase to cough-up any more cash and the noun hand-out in (2). The noun phrases high yielding preference shares in (3) ana flue-gas desulphurization equipment in (4) indicate a higher level of formality. The vocabulary in the first two examples is colloquial and informal while it is specialised and technical in the second two examples. However, the present study will not attempt to deal in any systematic way with such lexical differences. The focus of attention will be on variation in the syntactic structure of noun phrases. Noun phrases are very powerful devices. They can be extremely simple, as for instance the subject of (2) above, which consists of the single pronoun she, or they can be extremely complex. One fairly complex example is given in (3). The noun phrase some A $ 690m of high yielding preference shares issued during the acquisition of the Fox television stations in the US in 1986 shows multiple modifications as well as a modifier embedding. But it is by no means an extreme example. To summarise, then, there are three main objectives which this study aims to achieve. First, I want to reassess the analytical tools of sociolinguistics and of traditional stylistics in particular as to their applicability to syntactic variation. On the basis of this reassessment I will point out how the two fields can be made to benefit each other. Secondly, the noun phrase has been recognised as a style marker by several scholars (e.g. Aarts 1971; Varantola 1984; Raumolin-Brunberg 1991). In this study I want to further substantiate this claim by showing to what extent the noun phrase structure varies even within a very narrow range of styles. In order to achieve this, it will be necessary to give a fairly detailed description of the syntax of noun phrases in general. Thirdly, I want to add to our knowledge of the language of newspapers by describing its use of noun phrases as one small area of its syntax. This will show on the one hand that there is not just the much commented on two way distinction between broadsheet papers and tabloids, even though this is the main dividing line, but there are also two quite distinct categories of tabloid papers. It will also show that in many respects the language of the tabloids and the broadsheets does not differ as much as is commonly believed.

8

Introduction

This introduction was inevitably brief and programmatic. The following chapters will develop, expand and justify the claims sketched above. Chapter 2 develops the methodological framework. As a starting point it contrasts correlational sociolinguistics with traditional stylistics and with the ethnography of speaking. The notion of style is discussed in some detail, because it is used differently in sociolinguistics and in stylistics. In sociolinguistics it is usually taken to be a correlate of the amount of attention paid by a speaker to his or her production of speech. It is suggested that it should rather be seen as a correlate of the addressee(s). Thus stylistic differences are caused not so much by different amounts of attention, but by adapting to different audiences. This is contrasted to the notion of style in traditional stylistics, where it is seen as the frequency of occurrence of some linguistic feature in two or more contrasting text samples. Both traditional stylistics and correlational sociolinguistics relate linguistic variables to the extralinguistic context variables. In chapter 3, I shall explore the extralinguistic variables of my corpus by showing how the British daily newspapers can be categorised into three groups according to their targeted audiences. The broadsheet papers, which are traditionally called the "quality papers", are read by a readership of a high socio-economic status (up-market papers). The tabloid papers, which are traditionally called the "popular papers", split up into those that target a mass audience from a fairly low socio-economic level (down-market papers) and those that target an audience between these two extremes (mid-market papers). The differences in the socio-economic profiles of the respective readerships are reflected by numerous other differences, such as the main sources of revenue (advertising for the up-markets, copy sales for the down-markets), the number of sections, or the size of headlines. Chapter 4 outlines the syntactic structure of noun phrases. In this chapter the main constituents of noun phrases are distinguished, i.e. determiners, premodifiers, the head and postmodifiers with their subtypes, and the various subtypes of nouns are contrasted. Chapter 5 reviews the literature on syntactic variation in general and on variation in the syntax of noun phrases in particular. Some of these studies rely heavily on a Labovian framework, which was developed for phonological variables and assumes that there is no discernible semantic difference between alternating realisations of some particular variable (free variation hypothesis), whereas others reject the free variation hypothesis for syntactic variation. Arguments are given why the latter approach is preferable. On the basis of this and the previous chapter, a list of specific hypotheses will be set up that will provide a starting point for the second part of this book, the empirical tests.

Introduction

9

Chapter 6 shows how the complexity of noun phrases can be used as an indicator of style. Down-market papers use a distinctly higher percentage of names and pronouns and the noun modifications that are used tend to be simpler. There is a fairly strong tendency for unmodified nouns to occur in subject position whereas modified nouns occur in non-subject position. This tendency is weaker in the up-market papers. These results are used as a basis for a cluster analysis that groups together samples that are fairly similar within each cluster but sufficiently different across the clusters. Chapters 7 and 8 deal in more detail with various types of premodifier and postmodifier and show how they are used in the different categories of newspapers. Chapter 9 analyses constructions of the type Labour Leader Neil Kinnock versus the Labour Leader, Mr Neil Kinnock or Mr Neil Kinnock, Labour Leader. Noun phrase name appositions are one of the most clearly stratifying features distinguishing the various types of newspapers. The down-market papers and to a slightly lesser degree the mid-market papers strongly prefer a descriptive appositive with zero article to precede the name appositive, whereas the up-market papers generally do not use the zero article in this context and prefer the descriptive appositive to follow the name appositive. Chapter 10, finally, summarises the findings, ties together the different strands of argument, tries to evaluate the success of the theoretical framework, and generally assesses the methodological contribution made by this study to the fields of sociolinguistics and stylistics.

2. Variability in sociolinguistics and stylistics

2.1. Introduction One of the aims of this study is to show that the language used by newspapers shows a linguistic consistency that is greater within one particular section of one particular paper than it is across different sections or different newspapers, that is to say that the language produced by various journalists working for the same section of the same paper is often, in some relevant respects, more alike than that produced by journalists working for different sections or different papers. In short, I want to show some of the stylistic differences and similarities across the British national daily newspapers. Unfortunately, style is a notoriously ill-defined notion. The definitions given to it by researchers in the fields of sociolinguistics, and textlinguistics alone show a great variety, a variety that becomes almost bewildering if definitions in the fields of rhetoric and literary criticism are taken into consideration. However, things are not quite as bad as they seem, and I am going to offer a definition that subsumes many of the sociolinguistic approaches to stylistic variation, albeit at the cost of ignoring approaches in the fields of rhetoric and literary criticism. Furthermore it will allow us to see more clearly in what respects the different sociolinguistic approaches differ, and how they still can be said to be dealing with the same phenomenon. As a first step, I shall survey what could be called traditional stylistics, that is to say an approach to style that is fundamentally based on comparing frequencies of features in a given corpus with an explicit, or sometimes implicit, norm corpus. I shall then go on to demonstrate how such an approach relates to the two mainstream sociolinguistic methodologies of correlational studies on the one hand and of the ethnography of speaking on the other. This background will enable me to elucidate the notions of variety, in general and the notion of style, as one of its hyponyms, in particular. I reject the standard sociolinguistic views of style as a correlate of the amount of attention paid by the speaker to his or her speech, because of insurmountable difficulties in basing such a categorisation on operational criteria. In its place I suggest that the audience in the form of addressees, bystanders or overhearers should be taken as the decisive factor that can be correlated with linguistic differences to establish different styles.

12

Variability in sociolinguistics and stylistics

I shall round off this chapter by reviewing some recent approaches to the language of the media and placing them in the triangle of traditional stylistics, correlational sociolinguistics and the ethnography of speaking.

2.2. Three approaches to the concept of style 2.2.1. Traditional stylistics There is a lay notion of the concept of style, which equates style with the elevated and aesthetically pleasing forms that are used, for instance, by celebrated authors in their writings. Some newspapers, accordingly, are claimed to "lack style" altogether. This is of course not what traditional stylistics takes style to be. Every single text has got a style in as far as it has formal properties that can be compared with those from other texts. A stylistic analysis will try to single out those features that help to distinguish the texts under comparison. One particular feature may occur in only one text and not the other, or it may appear with a frequency that is appreciably different from one text to the other. In Winter's (1969: 3) words: A style may be said to be characterized by a pattern of recurrent selections from the inventory of optional features of a language. Various types of selection can be found: complete exclusion of an optional element, obligatory inclusion of a feature optional elsewhere, varying degrees of inclusion of a specific variant without complete elimination of competing features. The frequency of a feature has to be seen in relation to the length of the text so that we should talk in terms of density, that is to say the frequency of a feature within a well-defined stretch of text (cf. Enkvist 1973: 23). The length of the text may be defined in number of pages of printed text, for instance, or in number of running words, or any other operational measurement for the length of a text. The recognition and analysis of styles are squarely based on comparison. The essence of variation, and thus of style, is difference, and differences cannot be analysed and described without comparison. (Enkvist 1973: 21) The texts that are to be compared have to be sufficiently similar as far as their context of production is concerned in order to allow for a valid comparison. In Enkvist's (1973: 24) terms they have to be contextually related. Thus it probably does not make sense to compare an Old English poem, the Beowulf for instance, with a passage of twentieth century legal prose, but one might con-

Three approaches to the concept of style

13

sider comparing different instances of legal prose over a certain period of time, or one might compare some poems with a number of plays by the same author. The guiding principle will always be to keep as many non-linguistic features as possible constant over all the texts to be compared in order to be able to assign the linguistic differences with more confidence to those few features that do vary (cf. Enkvist 1980 for a categorisation of contexts from a stylistic point of view). The two texts under investigation can have the same status, and style is then seen as the difference between the two, but often one text is assigned the status of a norm against which the style of the other text is seen as a deviation. Enkvist broadens his definition to such an extent as to allow for norm texts that are implicit and exist only in the expectations which readers and analysts bring to bear on the text under investigation. This has the advantage of bridging the gap between linguistic stylistics and literary criticism, because it shows the essential similarities between the literary critic, who attempts to single out the unique features of a text, i.e. those features that deviate from expectations based on previous reading experiences, and the linguist, who compares empirically the features of clearly specified text samples. However, in the following I use the term traditional stylistics to refer more specifically to linguistic stylisticians who work with well-defined norm texts as a basis for comparison (see, for instance, the collection of articles in Dolezel and Bailey 1969). Traditional stylistics correlates language varieties with the context and situation of its production, but it usually does not make any distinctions between regional, social or historical varieties, diglossic levels or genres. All these varieties could be viewed by traditional stylisticians as styles (cf. Enkvist 1973, Crystal and Davy 1969). The failure to discriminate between different types of contexts is felt by many critics to be a serious shortcoming (e.g. Levinson 1988: 162), but this limitation is not inherent in the methodology. The methodology consists mainly in comparing texts and correlating the differences with the contextual differences. All that is needed to distinguish different types of varieties is a stringent categorisation of the contextual features. This problem is in no essential way different from the one encountered by the correlational sociolinguist, who, as I will show below, also has to categorise the contextual features in order to distinguish sociolects, dialects, historiolects and so on. Winter (1961) is an early and pertinent example of this approach. He tries to delimit styles on the basis of the relative frequency of certain syntactic features. The main features which he establishes for every text of his German corpus are the percentage of clause initial elements that are subjects rather

14

Variability in sociolinguistics and stylistics

than objects or adverbials, the percentage of words that are finite verbs (which gives a rough indication of clause length); and the percentage of finite verbs that occur in main clauses (which is a measure for sentence complexity). His large corpus consists of some 30 to 50 literary and scientific texts written within the last two hundred years (he does not use exactly the same corpus for all his tests). He categorises his texts into the four stylistic groups of stage prose, fictional prose, factual prose and scientific prose. The results show that stage prose, which he takes to be a rough approximation to spoken language, has - in relation to the entire corpus - a consistently high percentage of subjects in clause initial position, a high percentage of finite verbs (indicating short clause length); and a consistently high percentage of finite verbs in main clauses (indicating little complexity). In all these cases the opposite is true for the scientific texts, whereas the two categories of fictional and factual prose show a far greater variation. Thus Winter takes his entire corpus as a norm and compares it with its subparts on the basis of the relative frequency of syntactic features. He calls the two styles that he can most confidently set off against the entire corpus primarily spoken and primarily written German (1961: 216). If his results are somewhat less than convincing it is exactly because he fails to categorise his data more rigorously. He completely ignores the historical differences and any possible regional differences, and he includes in his corpus, consisting otherwise entirely of literary and scientific prose, one sample of newspaper texts. A similar approach is taken by Krämsky in various publications (e.g. 1967, 1969, 1972b, 1985), in which he typically correlates the frequency of particular word classes with different styles of written English such as fiction (by Dickens, Wells, Lawrence and Warner), "colloquial" style (plays by Shaw, Wilde, Seymour and Osborne), and scientific style (textbooks ranging from linguistics to biology) (1972b: 36). To date the most sophisticated attempt to differentiate styles on the basis of the frequency of syntactic constructions is the work carried out by Biber in collaboration with Finegan (e.g. Biber 1985, 1986a, 1986b, 1988 and Biber and Finegan 1986, 1988). Biber's (1988) primary aim is to establish the relevant differences between speech and writing, but the implications of his work are far more pervasive than that. With the help of a tagging algorithm he analyses large sections of both the London-Lund corpus of spoken English and the LOB corpus of written English and counts the frequency of 67 syntactic features. Most of these features are attributed specific discourse functions. They include tense and aspect markers, place and time adverbials, various categories of pronouns, nominalisations, subordinations, various relative clause con-

Three approaches to the concept of style

15

structions, downtoners, hedges and discourse particles, type/token ratio, word length, and many others. The guiding principle in the selection of these constructions is firstly whether any discourse function has been claimed for a particular item in the relevant linguistic literature and secondly whether it could be identified with the computer program. The place adverbials (1988: 224), as a case in point, are searched for on the basis of a finite list of common adverbs based on the list given by Quirk et al. (1985: 516). Biber's list, however, does not include adverbial phrases introduced by about, between, in, opposite or on because these prepositions "often mark logical relations in a text". This is of course true, and the omission is systematic across all texts analysed so that it should not distort the results but it intuitively seems to leave out a rather large proportion of all place adverbials. In other cases, as for instance the distinction between past tense forms and past participle forms, Biber is forced to post-edit the computer count manually. The actual analysis then clusters the linguistic features into groups of features that co-occur with a high frequency in texts. He establishes seven such groups or factors. These factors are interpreted as textual dimensions on the basis of the shared discourse functions of their individual features as they have been established by other researchers in the relevant literature. In the next stage of the analysis, Biber computes for each factor a factor score for each text, and for each genre an average score for all its texts. The term genre is taken to refer to text characterisation on the basis of external criteria, that is to say it includes such categories as telephone conversation, press reportage, business letters and so on, whereas the term text type refers to categorisations on the basis of linguistic criteria such as the proposed factors or dimensions. It turns out that each dimension structures the entire set of texts in a different way. On dimension 1, personal telephone conversations are highly involved, whereas financial press reportage and natural science academic prose are highly informational. On dimension 2, sports news reportage and political press reportage are highly narrative, whereas technology and engineering academic prose are highly non-narrative. On dimension 3, the technology and engineering academic prose have highly explicit an,d situation-independent reference, whereas personal telephone conversations and sports broadcasts have situation-dependent reference (for a slightly more detailed summary and a review cf. Jucker 1989). This approach then is traditional stylistics on a higher level. Styles are established on the basis of the frequency of occurrence of syntactic features, but with the help of powerful computer programs. This allows the analysis of an

16

Variability in sociolinguistics and stylistics

enormous data base; it allows the inclusion of a great number of syntactic features, and makes it possible to group the syntactic features into factors according to well-defined mathematical procedures. The weakest point is presumably the limitation of the tagging program, which in several cases restricts the range of syntactic features that can be used.

2.2.2. Ethnography of speaking As in traditional stylistics, the sociolinguist tries to relate language usage to its social context and situation. But there are two quite distinct approaches within the broad domains of sociolinguistics, which approach this central question in ways that seem to be radically different. There are differences not only in the methodology but also in the aims. The research tools must be seen in relation to the aims of those using them in order to allow a fair assessment of their value. What to some researchers may seem to be un-illuminating research tools may in fact turn out to be highly successful if they are seen in the proper light of the aims they are designed for. One of these approaches is the ethnography of speaking. It was initiated by and is best known through the work of Dell Hymes (cf. Hymes 1974 Gumperz and Hymes 1986). Saville-Troike (1987) and Duranti (1988) give useful surveys of work done within this framework. Duranti gives the following definition: The ethnography of speaking (...) studies language use as displayed in the daily life of particular speech communities. Its method is ethnography, supplemented by techniques developed in other areas of study such as developmental pragmatics, conversation analysis, poetics, and history. (Duranti 1988: 210) In this approach, style is not taken to be a collection of arbitrary features, but rather as a strategy to use certain features with specific intentions, and thus the motivation for the stylistically relevant features becomes paramount. What is at issue here is the "cultural relativity of linguistic functions" ( Levinson 1988: 167). Hymes introduced the term "speaking" in the label "ethnography of speaking" because he is interested in the active, process-oriented aspect of language, and its strategic use in specific situations, in contrast to corpus studies which analyse language as a static end-product. Saville-Troike points out that there is a relative dearth of theoretical pronouncements in this approach because of the danger of premature theorising, which is particularly acute in this approach. Researchers in the ethnography of speaking must adopt crosscultural perspective and seek to formulate generalisations that are valid across

Three approaches to the concept of style

\7

cultures, such that hypotheses can be tested in new and as yet uninvestigated settings. Descriptive tasks include enumerating the kinds of speech events which are recognized or can be inferred in a community, the nature of boundary markers which signal their beginning and end, and the features which distinguish one type from another. (Saville-Troike 1987: 664) The collection of articles in Gumperz and Hymes (1972), for instance, contains many examples of detailed accounts of speech actions in specific and unique social and geographical settings. Dundes et al. (1972), for instance, investigate the strategies employed by Turkish boys in their verbal duelling rhymes. They analyse the rigid constraints that hold for the message form and for the participants. The message always consists of two parts, a challenge and a reply, and only teenage boys are allowed to participate. Characteristically there is a contrast between the overt topics of these verbal duellings - they centre around sexuality and virility - and the communicative goals or functions that are pursued by the participants. The overt topics serve a symbolic function for the actual goal which appears to be to outplay the opponent in wit and presence of mind. These duels require very specific linguistic skills and intimate familiarity both with this particular genre and with its wider cultural context. In this approach, then, style is not seen as a conglomerate of linguistic features, but as a strategy which is employed by participants to achieve specific goals in specific - and highly culture-dependent - settings. It is in this light that Traugott and Romaine (1985: 29) postulate that style should be viewed "primarily as a relationship between participants in speech events who, as individuals, negotiate speech acts and thereby create 'styles' strategically, but who also are exemplars of social roles and have relationships in larger social institutions beyond the frame of dyadic interaction, e.g. networks" (my emphasis). They present an overview of some of the pertinent approaches to "style" within sociolinguistics. They, too, stress the difference between Hymes' and Labov's work, but they point out another area of difference between the two. For Hymes the locus of style can be in the form of the message, the topic, the setting, the channel, the audience, etc, whereas Labov reduces style to variation on the level of setting. They object to Labov's view mainly on the basis that it casts the speaker in a passive role in that he merely reacts to changes in the perceived formality of the situation. Instead, Traugott and Romaine advocate an approach to style that assigns an active role to speakers who use and create styles in order to achieve communicative goals.

18

Variability in sociolinguistics and stylistics

Similarly, Levinson (1988: 183) argues that "on the strategic view, styles should be seen as systematically motivated, as essentially rational adaptations to certain contextual circumstances."

.2.3. Correlational sociolinguistics Correlational sociolinguistics has been developed and popularised in the linguistic community first and foremost by William Labov (in particular 1972a and 1972b), but also by Trudgill (1974) and Milroy and Milroy (1978) and others. This view of sociolinguistics depends on what Levinson (1988: 165) calls sociolinguistic alternates. These are two or more linguistic forms that are said to have the same meaning, with the all-important proviso that they vary in their use. They are used by different speakers or in different situations or both. Thus an attempt is made to correlate the use of the synonymous variants of a variable with such non-linguistic contextual features as the social class of the speakers, their regional origin, or the formality of the current situation they are in. Sameness of meaning between variables under investigation has been regarded as crucial for most of the studies in the field of sociolinguistics and dialectology. The early dialectologists went out to record words for particular concepts, for instance farm tools, and plotted the results on maps. In this approach it is essential that for exactly the same concept two different words are used, or, if pronunciation was to be investigated, that the same word with the same meaning was pronounced in a distinctly different way in order to show up a difference of variety. In the same way sociolinguists investigating different pronunciations and trying to correlate them with social backgrounds depend on sameness of meaning. The variant realizations of a linguistic variable do not encode different referential meanings. They co-vary with other units in the system and/or with a range of speaker variables such as social class, ethnicity, age and sex. (L. Milroy 1987: 94, her emphasis) Labov's (1972a) well-known investigation into the pronunciation of [r] in New York department stores is a case in point. The phrase fourth floor "means" exactly the same, whether it is pronounced with or without an [r]. As far as phonological variables go, this seems to be plausible. Recently, however, more and more attempts have been made to take variability studies beyond the confines of phonology (cf. Lavandera 1978, Sankoff 1988). Jacobson (1989) discusses a number of different approaches to syntac-

Three approaches to the concept of style

19

tic variation, illustrating mainly research efforts from Scandinavia. As soon as the comparatively safe realm of phonological variation is left behind, the question of sameness of meaning becomes an issue that divides researchers into two camps; those that try to choose non-phonological variables which stand in some sort of paradigmatic relationship even if it is recognised that they are not strictly speaking synonymous, and those that reject the alternates approach outside the confines of phonology. Cheshire (1982 and 1987) and Romaine (1982) illustrate the former approach albeit fairly critically by showing many of its inherent weaknesses, whereas Aarts (1971), Varantola (1984), de Haan (1987), and Raumolin-Brunberg (1991) exemplify the latter. [It is not] always illuminating (or even possible) in syntactic work to adopt the assumption which in quantitative phonological work is taken for granted: that the object of study is a set of surface variants expressing the same underlying semantic structure. (L. Milroy 1987: 143) I have used the term "sameness of meaning" rather loosely to refer to the similarity that is required for two or more variants in a correlational study, but as pointed out above, it is only part of the meaning that is left constant, while the focus is on the remaining aspects of meaning, that is the contextual features such as social class. In view of the problems associated with defining the notion of synonymy, different terms have been suggested for that part of the meaning of variables that has to stay constant. The actual choice will obviously be influenced by the problem on hand, but the one with the best claim to universality is presumably referential sameness. This term makes explicit that the variants are required to have the same referential meaning while allowing for differences in connotative meaning and, more importantly, in social and geographical meaning.

2.2.4. The sociolinguistic triangle The three approaches mentioned above, traditional stylistics, correlational sociolinguistics and ethnography of speaking have many things in common. All of them try to relate features of linguistic production to the wider, nonlinguistic context in which they occur. In terms of their respective aims and methods, as I have briefly outlined above, there exist considerable differences. Nevertheless these differences are not irreconcilable. The three approaches should rather be seen as the three angles in a triangle, allowing for numerous intermediate positions, as shown in figure 2.1.

20

Variability in sociolinguistics and stylistics

Correlational sociolinguistics Υ

Ethnography ~7 of speaking

Traditional stylistics Figure 2.1. Sociolinguistic triangle

Both correlational sociolinguistics and traditional stylistics are firmly based on the comparison of linguistic features in different contexts. The former presupposes, in its epitomic form, the free variation hypothesis, that is the belief that the same can be said in different ways. In its less epitomic form, if lexical or even syntactic variables are investigated and the notion of sameness of meaning becomes more and more problematic, this approach moves in the direction of traditional stylistics, where the density of linguistic features is the terminus comparationis. The ethnography of speaking is equally distant from both correlational sociolinguistics and from traditional stylistics in that it investigates unique functions in unique contexts. It asks for the motivation of linguistic features without comparing alternative contexts. But again many examples of studies lying between the extremes can be found. The traditional stylistician may not be content to have found style markers that occur with distinct densities in different contexts, without going on to ask about the motivation of these markers and about how they are used strategically within a given context. In a similar fashion the correlational sociolinguist may not be satisfied with having established the synonymy of the variants, but may ask questions about the use to which they are put. Levinson (1988) excludes traditional stylistics from the above picture on the grounds that it cannot distinguish between style and other types of varieties. As mentioned above, I do not think that this is a problem that is inherent in the methodology of traditional stylistics, even if it has more often than not been disregarded in this framework. The problem is exactly the same as in correlational sociolinguistics: correlational, where it indeed always has been a primary issue. I shall turn to the problem of categorising different contexts into different types of varieties in the following section.

The notions "variety" and "style"

21

Levinson also warns against carelessly eclectic approaches, urging investigators to make up their minds in which way they want to approach the study of language in its social context. In the study of style, I think we must be open to both approaches, but given the divergent methodological consequences of each approach, we cannot afford to be carelessly eclectic. It is essential to be clear about whether, in a particular case, one is handling phenomena that are best treated as different realizations of the same function, or are best treated as distinct and incommensurable functions. (Levinson 1988: 173, his emphasis) While it certainly would be rash not to heed this warning, it might be equally rash to disregard one angle of the sociolinguistic triangle presented above at the expense of another. It is indeed one of the aims of the present study to show that a carefully eclectic approach to the language of newspapers can provide insights which a more dogmatic approach employing one single point of view could not do.

2.3. The notions "variety" and "style" The term "variety" is not as vague as it may seem. I take it to refer to a subset of language - either actually produced and possibly recorded or just potential - that is characterised by a cluster of linguistic features which can be associated with non-linguistic features delineating in some way the context of production of this corpus. Thus a variety is a theoretical concept that has no independent existence beyond its perception by the analyst. Hudson (1980: 24) defines "variety" very similarly as "a set of linguistic items with similar social distribution". This allows for entire languages such as English, Swahili or Polish to be considered as varieties or very specialised uses of language such as the language used by lower-class adolescents frequenting one particular play-ground in Reading (Cheshire 1982) or the language used by D-Js between playing records on BBC Radio One (Montgomery 1988). Ferguson (1971: 30) gives a much narrower definition: Any body of human speech patterns which is sufficiently homogeneous to be analyzed by available techniques of synchronic description and which has a sufficiently large repertory of elements and their arrangements or processes with broad enough semantic scope to function in all normal contexts of communication.

22

Variability in sociolinguistics and stylistics

This definition is also broad enough to allow for entire languages to be called varieties, but it excludes more specialised uses of language that are restricted to particular situations such as the Reading adolescents' play-ground speech or the BBC Radio One D-J speech, which presumably does not have "enough semantic scope to function in all normal contexts of communication". On the assumption that describable regularities exist for such specialised uses of language, the definition of the notion "variety" has to be broad enough to include them as well. The non-linguistic features that are used in delineating varieties are categorised, and they tend to serve as labels for the resultant varieties. If geographical features are used, the varieties are called dialects. They may again be large, such as American or British English or they may be very small, such as Reading English (Cheshire 1982) or Norwich English (Trudgill 1974), but it is always the analyst, whether a professional linguist or a non-specialist, that perceives linguistic differences and correlates them with geographic differences. What may by some be perceived as huge linguistic differences between the language of a Liverpudlian and a Cockney, may for others be all but indistinguishable (cf. Quirk et al. 1985: 17). Likewise the linguist may wish to point out the differences between varieties that are themselves large and heterogeneous as the above-mentioned examples of American and British English with tens of millions of speakers (see for instance Trudgill and Hannah 1982), or they may engage in investigations of very small speech communities, for example representative samples of all the sales assistants of three department stores in NewYork City with several hundred speakers at the most (Labov 1972a). In a similar fashion, varieties that are established on the basis of nonlinguistic features that relate to the social background of the speakers (or writers) of the language subsample under investigation, are called sociolects. Layperson and linguist alike regularly correlate linguistic features with the social class of the person using these features. Here, too, the varieties that are thus established do not have an existence of their own. They only exist insofar as an observer is aware of linguistic differences in two or more language samples and correlates these differences, rightly or wrongly, with the social background of the respective speakers. For the sake of clarity, I use dialect exclusively to refer to regional varieties and reserve sociolect for social varieties, thus avoiding the hybrid "social dialect" that is still sometimes used (e.g. Wardhaugh 1986: 40, 46). A dialect is thus defined as a regional variety without any consideration of its social prestige, whereas the Continental approach to dialectology still reserves the term dialect for non-standard varieties (cf. Ammon 1987).

The notions "variety" and "style"

23

Varieties can also be established on the basis of historical features by comparing subsets of language as produced at different points in time. Thus large varieties such as Old English and Middle English are distinguished, or smaller ones such as Early Modern English or Elizabethan English. Again there is no established size of the subsamples and no established distance between samples that are to be compared. Bell (1985: 109; 1988: 338; 1991: 132), for instance, compares the language produced by four different radio stations in New Zealand in 1974 and again in 1984. Quite clearly, the two samples of language can only be regarded as two different varieties because Bell was able to isolate a linguistic feature that differs from one sample to the other and because he correlates this feature with the time difference between the two. Unfortunately there does not seem to be an equivalent hyponym of the term variety for this type of feature correlation. Historiolect might be a suitable and rather obvious candidate. Such a term would indicate the methodological similarity between historiolects and other types of varieties. In many instances, even the same features will serve to distinguish different types of varieties. To the historian of English, thou lovest is an older form than you love; to the student of contemporary styles, it is a feature of a style that one might label as "Biblical" or "archaic". To those who find you ain 't characteristic of a social class it is a class marker, but to those in whose studies it correlates with a certain range of situations it becomes a style marker. ( Enkvist 1973: 19) All the varieties dealt with above are instances of interspeaker variation or variation according to the users. In all the examples, speech samples produced by different speakers or writers are under investigation with the possible exception of Bell (1985, 1988, 1991), where the same speakers may be represented in both the 1974 and in the 1984 sample. However, there is also considerable variation in the speech produced by individual speakers in different situations, that is to say variation according to the uses rather than to the users. It is the sum of all these intraspeaker variations that I want to call stylistic variation. A style, therefore, is a variety that is established on the basis of non-linguistic features that distinguish the speech as produced by the same speaker on different occasions. To set the notion "style" on a par with "dialect" or "sociolect" is, of course, not new. As witness the following quotation by Winter (1961: 194): Es hat den Anschein, als ob die linguistische Abgrenzung von Stilarten gegeneinander viel mit einer Abgrenzung von Dialekten gemein hat. Ebenso wie das Gesamtgebiet einer Sprache auf Grund von Isoglossenbündeln in einzelne Dialektgebiete geschieden wird, kann man die Gesamtheit einer

24

Variability in sociolinguistics and stylistics

Sprache - auf anderer Ebene - als aus verschiedenen Stilbereichen zusammengesetzt ansehen, die ebenfalls durch Isoglossenbündel begrenzt sind. [It seems that the distinction between types of styles has much in common with the distinction between dialects. The totality of a language can be split up into individual dialect areas on the basis of bundles of isoglosses, and in the same way it can be split up into discrete areas of style that are also delimited by bundles of isoglosses]. Defined in such a way, the notion of "style" does not make any predictions about what causes these differences, that is it does not explain them. Again there are similarities with other types of varieties. It has long been and still is a matter of controversy whether differences in historiolects, or rather language changes, can be explained (cf. Romaine's (1983) review article of Lass (1980), who is entirely pessimistic and denies the possibility of explanations for language change, and Lightfoot (1979), who is rather more optimistic and sets up a transparency principle in order to explain syntactic change). In the same way, the analyst may perceive a clear difference between intraspeaker varieties, that is to say styles, but to establish the cause for the variation may be a matter of guess-work. In historical linguistics, it has long been recognised that a time difference that correlates with linguistic differences does not in itself explain the linguistic difference. Unfortunately the same is not true of sociolinguistics, where it is still all too often assumed that social differences correlating with linguistic differences explain the linguistic differences. As pointed out in the previous section, traditional stylisticians very often do not distinguish between different types of varieties, calling everything "style". Neither are correlational sociolinguists always consistent in their usage of the terms. Some researchers see styles entirely in terms of class groups, irrespective of the context in which the speech samples are produced. For methodological reasons it is important that a clear distinction is made between interspeaker variation (geographical, historical, social) and intraspeaker variation (stylistic). Otherwise the terms would cease to be useful. Many researchers make a distinction between style and register. They use the former to refer to different levels of formality and the latter to refer to varieties that are characteristic of occupational and professional groups or of certain topics. According to this view style and register are, in principle, independent. Styles range from informal to formal whereas registers are varieties that are used, for example, by professional football players or the employees of a bus-company (cf. Trudgill 1983b: 101, but also Spillner 1987; Varantola 1984: 3; and Zwicky and Zwicky 1982: 214-216). Halliday (1988: 162) calls a register a functional variety and defines it as

The notions "variety" and "style"

25

"a cluster of associated features having a greater-than-random (or rather, greater than predicted by their unconditioned probabilities) tendency to cooccur and, like a dialect, it can be identified at any delicacy of focus." The fact that the features can have "unconditioned probabilities" suggests that there must be some sort of neutral, unmarked form or rather a norm from which particular registers can be seen as deviations. In sections 2.4 and 2.5 below, I will argue that intraspeaker variation is better seen as a correlate of the speaker's audience. The language of a physician, for instance, will vary, not so much according to his or her profession as according to his or her addressees, bystanders and overhearers, whether they are fellow physicians, medical students, patients or members of their own families. It would therefore be an oversimplification to posit a register of medical talk. The same criteria relating to the speaker's audience are used to distinguish different levels of formality. In contrast to Labov and others, who suggest that formality is a concomitant of the amount of attention paid to speech production, I will argue that the speaker's audience is a better guide to the different levels of formality. Hence, there is no need to distinguish the two types of variety, style and register. The former term will suffice to encompass all the varieties that have been envisaged under both of them. Newspaper language is a variety to the extent that it has linguistic features that distinguish it from other varieties. It is obviously part of the larger variety of media language as a whole, and - on a different level - it is part of the variety of written language. However, I will not be concerned to set it off against these larger varieties. My concern will be its internal variability. This means that individual articles or even a collection of articles by the same author cannot be the starting-point for my analysis. Instead the articles have to be combined into intuitively plausible sets according to non-linguistic criteria. The analysis will then reveal whether the chosen division is also linguistically relevant, that is to say whether it correlates with differences in the use of linguistic features. If the relevant information about the journalists were available, one could of course attempt to correlate the linguistic features with a writer variable such as social class in order to distinguish different sociolects within the domain of newspaper language. However, the differences of social class will not be very marked. Social class rankings usually depend on the person's employment, which differs for journalists only in as far as the importance of the paper differs. Thus a journalist working on a national paper will inevitably enjoy a higher social status than one working on a small local newspaper.

26

Variability in sociolinguistics and stylistics

What differs a great deal, however, as I will show in chapter 3 below, is the socio-economic readership of the audiences of the various papers. I am therefore, within the definition given above, dealing with intraspeaker variation rather than interspeaker variation, that is to say with stylistic variation.

2.4. Stylistic variation According to Labov (1972a: 208), the decisive factor for the stylistic variation is the amount of attention paid to the speech production: "Styles can be arranged on a single dimension, measured by the amount of attention paid to speech." The more attention speakers pay to their speech production, the more formal their styles will be. Informants in sociolinguistic interviews are likely to monitor their speech fairly closely because of the unfamiliarity both with the interviewer and the situation. Incidental speech that occurs during the course of the interview and is not addressed to the interviewer but to members of the family is less closely monitored and therefore more casual. A similar type of casual speech was elicited by Labov in the course of his interviews by asking his famous danger-of-death question. When informants had to describe a situation in which their life was in danger, they became emotionally more involved in the content of what they said, and therefore the self-monitoring of their speech was momentarily reduced. The more formal styles were elicited by three reading tasks in which the informants had to read a prose passage, a word list and a list of minimal pairs of words that in the vernacular are homophonous but are distinguished in the standard language. For a long time this conception of style was widely accepted without much argument and justification. In recent years, however, a number of substantial objections have been raised. It has been questioned whether it is really the amount of attention paid to the speech production which causes the variation in the linguistic forms and whether the two styles of unscripted speech and the three reading tasks really form a continuum. L. Milroy (1987: 173), for instance, has shown that the level of literacy of the informants can seriously impair any comparisons between reading styles. In Belfast, where she carried out extensive research together with J. Milroy and research associates, they found many working class informants who were illiterate or semi-literate only and who were reluctant to tackle the reading tasks at all. Some read through the word list at great speed without showing any signs of paying more attention to their speech but the same speakers sometimes did not manage to read out the prose passage.

Stylistic variation

27

[M] any of these same fluent word-list readers tackled the task of reading aloud a passage of continuous prose rather unhappily, in a halting, dysfluent manner. Certainly in these communities, where levels of literacy were relatively low and speakers were unfamiliar with the task of reading aloud, it was not possible to assume without further investigation that reading passage and word-list style simulated the attention factor in the rank-order proposed. (L. Milroy 1987: 173) Labov's conception of style was designed to mirror closely the variation that occurs in natural language production from the most formal to the most casual variety, ordered according to the amount of attention paid to the speech. However, if there is so much doubt about the attention factor, it is in fact unclear what kind of natural language production the three reading tasks are supposed to mirror. It may well be the case that the fairly formal interview style and the more casual style used in incidental talk during the interview and in the danger-of-death stories have got very similar real-life counterparts, but this is more than doubtful for the reading styles. Very few people, one would assume, regularly read out written texts to an audience, and even fewer regularly read out word lists. [W]hile "casual speech" in the linguistic interview may well be an analogue of spontaneous everyday conversation, it is hard to imagine a naturally occurring situation analogous to reading minimal pairs. The only time a speaker is likely to use minimal-pair style is when reading minimal pairs - a rather uncommon speech event. (Bell 1984: 150) The stylistic variation has always been of vital importance in sociolinguistic studies because it is only the interplay of the social variation with the stylistic variation that can reveal patterns of language change in progress. Usually a distinction is made between the variables that vary along the social axis only (indicators) and those that vary both on the social and on the stylistic axis (markers). The two differ in that speakers are aware of the significance of the markers, but not of the indicators. In more careful styles, they aim consciously to approximate their speech to the speech of the higher status social class, and thus increase the percentage of markers. The indicators, however, escape the conscious attention of the speakers, and therefore they do not show any significant differences in the different styles. Thus, according to the Labovian view of style, speakers are aware of the social differences in a class system, and they normally aspire to the prestige of the higher status class. Therefore they will use those language features that are commonly associated with the higher social prestige whenever they pay enough attention to their speech, or rather they will use increasingly more of these markers in line with the increasing amount of attention.

28

Variability in sociolinguistics and stylistics

Bell in a number of places (e.g. 1977, 1982, and in particular in his important 1984 paper) has suggested a rather different interpretation of intraspeaker variation, which offers a particularly plausible explanation for the variation in his corpus of media language which otherwise would be very difficult to explain. For this reason, I shall review his work in some detail.

2.5. Audience design Bell (1977) found that there are considerable style differences among news broadcast on different New Zealand radio stations. Sometimes the same speakers read the news on different stations, and they always show a systematic variation across these stations. In traditional methodology this variation cannot be accounted for, because in all instances they are using "reading style", and it cannot plausibly be claimed that the amount of attention paid to their speech differs according to the station on which they happen to broadcast. What is different though is the targeted audience of the different stations, and according to Bell this is the decisive factor. The intraspeaker variation is explained in terms of the social status, sex and age not of the speaker but of the hearers. The hearers are not just the addressees but include also auditors, overhearers and, to some extent, eavesdroppers, as shown in figure 2.2.

1st person 2nd person 3rd persons

Speaker Addressee Auditor Overhearer t Eavesdropper

Figure 2.2. Persons and roles in the speech situation (Bell 1984: 159; 1991: 91)

The speaker stands at the centre of the diagram because it is always the speaker's characteristics that account for most of the variation. On this level, we are talking about the interspeaker variation. Bell claims that the interspeaker variation is always paramount whereas the intraspeaker variation is derivative of the interspeaker variation because of the existence of indicators, which vary on the interspeaker level only.

Audience design

29

Bell (1984: 152-158) shows that in practically all the data on sociolinguistic variation published so far the overall interspeaker variation exceeds the overall intraspeaker variation. To take a hypothetical example, if for a particular variable there is a 75 per cent difference between the speech of the top and the bottom classes in casual speech, there will be at most and presumably rather less than a 75 per cent style shift for any of the classes. This leads him to formulate the following "style axiom": Variation on the style dimension within the speech of a single speaker derives from and echoes the variation which exists between speakers on the "social" dimension (Bell 1984: 151) Bell does not seem to consider the possibility that this may also be a reflection of the type of variability studies carried out so far rather than a reflection of universally valid regularities in speech communities. He does, however, cite two cases violating his style axiom. These are the hyper-correction by the second highest status group to a level over and above the level achieved by the highest status group in the most formal style (e.g. Labov 1972a: 114) and hyper-style variables in Tehran Persian reported by Jahangiri and Hudson (1982). In spite of the cross-over patterns reported by Labov and elsewhere, the overall intraspeaker variation is still smaller than the overall interspeaker variation in all the published data investigated by Bell. The variation in Tehran Persian shows a very considerable intraspeaker variation with very little interspeaker variation and thus is a true exception, even though there is only a really marked shift between the two reading styles and the two styles of unscripted speech. Within the individual modes there is very little variation. Therefore he concludes that this extensive intraspeaker variation is a reflection of the discontinuity of the two activities of talking and reading. All the other data investigated by Bell support his style axiom. However, there is no reason to exclude the possibility that the situation as described for Tehran Persian exists in other speech communities as well, even if such cases have as yet not been investigated or published widely. There does not seem to be anything unusual about speech communities in which there is very little sociolinguistic or interspeaker variation while its members control a repertory of widely disparate styles. Cases of diglossia would, of course, be the extreme cases, as for instance in German speaking Switzerland, where there appears to be little social differentiation while the two situationally bound varieties, Standard German and Swiss German, are markedly different. Nevertheless, even in such cases, it is plausible to attribute the intraspeaker variation to the influence of the audience. Bell distinguishes between addres-

30

Variability in sociolinguistics and stylistics

sees, auditors, overhearers and eavesdroppers. The addressee, as the one who is addressed, ratified and known is the main character among the speaker's audience (as indicated by figure 2.2). Next in importance is the auditor, who is known and ratified but not directly addressed. The overhearer is known but neither ratified nor addressed, and the eavesdropper is not even known by the speaker to be among the audience and consequently can be neither ratified nor addressed (cf. Bell 1991: 90-94). As outlined above, Bell claims that the variation induced by all the hearers (the intraspeaker variation) is smaller than the interspeaker variation. And he goes on to assert that within the intraspeaker variation the influence of the addressee is the most significant. However, in spite of the various studies quoted by Bell (1984: 163-169, e.g. Giles and Smith 1979; Coupland 1984; Thakerar et al. 1982), which support the ordering of the hearers as suggested in figure 2.2, it is again doubtful whether this should be a universal regularity of all speech communities. No matter how many studies confirm the suggested - and eminently plausible - ordering of the hearers, there can be no proof that in some other, as yet uninvestigated, speech community a different ordering may not obtain. The variations considered so far are all responsive because they represent in some way a reaction to the speech of the audience. There are, however, cases in which speakers do not have direct feedback from their audience and, in the absence of such cues, reinterpret the situation by choosing a particular style. In such cases Bell (1984: 161) talks about initiative speaker variation, which occur as "as a response not to the immediate audience but to certain third persons not physically present. These are reference groups, who are absent but influential on the speaker's attitude." These are cases of referee design (Bell 1991: 126-146). One obvious set of data which can be reinterpreted within this framework is Labov's (1972a) classic investigation of the social stratification of M in New York City department stores. Labov does not make any claims as to the social stratification of the sales assistants whose language he surveyed. On the contrary, he quite explicitly assumes that the employees of all three stores would probably fall into the same socio-economic group (1972a: 48). They are stratified according to the location of the shops, the papers in which they advertise, their price policies, and the physical appearance of the stores, that is to say the ratio, in impressionistic terms, between the available floor space and the merchandise on display. Thus the employees apparently adapt their language to the clientele of their shops, (cf. Coupland's research (1984) of how a travel assistant converges to the language of her customers.) The encounters in Labov's rapid and anonymous elicitations were so short that the employees

Audience design

31

can hardly be said to converge to the language of the actual customer. In that case rather less variation might have been expected as the researcher in all instances was Labov himself. Thus we must assume that the stratifying factor was some sort of model customer corresponding to the targeted clientele of the individual store. Bell investigated the stylistic variation of New Zealand news readers on different radio stations. Some newsreaders may read the same news on two or more stations that use the same suite of studios. The amount of intervocalic /t/ voicing, for instance, varied according to the station on which the news was read (cf. Bell 1977: 347-359; 1982: 162 and 1984: 171). Figure 2.3 shows the amount of intervocalic /t/ voicing of four newsreaders on two different stations. Station YA carries the national programme network from Wellington, which appeals mainly to an older audience with above average education, whereas ZB is the local community network station, which appeals to a middle status and mainly family audience in the thirties and forties age group (Bell 1982: 152 and 157). Percent 50 r 40 30 20 10 0

YA ZB one

YA ZB two

YA ZB three

YA ZB four

Figure 2.3. Percentage of intervocalic /t/ voicing by four newscasters on two New Zealand radio stations (Bell 1982: 162)

These variations are initiative because the newsreaders do not have any direct feedback from their audiences. All media language is initiative style design. It creates the relationship between communicator and audience, rather than responding to an existing relationship. This holds supremely for radio, where announcers rely solely on their speech to project whatever relationship they have with the audience. Their style draws its effect from the norms of who such a style is addressed to in face-to-face interaction. (Bell 1984: 192)

32

Variability in sociolinguistics and stylistics

As mentioned above, Bell developed the audience design model with the explicit aim to account for variation in media language. Journalists, whether on television, radio or in the print media, adapt their language to their targeted audiences, without changing the amount of attention paid to the production of speech. What is perhaps surprising is how easily many studies carried out in a Labovian framework can be reinterpreted in an audience design format.

2.6. Some recent studies in media language In the last section of this chapter, I want to review some pertinent studies on media language by showing how they fit into the sociolinguistic triangle presented in section 2.2 above, and how they can be interpreted in an audience design framework.

2.6.7. Crystal and Davy - "Journalese" versus "normal" English Crystal and Davy (1969) point out that the term "journalese", useful as it is, is also deceptive in that it suggests far more coherence than there actually is among the entire body of language that comes under its scope. To make their point they compare two newspaper articles, written in 1965 about the introduction of computers into weather forecasting and published in the Daily Express and The Times respectively. They note that the sentences in the Times article are longer and that there are more sentences per paragraph. Within the statement-type sentences, they note three important features. Verbs of speaking, such as say, declare or explain are often inverted with the subject as in Said Dr Mason .... The second feature is the great number of adverbials that appear in emphatic clause-initial position rather than in the unmarked post-verbal position even in contexts which do not suggest a particular emphasis for the adverbial. Thirdly they observe that there are hardly any instances of coordination at clause level in the Daily Express, article but that such instances are common in the Times article, which also has more subordinating constructions at the clause level. There is the same amount of subordination within noun phrases but the Daily Express uses more non-finite -ing clauses and relative clauses, whereas The Times uses more non-finite -ed clauses. This approach to style must be placed in the middle between traditional stylistics and ethnography of speaking. Crystal and Davy relate the linguistic features to their non-linguistic contexts. But they also assume, in the larger context of their investigation of English style, that the linguistic form of an

Some recent studies in media language

33

utterance will carry direct information on - among other things - the regional, social and historical provenance of the speaker or writer (1969: 81). They want to know why a particular form was chosen rather than another. In this approach the comparative element recedes somewhat, even though it is implicitly still present in that particular features derive their significance from the fact that they appear in one variety but not in another. The aspect of unique functions in unique contexts comes more into focus. In Crystal and Davy's words: The aim of stylistics is to analyse language habits with the main purpose of identifying, from the general mass of linguistic features common to English as used on every conceivable occasion, those features which are restricted to certain kinds of social context to explain, where possible, why such features have been used, as opposed to other alternatives and to classify these features into categories based upon a view of their function in the social context. (Crystal and Davy 1969: 10) The usage as it is attested in their two samples of newspaper language is always contrasted to what they call "normal usage". Thus they note "the presence of much more complex pre- and postmodification than we normally hear or write", or that "commas are absent from many places where they would normally be expected" (1969: 186 and 178, my emphasis). Elsewhere they suggest that inexperienced analysts start out by comparing their texts to one particular variety set up as a norm for this purpose, maintaining that the more experienced analyst does not need to resort to this. "The norm should not be taken as an obligatory category of stylistic theory, only as an optional procedural device" (1969: 91, fn 8). Thus Crystal and Davy compare two different texts, and they relate both to an implicit norm, but they also analyse the features of newspaper language as unique features in a unique context, trying to assess their impact on this particular type of English. The difference is that an ethnography of speaking approach would not content itself with a list of register features, but it would try to establish the intentions behind these features, on the assumption that they are strategically used in order to achieve specific aims. The variation that exists is clearly seen as a result of the journalists adapting their language to the targeted audience: "the disparity can largely be explained by reference to the very different audiences envisaged by the two papers" and "not only is different language necessary to suit the different temperaments of their respective audiences, but a different type of information is also given" (Crystal and Davy 1969: 174 and 189). In spite of the lack of explicit information on the respective audiences of The Times and the Daily Express, Crystal and Davy clearly anticipate an audience design framework.

34

Variability in sociolinguistics and stylistics

O'Donnell and Todd (1980: 85-100) give a very similar comparison between The Guardian and the Daily Mirror, noting some outstanding features in a non-quantitative, impressionistic way. Among other things they mention that the headlines in The Guardian have a tendency to avoid finite verbs, whereas the Daily Mirror tends to avoid verbs altogether in its headlines. They observe the differences in referring to the then British prime minister. The Guardian uses either Mrs Thatcher or the Prime Minster, whereas the Daily Mirror prefers Premier Margaret Thatcher or just Maggie. And they list several lexical items which they claim are more likely in papers like the Daily Mirror than in The Guardian such as horror, which is used for every more serious crime or accident or probe, which applies to all types of investigations. Further similar items are rap, pact, closure, claims, call, cuts, and freeze.

2.6.2. Carter and Ghadessy - Newspaper vocabulary Carter (1988) and Ghadessy (1988) concentrate on features of the vocabulary in British newspaper language. Carter analyses a home news article of the Daily Mail, whereas Ghadessy uses a larger corpus of articles from the sports section of The Times. Carter first defines the notion of core vocabulary (1980: 9f)· Core words tend to have antonyms, they are less collocationally restricted, they are more basic in that they are regularly used to define non-core words, they carry less emotive meaning, are less discourse- or register-specific and they are often superordinates rather than hyponyms. Carter takes it for granted that "newspaper reports should ideally report the facts in as core a vocabulary as possible" (1988: 10, his emphasis). On this basis he discusses in detail a front page article of the Daily Mail on the Labour Leader, Neil Kinnock. This article is seen as a gross deviation of objective press reporting, and the notion of core vocabulary is used to show where these deviations occur. He shows that there are a fair number of non-core words which are either markedly formal or markedly informal, or which are recognisably and usually negatively evaluative. As examples he lists among others snub, ducking, row, buttonholing, posing, and trendy. It may be questioned, however, whether journalists really should stick to a core vocabulary as defined by him. A non-core word like corpulent, which has not as clearly an antonym as, fat and which is collocationally more restricted, may nevertheless be more appropriate in a specific context. The same is true for words like snigger, grin, smirk or beam, which are all more specific - and therefore less core - than smile. Carter also criticises the use of metaphors as for instance in the opening sentence of the article:

Some recent studies in media language

(1)

35

NEIL KINNOCK, just elected Labour's youngest leader at 41, saw an old party tide threaten to swamp his new beginning last night (Daily Mail, 8 October 1983, quoted by Carter 1988: 11)

Carter (1988: 12) comments that "this statement is rhetorically marked by numerous metaphoric devices and figures which contrast with and contravene expectations of a plain style". While it is true that these devices are used to produce a highly tendentious article in this particular instance, it should not be overlooked that it is not the use of non-core vocabulary and of metaphors itself that makes it tendentious but the way in which these devices are employed. Carter does not use a second corpus in order to compare the language used in this article but he seems to have a very clear notion of a norm from which this article is seen as a deviation. Ghadessy (1988) points out the importance of written sports commentaries as a discourse genre. There are many specific words which do not exist or are used differently outside this genre, such as box, free kick or hat trick. It is a characteristic feature of written sports commentary that it narrates events which have already taken place, and that "although it is reporting a new event, there is a large body of knowledge and values assumed to be shared by the writer and the reader. The specialist terminology used need not be explained throughout the report unless a new term is coined" (Ghadessy 1988: 21). In direct contrast to the article by Carter (1988), which immediately precedes this article, Ghadessy suggests the term core vocabulary to apply to this specialist terminology in as far as these words have a currency even outside the English speaking countries. Ghadessy also distinguishes between neutral, factual language and emotive, subjective language. Carter tries to capture this distinction with the terms core versus non-core vocabulary, whereas Ghadessy talks of uninvolved and involved language. Ghadessy lists in detail many of the syntactic patterns that sports terminology enters into in his corpus of soccer commentaries of The Times. He does not compare these uses to any other type of newspaper language or to any kind of norm, but stresses that this would be the next and essential step. "A variety in the sense of 'register' can only be established when the findings are compared with the findings based on the analyses of other registers" (Ghadessy 1988:34).

36

Variability in sociolinguistics and stylistics

2.6.3. Ferguson - Sports announcer talk Ferguson (1983) investigates the peculiarities of a contextually very restricted type of speech, the speech of radio broadcasters reporting on a game of baseball in progress. He identifies selected syntactic characteristics for which he claims the status of register markers. Sentences often lack a sentence-initial noun phrase (e.g. [It] hit on the foul line), a copula (e.g. McCatty [is] in difficulty) or both (e.g. [It's] a breaking ball outside) (1983: 159). This feature is also prominent in other registers that rely on terse and concise formulations such as event-reporting headlines or advertising language. The most characteristic feature of sports announcer talk, according to Ferguson, is the inversion of subject and predicate under certain conditions. The subject is typically a player's name, while the verb is the copula or a verb of motion as in the following two examples Holding up at third is Murphy and Pete goes to right field and back for it goes Jackson (1983: 160, 161). There are two ways of expressing a result that "are very frequent in SAT [i.e. sports announcer talk] and rare in other kinds of talk" (1983: 161). They are the pattern for + noun and to + verb as in Joe Ross's caught it for a touchdown and And it gives us a double to Mumphries to lead things off (1983: 61). Sports announcer talk is further characterised by heavy noun modifications, which are, according to Ferguson, "more typical of written English" (1983: 163), such as David Winfield, the 25-million-dollar man, who is hitting zero, five, six in this World Series, or The quiet Texan Tommy John delivers. This feature contrasts with the conversational character induced by the frequent deletions mentioned above. Ferguson (1983: 163) asserts that "the referential function of adding incidental background information about persons mentioned in the discourse" could be realised by more typically conversational devices such as adding the information in a separate sentence. Ferguson further mentions tense usage and notes that the simple present is preferred to refer to short actions that are taking place at the time of speaking, especially in sports characterised by successions of rapid events, whereas the progressive is used to refer to actions of extended duration especially in sports characterised by continuous events such as boat racing or horse racing. But he concedes that this usage can hardly be taken to be characteristic of sports announcer talk because it "seems to be in full accord with the general analyses of the semantic values of English verb categories" (1983: 164). The last feature analysed by Ferguson is the use of routines, and in particular the routine of "giving the 'count'". The format is used invariably to give the number of balls followed by the number of strikes, both in cardinal num-

Some recent studies in media language

37

bers, conjoined by and. Zero is given as oh, or, if it refers to the number of balls, nothing. Optionally this may be followed by the preposition to and the name of a player as in the following examples: Count of one and one to M; Two and oh; and Nothing and one count. Ferguson has to be placed somewhere in the middle of the sociolinguistic triangle. He, implicitly or explicitly, adopts all three points of view. He frequently refers to "other kinds of talk" (161) or "the registers of casual conversation" (158). Thus he invokes a kind of norm with which he contrasts sports announcer talk by pointing out constructions that exist in one but not the other or are far more frequent (in impressionistic terms) in one than in the other. In this respect his approach owes a lot to traditional stylistics. But he also compares different realisations of the "same function" as in the cases of the result expressions and the heavy modifiers. In this respect, then, his approach has strong similarities with correlational sociolinguistics. However, it is only the realisations that are attested in sports announcer talk that are of interest to Ferguson. He is fairly vague and non-committal about their realisation in other varieties of language. The function "giving the 'count'", for instance, is of interest mainly because of its uniqueness in this particular type of talk. Hence there are also strong affinities to the ethnography of speaking.

2.6.4. Wallace - Sports versus news section Wallace (1977) compares the language of the news section with the language of the sports section in the two papers Chicago Tribune and Champaign- Urbana Courier. On the basis of style manuals for journalists, he selected five features for his comparison, two non-quantitative ones and three quantitative ones. The first nonquantitative feature is the use of expressions that add "colour" to the stories. These can be isolated lexical items that provide additional descriptive detail while not being essential to the facts of the story, as in Ford leads Carter by a shaky single point or the campaign and debates have failed to produce any exceptional enthusiasm (1977: 54). In this respect there is little difference between the news section, in which these examples appeared, and the sports section with examples like Michigan had to scrap for victory or this yawner of a World Series (1977: 55). He further notes the use of technical vocabulary, which may entirely elude the uninitiated, such as hat trick or sacrifice fly in the sports section and detente or filibuster in the news section. He also notes that the more unified subject matter of the sports section and the fact that many of its readers will be regular observers or fans lead to the greater freedom for journalists to use technical expressions without glosses.

38

Variability in sociolinguistics and stylistics

As a second non-quantitative feature he compares what he calls "descriptive quote words", by which he means speech act tags which are not as neutral as say, tell or state, but add some contextual or attitudinal dimension such as testify, accuse or argue. He finds that there are more descriptive quote words, and they are found in greater variety, in the Chicago Tribune than in the Champaign-Urbana Courier. In the former, but not in the latter, this feature could be used to differentiate the two registers: news section and sports section. To summarise his findings for the non-quantitative features, he maintains, somewhat unilluminatingly, that "news and sports registers do vary in vocabulary, but that the strategies of selection of the lexical items for adding color are the same in both" (1977: 59). As a first quantitative feature, Wallace compares the ratio of sentences per line as a measure for sentence length, and he used a z-test to test for statistical significance. In both papers the sentences are shorter in the sports section, but the difference between the news section and the sports section proves to be statistically significant at the one per cent level only for the Chicago Tribune while it is not significant at this level for the Champaign-Urbana Courier. The second test compares the number of passive verbs in proportion to all finite verbs. In both papers there is, at a level of statistic significance, a higher proportion of passive verbs in the news section. For this parameter, Wallace sets up an additional corpus for "(approximately) all American English" (63) or - as he calls it elsewhere - "a random sample of texts" (62). This corpus contains, among other texts, White House transcripts, some pages from a book on linguistics and from another one on biochemistry, and a Playboy article. The news sections of both the Chicago Tribune and the Champaign-Urbana Courier turn out to have more passives than this " norm", whereas the sports sections of both papers have fewer. The third quantitative feature, finally, are what Wallace calls mummy passives. The recommended structure of news stories gives the main points at the very beginning of the article and relegates background material to subsequent parts, on the assumption that this structure facilitates the quick browsing of an article by readers with not enough time for a thorough reading. The hypothesis was that there is a significant difference in the proportion of passives in the first four paragraphs (i.e. mummy passives) and the background passives in the rest of the article. But it can only be supported for the news section of the Chicago Tribune its sports section and both sections of the ChampaignUrbana Courier do not show any significant difference in this respect. On the basis of these five features, Wallace (1977: 67) concludes: "There is then support for positing that news and sports stories represent different regis-

Some recent studies in media language

39

ters by their use of language, but that this variation occurs within the restricted language of newspapers." Wallace thus carries out a meticulous comparison, partly quantitative and partly non-quantitative, of two different texts. For some features moreover he uses an explicit norm as a basis for comparison. He compares the relative frequency of syntactic features, but he does not investigate alternate realisations of elements sharing the same meaning. In this respect, he works very much in the framework of the traditional stylisticians. His comparisons are subjected to rigourous statistical testing, which is not always the case in traditional stylistics. In further contrast to some exponents of traditional stylistics, Wallace uses a carefully delimited corpus. It is restricted to a "Midwestern dialect of American journalese" (1977: 71, fn 13), and thus should have considerable internal coherence, while the internal variation that actually exists is focused on in the investigation. The comparative norm of "(approximately) all American English", however, seems to be rather dubious in its mixture of contextually almost entirely unrelated texts.

2.6.5. Floreano - British newspapers and radio news Floreano (1986) sets out from the premise that there is a well-known difference in the British newspapers between the "quality" papers and the "popular" papers (I will say more on this dichotomy in chapter 3 below). This difference manifests itself not just in the outward appearance and in the number of articles that appear in one issue but also, among other things, in the language of the respective papers. Floreano tries to ascertain whether there is a similar difference across the language of different radio channels. In his newspaper sample he uses data from The Guardian and The Daily Telegraph to represent to "quality" papers and The Sun and the Daily Mirror to represent the "popular" newspapers. In the radio sample, he uses data from the BBC Radio One news bulletin Newsbeat and the BBC Radio Four news bulletin Six o'clock news. He gives the readership profiles of the newspapers in terms of social class, sex and age. For the radio channels he unfortunately can only provide exact data for the variables sex and age but not for social class. However, on the basis of the available information about the respective audiences, he hypothesises that Radio One, which broadcasts mainly new pop music, will linguistically be similar to the "popular" papers, whereas Radio Four, which is mainly a word channel, devoted to current affairs, news, radio plays and educational programmes, will be similar to the "quality" papers. He correlates the audience profiles with several linguistic variables such as sentence length, sentence complexity, noun phrase modifications, and the use

40

Variability in sociolinguistics and stylistics

of different tenses. As expected he finds that sentences are longer in the "quality" newspapers than in the "popular" newspapers. But the difference between Radio One and Radio Four seems to be less marked. The noun phrase modifications are more complex and there are fewer pronouns in the "quality" newspapers. In terms of noun phrase complexity, both the Radio One and the Radio Four news bulletins are more similar to "popular" newspapers than to "quality" newspapers. There is, however, one difference that is clearly paralleled in both media in that Radio One uses more pronouns than Radio Four. Floreano uses a great range of linguistic and non-linguistic variables, which is both the strength and the weakness of his study in that he can explore many potentially fruitful paths but has to face the inherent danger of getting lost in the wilderness. None of his linguistic variables relies on a paradigmatic relationship, and thus, within the sociolinguistic triangle, his approach must be classified as traditional stylistics but he at least tries to provide a coherent characterisation of his samples in non-linguistic terms, which is a feature of correlational sociolinguistics.

2.6.6. Ryden and Bell - Determiner deletion Ryden (1975) and Bell (1985) both investigate the use of noun phrase name appositions in newspaper language and in particular the spread of phrases like Opposition Leader Neil Kinnock with a descriptive noun phrase appositive without a determiner preceding the name. This format is relatively recent and is in Britain largely but not entirely restricted to the two categories of tabloid papers. Chapter 9 will deal with noun phrase name appositions, and therefore I will have to review Ryden's and Bell's research in more detail in that context. Here it will suffice to point out how they fit into the sociolinguistic triangle. In spite of considerable methodological differences, they are both examples of correlational studies. The noun phrase name apposition is the variable and the four patterns with preposed or postposed descriptive appositive either with or without a determiner are the variants, which are identical in referential meaning. The distribution of the patterns is then correlated with different types of newspapers, or with different historical stages. Ryden uses a descriptive framework and is mainly interested in pointing out the marked differences in the frequency of the patterns in the different papers and according to a semantic classification of the descriptive appositive as "political noun", "sports noun" or "other noun" (Ryden 1975: 16f)· Even though he points out that the pattern with preposed descriptive appositive without a determiner is most common in the "popular press" (1975: 37), he

Some recent studies in media language

41

does not correlate the linguistic variable in any systematic way with the nonlinguistic characteristics of the newspapers. Bell, on the other hand, uses the concept of a variable rule. This rule deletes a determiner in the preposed descriptive appositive of a noun phrase name apposition. The frequency of rule application is then correlated with various extra-linguistic factors. Geographically, the rule applies at different frequencies in different countries. While it applies almost invariably in American English newspaper language, there is considerable variation in its British English counterpart. Socially, the rule application depends - within a British English context - on the readership profile of the newspapers. And historically, the frequency of rule application has increased steadily in certain segments of the British media. As I will point out in detail in chapter 9, Bell's approach has some undesirable consequences because he fails to take proper account of the different preferences for either preposing or postposing the descriptive appositive. In my sample corpus of home news section articles of the Financial Times, for instance, there are 126 instances of noun phrase name appositions, two of which are preposed without a determiner. As there are only five instances of a preposed descriptive appositive that do have a determiner, the frequency of rule application, which only takes into account preposed descriptive appositives, turns out to be 29 per cent. By comparison in the home news section of The Guardian, there is one instance of a deleted determiner in 147 noun phrase name appositions. But The Guardian prefers preposing and therefore has 74 instances of preposed descriptive appositives with undeleted determiners, which gives a rule application of 1.3 per cent. Hence, the differences in rule applications must be considered to be highly deceptive. Nonetheless, Bell's approach must also be commended because he correlates the linguistic variation in a more systematic way with different dimensions of non-linguistic factors, and because this led him to the formulation of the audience design framework, which offers a more plausible account of stylistic variation than alternative attempts so far have succeeded in doing, and not just in the area of media language.

2.6.7. Verschueren - Metapragmatic metaphors Verschueren (1985) investigates media language by presenting a case study rather than by compiling a random sample. For this purpose he chose an incident with clear international significance set far enough in the past "so that a proper perspective is readily available on what really happened" (1985: 33). He found this requirement best fulfilled in the U-2 incident, in which an

42

Variability in sociolinguistics and stylistics

American spy plane was shot down over Soviet territory in May 1960, with serious effects on East-West relations. As a corpus he used all the reports on this incident and its aftermath published by The New York Times. The entire incident, he argues, derives its significance for the media not just from what actually happened but to a large extent from the verbal reactions and counter-reactions to the events in the form of statements issued by the two governments in Washington and Moscow. Thus he investigates the use of meta pragmatic terms or linguistic action verbials which he defines as "verbs and verb-like expressions used to describe (aspects of) linguistic action" (1985: 34, his emphasis). Verschueren intends his study to reveal how members of one culture (Americans), as members of that culture and as professionals whose work is governed by a range of restrictions imposed on them by an established institution (journalism in a free-press tradition), interpret and present the communicative behavior (constituting the core of the reported event) displayed by other members of their own culture (American politicians and spokesmen) as opposed to members of a different - even antipodal - culture (Soviet political leaders and their representatives). (1985: 38) In order to do this, he opposes "neutral" linguistic action verbials such as said, declared, or commented, which might be expected to be particularly common in a free press tradition priding itself on being "objective" and "impartial", to metapragmatic metaphors such as admitted, pointed out, or seized this opportunity also to repeat. In this way, he investigates whether the framing of the communicative events was biased. The use of a certain linguistic action verbial depends both on the reported communicative event itself and on the journalist' s interpretation of this event. The investigation reveals a marked difference in the reporting of the verbal behaviour of the two leaders, Nikita Khrushchev and Dwight Eisenhower, which apparently corresponded to a large extent to the facts, however, "it completely disregards the fact, usually not so much neglected where other types of human conduct are concerned, that the same surface activities (in this case utterances) may carry different weights in different societies" (1985: 98). Therefore it may be accurate to report Khrushchev as having "exploded" in public, but this should be seen in the context of the culture-specific norms of speaking styles. Eisenhower, on the other hand, is shown in a more statesmanlike way because of his more restrained verbal behaviour. Verschueren does not compare the language used in his corpus with a norm of any kind. He is interested in the use of metapragmatic metaphors in this particular and carefully selected context, which is not necessarily representative of newspaper language at large. His approach, therefore, is an exam-

Some recent studies in media language

43

pie of the ethnography of speaking because it investigates unique functions in unique contexts. The motivation for the individual functions is one of the central issues.

2.6.8. Lüger - Newspaper text types Lüger (1983) presents a textbook introduction to German newspaper language. He reviews a considerable number of studies of newspaper language, predominantly but not exclusively German newspapers. He distinguishes normative or prescriptive treatments, studies of newspaper ideologies, studies of language comprehension, studies of the function of newspapers in a social context, and finally linguistically motivated studies. What is noteworthy about his review of the linguistically motivated studies, is the way in which he splits up the field. He distinguishes three types of approaches on the basis of the status assigned to the concept newspaper language (1983: 22f). Some researchers apparently take it to be a good approximation to everyday language in general, providing, for example, an easily available corpus for diachronic investigations. None of the approaches reviewed above would fall into this category. The second type of studies recognises newspaper language as one particular variety of language that can be compared with other varieties such as radio or television language. And the third type of studies is devoted not to newspaper language as a whole, but to one single newspaper or magazine, and does not make generalisation about newspaper language in general. Verschueren (1985), dealing with the language of The New York Times, falls into the third category while all the other studies reviewed above would fall into the second category. Lüger's main concern, however, is to show newspaper language as it is used in different sections of newspapers. He deplores the restricted view of other scholars studying newspaper language, because most of them restrict themselves to the lexical choices and/or to sentence grammatical considerations without appreciating the textuality of newspaper language. On the basis of the underlying intentions, he distinguishes five basic classes of texts with several subtypes. Informative texts contain mainly facts without any evaluation. Its subtypes are hard news, soft news, news in brief, report (Bericht), personal report (Reportage), critical review of a state of affairs (Problemdarstellung), and weather report. Persuasive texts, on the other hand, evaluate, give opinions and argue explicitly for these opinions. Here the subtypes are leader, commentary (Glosse) and review.

44

Variability in sociolinguistics and stylistics

Instructive texts give practical information and guidance, for instance on gardening, (Handlungsanleitung) or counselling and advice (Ratgebung), for instance on food or on cars. The latter differs from the former in that it leaves more freedom to the readers by taking into account varying needs. Hence a recipe in which a number of steps must be taken in order to achieve the desired result would be an example of the former, whereas an article on a number of ski resorts in the alps which points out the respective advantages and disadvantages would be an example of the latter. Dyadic texts have two dominant and distinct text intentions. These are either interviews with a succession of questions and answers or agony columns in which a question concerning matters of the daily life is posed by a reader followed by an answer supplied by an "agony aunt". Contact creating texts, finally, are designed to arouse the readers' or the potential readers' interest with the aid of photographs, catchy headlines and the like. Such texts appear mainly on the front page of newspapers and have an advertising function, particularly so for the down-market papers, which in Germany depend almost exclusively on street sales whereas the up-market papers are mainly subscription papers. Lüger recognises that many newspaper articles are not unambiguous instances of any one of these text types. They may have a mixture of more than one intention underlying them. But Lüger's point is that usually one intention is predominant to such an extent that a classification is possible. What Lüger offers in effect is not so much an analysis of unique functions in unique contexts, but a classification of possible contexts in which the descriptive details as regards the lexical and grammatical choices serve the sole purpose of classifying the contexts. The distinction between a report (Bericht) and a personal report (Reportage), for instance, is mainly based on the fact that the latter employs a more personal point of view, which is indicated by deictic elements suggesting that the reporter is an eyewitness of the reported events at the time of writing the article. Hence it is not an empirical result that writer-centred deictics occur in a personal report (Reportage), but a matter of definition. For this reason, Lüger's classification can only serve descriptive purposes by drawing attention to the multitude of varying intentions lying behind different newspaper articles. Any other classification of newspaper texts may produce categories that are just as plausible.

Some recent studies in media language

45

2.6.9. Conclusion This brief review of some pertinent studies of the language of the media is not designed to do full justice to the depth and breadth of these studies. What I have tried to argue is, on the one hand, how all the different approaches necessarily influence - and in some sense also restrict - the range of their possible findings. On the other hand, I hope to have also demonstrated that important insights can be gained within quite disparate methodological frameworks. I began my review with Crystal and Davy (1969), who are the best representatives of traditional stylistics. Carter (1988) and Ghadessy (1988) also fall into this category even though they fail to provide a cogent comparison of their chosen text or texts with similar texts or with some kind of norm. Floreano (1986) also works within a traditional stylistics framework, but he tries to correlate the linguistic variables systematically with non-linguistic ones such as social class, sex and age of the actual or targeted audiences. Ryden (1975) and Bell (1985), in spite of their methodological differences, both illustrate a correlational approach. Ferguson (1983) and Wallace (1977) are more difficult to classify because they combine different points of view. Verschueren (1985) and Lüger (1983) are more clearly examples of the ethnography of speaking, because they analyse their corpus in functional terms, showing how language is used strategically in a particular social and communicative situation in order to achieve certain goals. All these approaches - within their restricted perspectives - yield important insights into a particular variety of language. And to this extent it would be unwise to choose one approach at the expense of another. Thus I advocate a carefully eclectic approach. Such a view is most likely to safeguard the analyst from partial sightedness, and it certainly safeguards against taking any analytic framework at face-value. What is needed are not purist approaches but approaches that are as clear and explicit about their methodological repercussions as possible. This will not do away with the truism that the methodology to some extent always influences the results but it will at least go some way towards making clear the extent to which it does so.

3. Socio-economic differentiation of the British dailies 3.1. "Qualities" and "populars" British newspapers are traditionally classified into "qualities" and "populars". Newspapers like The Times or The Guardian are said to be quality papers because they observe high standards of newsreporting, whereas papers like the Daily Mirror or The Sun are said to be popular papers because they appeal to a far larger readership. This dichotomy is unsatisfactory because it combines a term making a value judgment with a term referring to the alleged appeal of certain newspapers. This conceals the fact that the two types of papers address a different audience and therefore have quite distinct aims. What counts as "quality" in one type of paper may not be desirable as an aim for the other type of papers, and there are considerable quality differences between papers of the same category. Furthermore, the dichotomy wrongly implies that "qualities" are not, or less popular. A look at the circulation figures, however, reveals that "popularity" does not provide a reliable criterion to distinguish between newspaper categories. Today, for instance, which is commonly regarded as a popular, has a markedly smaller circulation than most quality papers. Table 3.1. Circulation of British national daily newspapers British national dailies The Times The Independent The Guardian Financial Times The Daily Telegraph Today Daily Mail Daily Express The Sun The Star Daily Mirror

circulation average Oct 87 to Mar 88 449,531 376,750 474,017 298,400 1,154,018 350,257 1,800,539 1,690,538 4,098,234 1,041,648 3,068,240

Figures of certified average net sales per issue as published by the Audit Bureau of Circulation

48

Socio-economic differentiation of the British dailies

Another dichotomy that is commonly employed to distinguish newspaper categories differentiates between "broadsheets" and "tabloids" referring to the format of the papers. Tabloids are smaller and measure some twelve by fourteen inches (30 by 40 cm), whereas broadsheets measure some fifteen by twenty-four inches (38 by 61 cm). Both dichotomies conceal the fact that the tabloid newspapers, which are traditionally regarded as "populars", fall into two quite distinct groups, if, among other things, the readership appeal of the papers is taken into consideration. Harry Henry (1983) has introduced the terms up-market, mid-market, and down-market. This terminology refers to the socio-economic classes of the people who read a particular paper. The up-market papers are exactly those that are published in broadsheet format, whereas the tabloids fall into the categories mid-market and down-market. Table 3.2. Categories of British daily newspapers broadsheet papers

tabloid papers

up-market papers The Times (T) The Independent (I) The Guardian (G) Financial Times (FT) The Daily Telegraph (DT)

mid-market papers Today (To) Daily Mail (DMa) Daily Express (DE) down-market papers The Sun (S) The Star (St) Daily Mirror (DMi)

As will become evident in the next section, and indeed throughout this study, the main division is between the broadsheets and the tabloids. I shall therefore continue to use these terms whenever the distinction between the two classes of tabloids is not at issue. Otherwise I shall be using the terms referring to the socio-economic profile of the individual papers.

3.2. Readership profiles While it is true to say that all newspapers are read by members of all social classes, it is also true, and far more obvious, that the proportions in which the different social classes are represented in the readerships of different newspapers vary to a very considerable extent.

Readership profiles

49

The National Readership Survey (NRS), which regularly publishesreader ship profiles for all national newspapers, distinguishes between the six socioeconomic classes A, B, Cl, C2, D, and E. These classes are based on occupational data, on the assumption that the occupation of the head of household (i.e. the person who either owns the accommodation or is responsible for the rent) or the chief wage earner (i.e. the senior working member of the household) gives the best indication of the informant's social standing in the community. The NRS defends this approach on several grounds (cf. Monk 1985). First, it is difficult to collect accurately the data which might be suggested as a substitute such as the income. Second, there are considerable difficulties with any attempt to combine several criteria because it is unclear how the different criteria should be weighted. The age at which somebody finished his or her education (the so-called terminal education age or TEA), for instance, gives a good indication of the social grade among the young, but it is far less reliable among the older generation. In such a case it is particularly difficult to determine how the factors occupation, terminal education age and age should be interrelated. And third, occupation provides the best general purpose grading as far as the aims of the NRS are concerned. People using social grade classifications normally have a limited number of specific interests. Thus, the marketer of consumer durables expects the grading system to be biased towards the probability of acquiring such durables: the newspaper editor expects it to be biased towards people's interest in different subject matters: and the person marketing services wishes it to be biased towards use of such services. Although such wishes are understandable, it is clearly impossible for a general social stratification system to meet all specific objectives. What might be considered surprising, however, is the degree to which a general system of grading is still, in our modern society, one of the best predictors in all these areas. ( Monk 1985: 3) Table 3.3 (see next page) gives a survey over the six classes with a very small number of examples for each class. The readership profiles as they are periodically published by the Joint Industry Committee for National Readership Surveys (JICNARS) indicate how many per cent of the readership of one particular newspaper belong to which socio-economic class. These figures must always be seen in relation to the percentages for the adult population as a whole. Thus it is estimated that only three per cent of the entire adult population belong to class A 15 per cent are class B 23 and 28 per cent are Cl and C2 respectively 18 per cent are classified as D and 14 per cent are at the lowest level of subsistence in class E. Figure 3.1 represents the readership figures in graphical form.

50

Socio-economic differentiation of the British dailies

Table 3.3. Social grading according to the National Readership Survey (JICNARS April 1985) Grade

Status

Head of household's occupation

upper middle class

Higher managerial, administrative or professional, such as bank branch manager, senior civil servant, commercial airline pilot or university professor Intermediate managerial, administrative or professional, such as bank clerk with special responsibilities, senior officer in local government, air traffic controller or university lecturer Supervisory or clerical, and junior managerial, administrative or professional, such as bank clerk, not professionally qualified social worker, driving instructor or student on grant Skilled manual worker, such as bricklayer, carpenter, bus inspector or brewer Semi-skilled and unskilled manual worker, such as fisherman, ticket collector, milk roundsman or caretaker State pensioner or widow (no other earner), casual or lowest grade worker

B

middle class

C1

lower middle class

C2

skilled working class

D

working class

those at lowest level of subsistence

Population [

τιιιιιιιιιιιιιιιιιιιιιιιικ^^

"IP

IIIIIIIIIK'gSSSggggSggg^^

§

FT lllllllllllllllllBS8SgSS8SS8&9SS8^^ llllllllllllllllira&gggggggg^^

B

To

DE ·?9g&x?ΦOδs^^^^s^^\^^^^c^^ Sun Star DMi 10

20

30

40

50

60

70

80

90

100

ΙΑ

EZ3C2

EI3D

Figure 3.1. 1987 class profile of the British national daily newspapers. For the exact figures see table A3.1 in appendix III. (Source: JICNARS).

Readership profiles

51

This figure shows that the papers clearly fall into three groups. They can be defined either in terms of their AB readership or alternatively in terms of their combined ABC1 readerships. All broadsheet papers have got an ABC1 percentage of around 80 per cent, which compares with 41 per cent for the population as a whole. The mid-market papers have got percentages of just over 50 per cent with the exception of Today, which has only 40 per cent. But even Today's figure is considerably higher than those for the down-market papers which are 25 per cent and below. This shows that there is a remarkable consistency of readership profiles within the individual groups of newspapers, and that there are marked differences between the papers of different groups. No paper follows the pattern of the population as a whole. Today has the most similar pattern, but it does not seem to benefit from this in terms of circulation. Thus it is true to say that the British press not only reflects but actually exaggerates the differences in social class and education of the nation as a whole. The differences are so obvious that statistical tests might appear to be superfluous. A one-way ANOVA (analysis of variance) for the ABC1 percentage yields an f-ratio of 138.3, p < 0.001, which means that it is extremely unlikely that the differences between the three groups can be attributed to chance.2 In terms of circulation the three groups also follow similar patterns, albeit less clearly, as has been pointed out above. The up-market papers have on average fairly small circulation figures. The down-market papers on average have got considerably higher figures with the mid-market papers lying somewhere in between. The one-way ANOVA for the circulation figures yields an f-ratio of 5.4, p = 0.032. This means that the probability for the differences in circulation to be a matter of chance is smaller than five per cent, but it is considerably higher than for the ABC1 percentage. This conforms to the intuitive perception that there are marked differences between the circulation figures in the three types of newspapers, as is expressed in the term "populars" for the tabloids, while it is also intuitively clear that the differences are not as clear-cut as those for the readership profiles as they are set out in figure 3.1. Bell (1977: 125) has noted the difficulty of equating the actual readership of a newspaper as established by a readership survey with the readership aimed at by the journalists. The journalists may have an inaccurate picture of their readership. Within an audience design framework, it must be expected that it is the journalists' perception of their audience rather than the actual au-

52

Socio-economic differentiation of the British dailies

dience which potentially influences their linguistic production. However, the distinction between the three categories of papers is so categorical that it is implausible to assume major mistakes in this respect. Journalists may well under-estimate, or possibly over-estimate the intelligence of their readerships but they will be in no doubt about the socio-economic market segment their newspaper is aimed at, particularly so since in recent years there has been very little change in the structure of the newspaper market and its segmentation, in spite of frequent newspaper articles reporting slight shifts in circulation figures. Henry (1987a) gives ample evidence that the overall picture remains remarkably stable. The average daily circulation of all national dailies and the respective share of the individual categories of papers has remained practically unchanged for the last twenty-five years. In the words of Henry (1987a: 23): Though there have been variations - up and down - both in total circulations and in the circulations of the separate newspaper categories, they are so small that a dispassionate observer might be inclined to wonder what all the fuss is about. There is a fairly strong connection between the two sets of figures given in figure 3.1 and in table 3.1, i.e. the circulation figures and the readership profiles, respectively. Figure 3.2 plots the two sets of figures against each other. This reveals quite clearly that the market segmentation is by no means continuous, but compartmentalised. 100

up-market 75 l/l L_ 01

α οι Ο QQ

mid -market 50

down-market 25

DMi

Sun

circulation average in millions Figure 3.2. Relationship between percentage of ABC1 readership and average circulation (Combination of the figures quoted in figure 3.1 and table 3.1)

Readership profiles

53

Mander (1978) notes that all the newspapers that ceased publication in the late fifties and the sixties, a time of fierce competition, particularly in the early part of this period, suffered from a lowish circulation and a low ABC 1 percentage. They would all appear in the low left-hand corner of the figure 3.2, an area which Mander (1978: 77) calls "a sort of Bermuda Triangle - the graveyard of mastheads that have disappeared under the increased costs and diminished revenues." The reason for this "Bermuda Triangle" lies in the way in which the newspapers are financed. All of them derive their revenue partly from their copy sales and partly from advertising, but in very different proportions. The upmarket papers get almost two thirds of their revenue from advertising, both display advertising and classified advertising, whereas the down-market papers get more than three quarters of their total revenue from the sales revenue. The high socio-economic status of the readers of up-market papers makes it more interesting for advertisers to place their adverts in these papers, even though it is far more expensive to advertise in the up-market papers. It is therefore of paramount importance for the up-market papers to appeal to the affluent, well-educated socio-economic classes in order to attract advertisements to their expensive pages. For down-market papers, on the other hand, it is more important to reach a maximally large readership. Circulation policy for all of the quality press is not to create as big a circulation as possible, but to maintain a sale sufficiently high, and of an appropriate nature, to command a slice of advertisement markets at an acceptable advertisement rate. (Mander 1978: 76) Figure 3.3 gives an indication of the substantial differences of revenue patterns across the three types of newspapers. For the up-market papers it cannot

Up-market Mid-market '/////////////λ KV\\\\\V

V//////A Down-market ////Λ \\\1 10

E-2 Copy Sales

20

30

pence

E3 Display advertising

122221 Classified advertising Figure 3.3. Net revenue per copy sold in 1986 (pence). Source: Henry (1987b: 23).

54

Socio-economic differentiation of the British dailies

be the major aim to increase their circulation because it broadens and thus dilutes the readership, which makes it less attractive to prospective advertisers. To simplify a little, the down-market papers sell their papers whereas the upmarket papers sell their readerships. Figure 3.3 shows again a very considerable difference between the broadsheet papers and the tabloids and a noticeable but less pronounced difference between the down-markets and the mid-markets. There is also some variation in the readership profiles of the individual newspapers if age and sex of the readers are taken into consideration but figure 3.4 and table 3.4 below show that the variation is not related to the newspaper categories. Population NIIIIIIIIIIIIINIIINIIIK'QOSOOC^ T iiiiiiiiiiiiiiiiiiiiiBsssKg^^ I

•miimnnmiiBsggaggg^^ FT DT To DMa miiiimiiiiimiRsassssgE^^ DE Sun •MiiiiiiiiiiiiiiiiiiiiB^^ Star DMi 10 20 30 40 0

-24

25-34

^245-54

CH55-64

50

60

70

80

90

100

Figure 3.4. 1987 age profiles of the British daily newspapers. (Source: JICNARS) For exact figures see appendix III, table A3. 2.

The Daily Mirror has got an age profile that is quite close to the profile of the population as a whole. In the other two down-market papers, the two youngest age groups are clearly over-represented and the two oldest age groups underrepresented, whereas the age groups from thirty-five to fifty-four are represented to the same extent as they are in the population as a whole. In the age profiles of the Daily Mail and the Daily Express, the older age groups are slightly over-represented, whereas Today has more than half of its readers in the two youngest age groups. This can probably be explained by the fact that Today is a very young paper, launched in 1986, and therefore does not have a long established readership.

Readership profiles

55

Among the up-market papers, there is again considerable variation. The Financial Times, which is clearly aimed at a specialised readership, has seventy per cent of its readers in the age groups from twenty-five to fifty-four, which make up only forty-eight per cent of the entire population. These three age groups are clearly over-represented whereas the other three are underrepresented. The Daily Telegraph is also over-represented in the middle age groups, albeit only slightly so, and only the youngest age group is clearly under-represented. The Times and The Guardian both have higher percentages in the two youngest age groups than the population as a whole and for both papers only the age group of over sixty-fives is clearly under-represented. These differences in the age profiles of the individual papers, however, are not systematic across the three classes of papers. A one-way ANOVA applied to the percentage figure of the under 34-year-old readers shows the variation to be not significant. The same can be said for the proportion of male and female readers for the individual papers. All papers have got more male readers than female ones even though in the population as a whole there are more women. Table 3.4. Sex profile of readership, 1987 British national dailies Population up-market papers The Times The Guardian The Independent Financial Times The Daily Telegraph

% men 48 63 58 67 73 54

mid-market papers Today Daily Mail Daily Express

62 51 52

down-market papers The Sun The Star Daily Mirror

53 59 55

Source: JICNARS

The mid-market papers, with the exception of Today, follow the pattern of the population more closely than the other papers. The percentage of male readers is slightly higher for the down-market papers and again very divergent for the

56

Socio-economic differentiation of the British dailies

up-market papers. The seventy-four per cent male readership of the Financial Times mirrors its more specialised appeal and the fact that the professional people who this paper is aimed at are still predominantly male. The Times and The Guardian again have fairly similar patterns, both with a high proportion of male readers, whereas The Daily Telegraph follows its own pattern with a somewhat lower percentage.

3.3. Differences in appearance and coverage There are further obvious differences between the mid-market and the downmarket papers which suggest that the two types of papers should not be lumped together under one heading. As it has been pointed out above, the mid-market papers offer more sections than the down-market ones by including a regular foreign news sections, which none of the down-market papers does, and a regular business and finance section, which only The Sun does on a regular basis. The down-market papers, in contrast to the mid-markets, furthermore publish pictures of sparingly dressed women on page 3 of all their issues. None of the mid-markets does this. The consumer magazine Which! (Aug 1988) published a survey over all British national daily newspaper pointing out which paper or papers did best in particular categories. The comparison was based on the number of articles published in a category rather than the quality of the articles themselves. A distinction was made between news articles and features for every category. The Sun did best in news about TV, in "celebrity doings" and it had the highest amount of prizes offered to its readers. The Star had the most sex stories, and the Daily Mirror was strongest in human interest news, in disasters (more than 30 disaster stories in ten days) and in royal stories, and it had the most special offers and vouchers for its readers. The Daily Mail was strong on news on religion and on human interest features, the Daily Express was good at consumer information and at features on the royalty, whereas Today had more media news, more property features and more features on "celebrity doings" than any other paper. The Daily Telegraph was best for news in the categories law and order, employment, property and consumer information, and it covered a higher number of different sports than any other paper. The Times was strongest for news in the categories politics and science, it had the most features on law and order and it had more pages on sports than any other paper. The Independent published the highest number of obituaries and it was best on news on defence and features on politics. The Guardian was strongest on education, and it had the most news about the environment and about health and social welfare, it

Differences in appearance and coverage

57

had the most features in the category arts, TV and radio, and it had the best coverage of foreign news. It also published more letters to the editor than any other paper. The Financial Times, not surprisingly, was best in the category finance, business and economy. It also had the most features on industry and trade unions. Which? also established that it was the Financial Times that had the most printing errors. The Guardian, which came second, once even managed to spell its own name wrong. Today had the fewest printing errors. The Daily Telegraph was more often first with the news, whereas The Guardian and The Independent were most often late with the news. The results on the league table of coverage as established by Which? are indicative of the priorities assigned to different categories of news by individual papers. All the tabloid sized papers, for instance, give a very British view of the world with very few foreign news stories. The different performances in terms of the number of printing errors, on the other hand, are presumably more indicative of the production methods that are used for one particular paper. Which? carried out its investigation before the Financial Times switched to modern composition techniques. So there should be fewer printing errors now. The size of the main front page headline also varies systematically according to the type of newspaper. Its height is 10 mm in the Financial Times and about 15 mm in the other four up-market papers. In the mid-market papers it varies from about 30 to 45 mm, and in the down-markets it is generally greater than 45 mm and can be up to about 100 mm. See also Crystal and Davy (1969: 173-192) and O'Donnell and Todd (1980: 85-100) for a description of the differences between the tabloids as a whole and the broadsheets. In October 1987, all the down-markets had a cover price of 20 pence. The mid-markets cost 22 pence (with the exception of Today which cost 20 pence), and the up-markets cost 25 pence (with the exception of the Financial Times, which cost 45 pence). In the first half of 1987, the up-market papers accounted for 16.8 per cent of the entire national dailies circulation. The mid-markets accounted for 24.5 per cent and the down-markets for 58.7 per cent. The comparable figures for 1965 were 12.4 per cent, 39.8 per cent and 47.8 per cent. Thus there has been a considerable decrease of the mid-market segment from 40 per cent to 25 per cent, which constitutes a 40 per cent drop of its 1965 circulation. More than half of this loss in copies sold went to the down-market papers, which increased their market share to over nine million, an advance of 18 per cent over its 1965 figures. The up-market papers also gained from the losses of the midmarket papers. They increased their 1965 circulation as much as 30 per cent.

58

Socio-economic differentiation of the British dailies

Table 3.5. Newspaper circulations in Britain in 1965 and 1987

1965 up mid down total

2.0 6.4 7.7 16.1

1987 12.4% 39.8% 47.8%

2.6 3.8 9.1 15.5

change

16.8% 24.5% 58.7%

+ 0.6 -2.6 + 1.4 -0.6

+ 30% -40% + 18.2% - 3.7%

In millions. Source: Henry 1983: 488 chart 4 and 1987a: 23 chart 1.

3.4. Conclusion The British national daily newspapers can be split up into three groups, the up-markets, the mid-markets and the down-markets, on the basis of their socio-economic readership profiles. More than two thirds of the readership of the down-market papers are members of the working class (C2DE). Well over half the readership of the mid-market papers are members of the lower middle class and the skilled working class (C1C2). And 50 per cent or more of the readerships of the up-market papers are members of the middle middle class and the upper middle class (AB). But it must again be pointed out that all the papers are read by members of all the social classes. What differs are the percentages with which particular classes are represented in the readership of individual papers. It is, however, reasonable to assume that the editors of the individual papers will strive, in content and style, to cater for the majority of their readers in order to prevent them from switching to another paper. The papers also vary considerably in the way they are financed. The upmarkets, which draw the greatest part of their revenue from selling advertising space, have to attract a readership that is as exclusively up-market as possible, whereas the down-markets, which earn their money mainly through the copy sales, try to appeal to as large a readership as possible. This explains why the segmentation of the British national dailies does not just mirror but exaggerate the existing differences in social class and education of the entire population. There are clear non-linguistic differences in the appearances of the papers and in the priorities they assign to particular categories of news, which reflect the interests of the targeted readerships, or, presumably, what the editors and journalists perceive to be the interests of their targeted readerships. It is only natural to expect that these considerable differences across the three types of papers will be reflected in the language used by the individual papers.

4. Structure of noun phrases

4.1. Introduction It is the aim of this study to examine syntactic variation within the noun phrase across different styles of written English, or more specifically across different styles of newspaper English. In order to be able to show where variation can occur within the noun phrase, it is necessary to spell out in detail its internal structure. In spite of the fact that the verb phrase has always enjoyed greater popularity with linguists, there is no dearth of studies on the English noun phrase. For generative grammarians, the noun phrase received some attention within the X-bar framework. Jackendoff (1977), for instance, subjects the noun phrase to a close analysis in order to show its similarity with all the other major phrasal categories. The noun phrase also received attention from generative grammarians in discussions about the origin of nominalisations (cf. Chomsky 1970). The generative semanticists, on the one hand, argued that nominalisations were the result of transformations, whereas the interpretive semanticists argued for nominalisations to be given separate entries from their verbal equivalents in the lexicon (cf. Newmeyer 1986: 106-138 for a short historical account of the controversy). However, I shall be using a more traditional approach, and for ease of reference I shall take Quirk et al.'s (1985) terminology as a starting point. In several instances I will have occasion to point out what I perceive as weaknesses or shortcomings in their classifications, but in many other respects I cannot even attempt to give as detailed a presentation as they do. They devote four chapters (5, 6, 7 and 17) and over 350 pages to the noun phrase and its constituents. Within the scope of the present study I must perforce be selective and concentrate on those issues that are central to the variability of the structure and complexity of noun phrases. The italicised parts in the following sentences all constitute a noun phrase. (1) (2) (3) (4)

Gustav Husak stepped down as party chief. The old man stepped down as party chief. He stepped down as party chief. Gustav Husak, 74, who restored Czechoslovakia to Communist orthodoxy after the brief enlightenment of the 1968 Prague Spring, stepped down as party chief. (DT 18.12.1987, 1.4)

60

Structure of noun phrases

Noun phrases can be realised as nouns, names or pronouns. Nouns and names - but not normally pronouns - can be modified in various ways. The head of the noun phrase is preceded by two slots and followed by one slot, all of which can but need not be filled with linguistic material. It is preceded by the determiner and the premodifier slot and followed by the postmodifier slot. These slots can all be further broken down. The determiner consists of a predeterminer, a central determiner and a postdeterminer, whereas the premodifier consists of a precentral, a central, a postcentral and a prehead. Again, some of these slots may be realised in individual examples while others may remain empty. The constituent analysis of noun phrases relies to a great extent on paradigmatic relationships. There are no obvious candidates for free variation within the noun phrase. In generative terms, a transformation with a claim to be meaning preserving would be the most likely candidate for free syntactic variation. Romaine (1982), for instance, discusses her data of Middle Scots relativisers explicitly in these terms. However, other researchers have pointed out that it is unlikely that there are many transformations that operate within the boundaries of the noun phrase, as for instance Selkirk (1977: 286) who claims that the investigation of noun phrase structure - the determination of the constituents that compose the noun phrase - is not without its difficulties, for few transformations, those invaluable tools of constituent structure analysis, ever apply within this domain.

4.2. Premodification The heads of noun phrases can be modified by a wide range of elements that either precede or follow it. I shall turn first to the premodifying elements. Most commonly they are adjectives, participles or nouns. These elements can be very simple or very complex. As Huddleston (1984: 256f) points out: Strictly speaking, the modifiers should not be analysed (immediately) as words but as phrases, for although they are most frequently single words, there is in general the potential for them to have their own dependents. In the phrase a supersonic aircraft captain, the noun captain is modified by another noun aircraft, which is in turn modified by the adjective supersonic. This gives structure (5) rather than (6): (5) (6)

the [[supersonic aircraft] captain] the [supersonic [aircraft captain]]

Premodification

61

The structure in (6) would suggest that the referent is a captain of an aircraft and that he or she is supersonic, which is clearly nonsensical. Thus the head of the noun phrase is modified by one single phrase rather than by two individual words. The category of this phrase, however, is controversial. In spite of being a higher level category than a single word, it cannot be a fully fledged noun phrase because it does not allow its own determiner. (7)

*the [[a supersonic aircraft] captain]

I shall not postulate a new name for a category of this level. It suffices to realise that it must be assigned "to a distinct phrase category intermediate in the constituent hierarchy between noun and NP" ( Huddleston 1984: 258). Many attempts have been made to systematise and categorise all the possible semantic relationships holding between nouns and their premodifiers (e.g. Ljung 1970; Levi 1978; Warren 1978, 1984; Aarts and Calbert 1979; and Warren (1984: 7-17), who gives a useful summary of work in this area). Aarts and Calbert (1979), for instance, distinguish 13 different types of relations such as experience (weary reader), cause (infuriating behaviour) or manner (fast train). However, I shall be concerned with such attempts only to the extent that they shed light on the possible ordering of the elements in noun phrases with multiple modification. There have also been many attempts to make generalisations about the permissible or the preferred orderings of premodifiers in noun phrases with multiple premodification (e.g. Goyvaerts 1968 (with a review of earlier work), Fries 1970; Martin and Ferb 1973; Abberton 1977; Coates 1977; Bache 1978; Dixon 1982; Coulter 1983; and Warren 1984). Such attempts, of course, depend on some categorisation of the premodifying items. Huddleston (1984: 259), for instance, notes that "modifiers denoting 'gradable' or 'scalar' properties precede those denoting 'non-gradable' or 'categorial' properties". Goyvaerts (1968) categorises in a similar fashion, although his categories do not offer hard and fast criteria. He postulates (1968:18) the principle of "the broader, more comprehensive, more general, commoner, before the more specific, more particularising, more detailed, less common". Coates (1977: 15), realising how difficult any succinct categorisation turns out to be, uses syntactic criteria: "Whatever criteria are used, one clear generalisation emerges: nominal modifiers (i.e. nouns and denominal adjectives) occur close to the head, while adjectival modifiers occur further away." (her emphasis). Halliday (1985: 163) calls premodifying items epithets or classifiers. Epithets are either experiential if they indicate some quality of a subset of possible referents denoted by the head of the noun phrase more or less objectively, such as the adjectives blue, fast, long; or else they are attitudinal if they ex-

62

Structure of noun phrases

press the speaker's subjective attitude towards the referent, such as the adjectives splendid, silly, fantastic. He notes that attitudinal epithets tend to precede experiential ones. Classifiers, on the other hand, indicate a particular subclass of the thing in question: electric train, passenger train, or wooden train. This distinction correlates with some syntactic differences. Only epithets can be moved to the front of measure expressions in partitive constructions whereas classifiers cannot be moved. It is possible to say both a cup of strong tea and a strong cup of tea even though the adjective strong, which is an epithet, clearly modifies tea and not cup. In contrast, we can only say a slice of brown bread but not a brown slice of bread, because here the adjective is a classifier, which indicates what type of bread we are talking about (Halliday 1985: 174). The epithets that are possible in front of the measure expressions must be connected to their quantifying function, as in a large cup of tea. Dixon (1982: 15-34) offers a very sensitive treatment of the ordering of premodifiers which is firmly based on semantic features of the different adjectives. He distinguishes seven types of adjectives, that is value (good, bad, excellent); dimension (big, long, thick); physical property (hard, heavy, rough, hot); speed (fast, quick, slow); human propensity (jealous, happy, kind); age (new, young, old); and colour (black, white, red). If there are more than one adjectives in one noun phrase, they will, "in semantically unmarked cases" (1982: 25), appear in the order suggested by this list. However, Dixon restricts his list to the most typical adjectives saying very little about premodifying nouns as in oatmeal dog food. And he does not consider adjectives that can function both as a classifier (the fast lane of the Mil) or as an epithet (an extremely fast bicycle race). Quirk et al. (1985: 437, 1337-1342) distinguish four premodifier slots, which can be filled by specific types of premodifier. The four slots, or premodifier positions, are called precentral, central, postcentral and prehead. In contrast to the categorisations mentioned so far, this approach is based on syntactic tests. They claim that the four positions or zones correlate largely with semantic classes (1985: 1338), but the correlation is indirect and not always very obvious. The central position is reserved for the central adjectives. They belong most typically to their word category because they satisfy all the relevant tests. They can occur both predicatively and attributively, they can be intensified by very and they are gradable in that they have comparative and superlative forms. (8)

a big event - the event was big - an exceptionally big event - the biggest event

Premodification

(9) (10)

63

an interesting event - the event was interesting - an exceptionally interesting event - the most interesting event a peaceful event - the event was peaceful - an exceptionally peaceful event - the most peaceful event

These examples illustrate three subtypes of central adjectives. Adjectives can be underived such as big in (8). Others would be great, new, hard, cold. The adjective interesting in (9) is deverbal because it is based on the verb "to interest", and peaceful in (10) is denominal because it is based on the noun "peace". Quirk et al. (1985: 1338) claim that there is a tendency for central premodifiers to occur in the order nonderived - deverbal - denominal if two or more central premodifiers of different subtypes are used. Thus the examples under (11) are perhaps somewhat more natural than those under (12) (11) (12)

a big interesting event - an interesting peaceful event - a big and peaceful event - a big, interesting and peaceful event an interesting big event - a peaceful interesting event - a peaceful and big event - a peaceful, interesting and big event

The central position is preceded by the precentral, which is reserved for peripheral, nongradable adjectives that are most typically intensifiers such as entire, certain, and complete. These are peripheral adjectives because they satisfy only some of the relevant tests. They cannot be used predicatively but only attributively, they can usually not be intensified by very, and they do not have comparative and superlative forms. (13) (14) (15) (16)

the entire event - sheer enjoyment - a mere bungalow *the event was entire - *the enjoyment was sheer - *the bungalow was mere *the very entire event - *a very sheer enjoyment - *a very mere bungalow *the entirest event - ?sheerest enjoyment - ?a merest bungalow

Again a subclassification is possible. The intensifiers can be emphasisers such as certain, definite, plain, pure, sheer; they can be amplifiers such as absolute, entire, extreme, perfect, total; or they can be downtowners such as feeble or slight (Quirk et al. 1985: 1338). However, there are several idiosyncrasies among them. Some of them are possible in predicative position (the silence was total, the heat was extreme, the damage was slight), they may allow being intensified themselves or they may take comparatives and superlatives (his very feeble attempt, under most extreme pressure, moments of purest joy, the most certain failure}. Thus they defy any far-reaching generalisation, but they have in common that they always precede other premodifiers.

64

Structure of noun phrases

The postcentral position contains participles and colour adjectives, such as retired, sleeping, red, and pink. These elements are again not central adjectives in that they do not satisfy all the relevant tests. They can be used both predicatively and attributively. They are usually not gradable and cannot be intensified. (17) (18)

an established fact - the fact is established - a pink bicycle - the bicycle is pink *a very established fact - *the most established fact - ?a very pink bicycle - ?the pinkest bicycle

In the right context basic colour adjectives such as red, pink, green seem to be both gradable and intensifiable but this is not the case for their hyponyms such as scarlet, mercury, zinnobar or jade. Hence (19) is possible but not (20). (19) (20)

This is the pinkest bicycle I have ever seen. *This is the most mercury car I have ever seen.

The prehead position finally is the position nearest to the head of the noun phrase itself and contains the "least adjectival and the most nominal items" (Quirk et al. 1985: 437). Among these there are denominal adjectives such as nationality adjectives like Scottish or Chinese, and "denominal adjectives with the meaning 'consisting of, 'involving', 'relating to', e.g. experimental, statistical, political, statutory" (1985: 437). This comes closest to a semantic definition, but again there are syntactic constraints that hold for this position or zone. The elements in this zone are not gradable, and they cannot be intensified. (21) (22)

privatisation plans - French cheese - the annual meeting *the plans are privatisation - *very privatisation plans - *very French cheese - but the cheese is French - *a very annual meeting - but the meeting is annual

The examples in (22) show that premodifying nouns cannot be used predicatively and that they cannot be graded or intensified. The denominal adjectives in this position also cannot be graded or intensified, but they do allow predicative position. The nouns in this position may form quasi-compounds. Very often it is difficult to decide whether a given combination of two nouns should be regarded as a compound or a head of the noun phrase with a premodifying noun in prehead position. I will say more about compounds below. The elements in this position have received particular attention from seve al linguists who use various names for them such as transpositional adjective

Premodification

65

(Marchand 1966), first element of an N + äff N compound (Ljung 1970), denominal adjective (Coates 1977), nonpredicating adjective (Levi 1978), or first element of a linking construction (Aarts and Calbert 1979). The terms are not always exactly coextensive, but they all refer to premodifiers that are not central adjectives and occur nearer to the head. The additional criteria depend on the point of view of the individual researcher some indication of which is given by the technical terms that are used for them. To summarise, the four positions or zones which contain premodifiers in noun phrases can be distinguished on the basis of four simple syntactic tests and on the basis of their respective position. Those elements that can be used not only attributively but also predicatively and that can moreover be intensified, e.g. by very, and that have comparative and superlative forms are central adjectives and appear in central position. Elements that function as modifiers but do not satisfy all these tests and are positioned to the left of central adjectives (either actually or potentially) are precentral adjectives. The remaining premodifiers have to be classified either as postcentral or as prehead. The difference between the two is largely morphological and semantic. Deverbal adjectives and colour adjectives are in postcentral position, whereas denominal adjectives and nouns are in prehead position. Quirk et al. generalise their observations about the ordering of premodifiers in a way that is very similar to Halliday's conclusion quoted above. They suggest that modifiers relating to properties which are (relatively) inherent in the head of the noun phrase, visually observable, and objectively recognizable or assessible, will tend to be placed nearer to the head and be preceded by modifiers concerned with what is relatively a matter of opinion, imposed on the head by the observer, not visually observed, and only subjectively assessible. (Quirk etal 1985: 1341) Halliday's observation that attitudinal epithets precede experiential ones, and that classifiers follow the epithets and immediately precede the head of the noun phrase accords well with the suggestion that relatively inherent and objective qualities tend to be preceded by subjective and noninherent modifications. However, it is important to notice that this is not a rule or a constraint, it is merely a tendency. On the basis of these observations it seems plausible to expect some systematic variation in the way in which the various newspapers use premodifiers. Intuitively one might expect that up-market papers use more so-called "objective" classifiers in prehead position and possibly fewer intensifiers in precentral position.

66

Structure of noun phrases

There is some evidence in the literature that this is indeed the case. Levi (1978: 226), for instance, suggests that there is a positive correlation between the formality of speech and the frequency with which nouns in premodifying position are replaced by denominal adjectives: The higher the stylistic or technical level of the discourse, and the higher the educational level of the speaker (and the addressee), the more likely that Morph Adj [i.e. Morphological Adjectivalization] will have been applied to produce a nominal adjective. For example, medical personnel would be more likely than laymen to speak among themselves of renal disease rather than kidney disease, just as engineers or geologists would be more likely than nonspecialists to use fluvial currents in place of river currents. On the basis of this quote one might rather expect more premodifying adjectives than nouns in the up-market papers, whereas Quirk et al.'s and Halliday's observations lead one to expect the contrary. On closer inspection, however, this is not necessarily so because the adjectives given by Levi and ascribed by her to a more formal level than the corresponding nouns are still classifying. Thus we would have to expect more classifiers in the up-markets than in the mid-markets or down-markets, but at the same time a higher proportion of adjectival classifiers. This makes clear that a comparison of premodifying nouns versus premodifying adjectives would not be sufficient unless the semantic function is taken into consideration. The empirical data which supports this hypothesis for my corpus of newspaper language will be given in chapter 7 below. For the moment it suffices to illustrate the regularity of the above-mentioned ordering tendencies with some illustrations picked out of my corpus at random. Table 4.1. Examples of multiple premodification del

(23) some (24) a (25) (26) the (27) no

precentral

central

postcentral

very

odd hard, glossy

dehydrated converted

great intense

modern personal

prehead

head

Teflon-like Soviet transport media

butter surface planes moguls vision

postmod

of life

References: (23): DMa 18-12-87, 9.5; (24): DMa 18-12-87, 9.5; (25): I 6-10-87, 8.1; (26): FT 22-1-88, 18.1; (27): DT 18-12-87, 10.5.

Premodification

67

The examples in table 4.1 all correspond to the ordering rules as set out above. Sentence (28), given below, contains a rather more complex, and presumably somewhat extreme example of noun modification. (28)

Sotheby's is selling off Elton's effects in Ά four-day, 2000-lot, you'veseen-the-catalogue, now-wear-the-T-shirt sale that kicks off the autumn saleroom season. (Sunday Times Magazine 7-8-88, 36.2)

There are four premodifiers, two of which consist of a numeral attached to a noun, while the other two are sentential in that they consist of an entire clause of their own with subject and predicate. The individual constituents of these four premodifiers are joined together by hyphens. All of them have to be assigned to prehead position. None of them can be regarded as an intensifiers, none of them could be graded, they are certainly not colour adjectives or nongradable participles, and therefore they must be in prehead position. They must be seen as classifiers, which do not just describe the head of the noun phrase but actually restrict its possible referents to a small subgroup. Premodifying elements in prehead position are often used as qualifiers, which means that they restrict the reference of the head of the noun phrase to a subset of the things it denotes. In many cases the resultant expression is fairly permanent, and is used regularly. Eventually the meaning of the combined expression may differ from the meaning that is derivable from the meaning of its constituents. In this case the term compound or nominal compound is often used. However, in spite of a great deal of attention devoted to this construction, there is little agreement on the exact definition of nominal compounds (cf. Bauer 1978; Dierickx 1970; Jankovic 1986; Johansson 1980; Leonard 1984; Osselton and Osselton-Bleeker 1962; Pennanen 1980; Sampson 1980; and Warren 1978). (29) (30) (31) (32) (33)

lighthouse - light music software - soft option hothouse - hot house blackbird - black bird darkroom - dark room

The first element in these examples is always the compound which is contrasted to the second element that is not usually regarded as a compound. Compounds tend to have a primary stress on the first element, whereas noun phrase combinations tend to have it on the head of the phrase. In writing this is often, but by no means invariably, reflected in the orthography. Compounds are spelled in one word or they are hyphenated, but noun phrase combinations are

68

Structure of noun phrases

written in two words. Quirk et al. (1985: 1569) note that despite the idiosyncratic spelling conventions, there is in general "a progression from open to solid as a given compound becomes established, and hence widely recognised and accepted as a 'permanent' lexical item". The criterion of primary stress is notoriously unreliable (see in particular Levi 1978; Pennanen 1980; and Sampson 1980). In deciding whether to treat combinations such as those in (29) to (33) as single word compounds or as premodified noun phrases, I accept spelling as a guide because I want to refrain from imposing my intuition on the data under analysis. To summarise this section on premodifiers, the most striking generalisation that can be made across all the elements that can be present in a noun phrase to the left of the head of the noun phrase, recalls Halliday's quotation given above. The noun phrase starts with the most deictic elements, the determiner, and more specifically with the predeterminers like all and both and the central determiners, i.e. the articles as such. The most central purpose of these elements is to help the addressee to pick out the correct referent either from the linguistic or the situational context. The cardinal and ordinal numbers are also important elements to establish the correct referent, by giving the exact number or the position in a sequence of elements. The next elements are first the more evaluative and then the more descriptive adjectives. Finally, in prehead position, there are the classifiers that restrict the set of referents of the head of the noun phrase. Not surprisingly, perhaps, it is this element that can be merged with the head of the noun phrase to form a compound, which by degrees may be regarded as a lexicalised item in its own right.

4.3. Postmodification It has often been noted that posthead modifiers are more explicit than prehead modifiers (e.g. Solinger 1967; Huddleston 1984: 267; and Quirk et al 1985: 1243f). There are basically three types of postmodifiers that vary considerably in their own rights as to their explicitness. The finite relative clause is the most explicit of all the modifiers. The infinitive clauses and the -ing or -ed participle clauses do not contain a finite tense element and are accordingly less explicit. Postmodification by a prepositional phrase is still less explicit as there is no verb indicating the argument structure of the modifier.

Postmodification

69

4.3.1. Relative clauses As mentioned above, relative clauses are the most explicit type of noun modifier. The head of the noun phrase is embedded within a subordinate argument structure with clear indications as to the function of the head of the noun phrase within this structure and with a clear encoding of the tense which allows the addressee to situate the proposition of the relative clause within the nonlinguistic context. (For detailed studies of relative clauses see in particular Hartvigson 1969; Ryden 1970, 1974; Aissen 1972; Taglicht 1972, 1977; Michiels 1975, 1977; Pagnoux 1976; Bresnan and Grimshaw 1978; Downing 1978; Weisler 1980; Young 1980; Olofsson 1981; Romaine 1982, 1984; Lehmann 1984; Andrews 1986) The distinction between restrictive and nonrestrictive modification is particularly pertinent for relative clauses because the difference is grammaticalised for this modifier. (34) (35) (36)

A promotions boss who fiddled a raffle in front of a 23,000 soccer crowd to win his sister a car was jailed yesterday. (DMa 18-12-87, 15.5) Teachers who go on strike will pay dearly for deserting their classrooms. (St 19-12-87, 2.3) The ban followed the wild scenes that erupted at Wembley two weeks ago after Frankham was stopped in the first round by Billy Sims of Hackney. (St 17-12-87, 32.5)

In sentences (34) to (36), the headnouns are modified by relative clauses given in italics. The modifier offers a way of adding a complete proposition about the referent. The relative pronoun encodes the rudiments of a gender distinction, in that who is restricted to human, or humanised, referents. In all these examples, the relative pronoun, referring to the same referent as the noun it modifies, functions as subject in the relative clause. In sentence (37), on the other hand, there is a zero relative pronoun, which functions as direct object of the subordinate clause. (37)

after all the criticism the Met men received after the October hurricane. (To 18-12-87, 5.1)

In all the examples (34) to (37) the relative clause is clearly restrictive in that the addressee would not be able to identify the intended referent without the information given therein. In (34) and (35) the noun phrase is indefinite, and the modifier restricts the set of possible referents to just one individual or to a specific subset. In (36) and (37), the noun phrase is definite. The addressee is taken to be able to pick out the correct referent on the basis of the identification given within the noun phrase itself.

70

Structure of noun phrases

In the next two examples the addressee does not depend on the relative clause in order to pick out the correct referent. The relative clause in both cases modifies a proper noun. (38) (39)

The editor of The Hindu, Mr N. Ram, who had flown specially from Madras, urged the president to set Mr Jeyaraj free. (G 31-10-87, 9.1) All the honours went to Southend who performed with great courage and no little skill and who could possibly have won the tie by a wider margin had they chosen so to do. (T 8-10-87, 42.2)

In the excerpt given under (38), the relative clause is bracketed by commas. The name of the individual is given in apposition to his role which conferred some newsworthiness on him. The entire apposition thus identifies the intended referent unambiguously. The relative clause simply adds some further information which may be essential for a clear understanding of the article of which this excerpt forms a part, but it is not needed for identifying the referent of the head of the noun phrase of its superordinate phrase. The same is true for example (39). In the context of the article, it is perfectly clear that the proper noun Southend refers to the football team of that name. Both relative clauses must therefore be interpreted to be nonrestrictive, even if the author or the responsible editor failed to separate them off by commas. The distinction between restrictive and nonrestrictive is valuable, but we should be prepared to view it as a gradient rather than as a dichotomy between two homogeneous categories. (Quirk etal. 1985: 1257) Relative clauses have long proved a favourite area for researchers interested in syntactic variation because they offer the most convincing example of free variation within the noun phrase. The choice of one particular relative pronoun in favour of another very often does not affect prepositional meaning, and therefore it can be assigned to some other factors - linguistic or nonlinguistic - that influence the choice. The most elaborate study of the variation of the relative pronoun is Romaine (1982).

4.3.2. Appositive clauses Appositive clauses bear some superficial resemblances to relative clauses. They can be restrictive or nonrestrictive, but unlike relative clauses they are always introduced by the same particle, i.e. that. This particle, again unlike relative clauses, does not have the function of subject or object in the clause it introduces.

Postmodification

(40)

(41)

(42)

71

Mikhail Gorbachev's new wave rolled on yesterday with the promise that Russia would stop pushing around her Iron Curtain neighbours. (DE 5-11-87, 4.2) the Government received warnings last week - while the siege at Peterhead Prison was going on - that Perth could be the next flash point. (16-10-87, 1.3) Another theory, that the fire had started in rubbish under the escalator, was also challenged when ... (T 20-11-87, 2.6)

Examples (40) and (41) are both restrictive appositive clauses. They would be not just semantically but also syntactically incomplete without the appositive clause. Example (42), on the other hand, shows a nonrestrictive appositive clause, which could be left out without rendering the sentence ungrammatical, even though it is semantically of vital importance. Without it, the reader would have no idea what kind of theory was meant. Example (41) is perhaps slightly unusual because the appositive clause is postposed to the end of the superordinate clause. There are two adverbials in apposition, one phrasal, the other clausal, that intervene between the modified head and its appositive clause.

4.3.3. Nonfinite clauses Nonfmite clauses as postmodifiers are less explicit than relative clauses because they do not encode the tense directly. They may include a participle, either present or a past, or an infinitive. The tense relations have to be inferred from the context. In the examples (43) to (47) nouns are postmodified by a -ed participle clause. (43) (44) (45)

(46)

(47)

Perth jail, built in 1859, is Scotland's oldest prison. (I 6-10-87, 1.3) He eventually received 4,800 from his six applications lodged under variations of his name (I 6-10-87, 1.5) Police were led to the spot by a man already charged with abducting and threatening to kill an 18-year-old slaughterhouse worker (To 1-288, 1.1) Kendall had been sprung from Gartree with convicted murderer Sydney Draper in an airborne operation masterminded by an East London godfather (To 11-2-88, 5.5) Analysts believe that these factors, underpinned by speculation that Northern Telecom might decide to increase its stake, which now stands

72

Structure of noun phrases

at 27.8 per cent, could soon drive the price of STC's shares above 330p, and even over 350p. (T 8-10-87, 25.6) In all these cases, the head of the modified noun phrase is understood to be the passive subject of the nonfmite clause. Examples (43) to (45) require a reading in which the time of the event reported in the modifying clause is anterior to that of the clause that contains the modified noun phrase. Thus, Perth jail was built a long time before it was the oldest prison of Scotland the applications were lodged before he received the British Telecom shares for which these applications had been made and the man had been charged before he led the police to that particular spot. In the examples (46) and (47), on the other hand, the actions in the containing clause and in the modifying clause are probably best understood as being simultaneous. An East London godfather was masterminding the prison escape at the time it actually took place, but there is presumably also a considerable amount of preparation involved which would have taken place anterior to the actual escape. Thus, both (4a) and (4b) would be possible renderings of the modifying clause in a more explicit relative clause. (46)

a. an airborne operation which had been masterminded by an East London godfather b. an airborne operation which was being masterminded by an East London godfather

In example (47) above, the speculation underpins the analysts' belief at the same time that they hold these beliefs. In this instance the simultaneous interpretation seems to be the only one possible. In contrast to the -ed participle clauses, in which the modified noun is understood to be the passive subject, in -ing participle clauses it is understood to be the active subject. (48) (49)

doctors switched off the machine keeping baby Alexander Davies alive (DMa 18-12-87, 2.1) Seven other people, including two firemen and a transport policeman, are still in hospital. (T 8-10-87, 1.2)

The temporal relation between the modifying clause and the clause containing it is again left open to contextual interpretation. In (48) the machine kept the baby alive until the moment when the doctors switched it off. Thus what is described in the modifying clause is anterior to the action described in the containing clause. In the example (49), the two clauses describe simultaneous states of affair. The people include those mentioned precisely at the time when they are in hospital.

Postmodification

73

Postmodifying infinitive clauses are less explicit than either type of participle clause because they do not encode the function which the head of the noun phrase fulfils within its argument structure. There are two basic types of postmodifying infinitive clauses. One type is similar to the -ed and -ing participle clauses in that the head of the noun phrase is situationally understood to be an argument of the modifying clause. In contrast to the -ed and -ing participle clauses, the head of the noun phrase can be understood as not just one given argument, but can function as either subject, object, adverbial or even, in some cases, complement. (50) (51) (52) (53)

... which will be removed as part of the new arms deal to be signed in Washington (DT 27-1-87, 1.3) All have covers which shout of a hundred and one things to do. (DE 511-87, 15.1) the UK Government had a fortnight to supply an extradition warrant (DE 20-11-87, 13.3) Bush joined President Reagan in fully supporting a "risky operation" to win the freedom of American hostages in Lebanon. (DMa 18-12-87, 10.1)

In the examples (50) to (53), the modified nouns are understood as one of the arguments of the infinitive clauses modifying them. In (50), the head of the modified noun phrase is understood to be the subject of the passive infinitive clause modifying it. This can be most clearly shown by contrasting it to the corresponding relative clause, which will be signed in Washington, in which the relative pronoun that is coreferential with the head of the noun phrase deal acts as a passive or surface subject. In example (52), the noun, things, can only be understood as the direct object in the postmodifying infinitive clause. The subject is left unspecified and must be recovered from the context, even though it is impossible to identify unambiguously the intended referent. The article from which excerpt (52) is taken reports about new glossy women's magazines. Thus the intended subject could be women in general, the readers of those particular magazines, or just whoever wants to do the things that are displayed on the covers of these magazines. In (52), the head of the modified noun phrase must be understood as an adverbial within the argument structure of the postmodifying infinitive clause. In this case, the intended referent of the empty subject is plain, it is identical with the subject of the superordinate clause. Sentence (53) is less straightforward. The head of the modified noun phrase can be understood to be the subject of the modifying clause, but it can also be

74

Structure of noun phrases

taken as an adverbial. These two interpretations can be demonstrated by the corresponding constructions with a relative clause, which makes the argument structure more explicit. (53)

a. Bush joined President Reagan in fully supporting a "risky operation" which will/should/might win the freedom of American hostages in Lebanon b. Bush joined President Reagan in fully supporting a "risky operation" by/through which they might/hope to/intend to win the freedom of American hostages in Lebanon

In (53a) the relative pronoun, standing for the noun, operation, is the subject of the relative clause whereas in (53b) it is governed by a preposition and is an adverbial. These two renderings of the infinitive clause in the form of a relative clause also reveal how much more explicit the latter is. Several possible readings are indicated both in (53a) and in (53b) but even more readings may be possible. With a relative clause, the author must be explicit and indicate not only the intended argument structure but also the intended tense and modality. In the infinitive clause, these aspects can be left open to interpretation. The second type of postmodifying infinitive clause is called appositive. It does not include the head of the noun phrase in its argument structure. This is equivalent to saying that it does not have a corresponding relative clause construction. Sentences (54) to (56) are typical examples. (54)

(55) (56)

Mr Gorbachev's decision to call on the Prime Minister on the eve of the historic Washington summit is seen in Whitehall as ... (DT 27-1187, 1.2) Few would take orders to buy and sell shares for less than a minimum investment of £ 50,000. (To 18-12-87, 2,6) Associated British Foods, which acquired 15 p.c. of Dee in the £686m deal which contained a provision limiting its freedom to sell before June 1988. (DT 18-12-87, 17.4)

In (54) and (55) the head of the noun phrase with the modification given in italics is a deverbal noun that takes an infinitive clause complement. The head of the noun phrase in (56) is rather deadjectival, but it also takes an infinitive clause complement. In all these cases the head of the noun phrase is clearly not part of the argument structure of the modifying clause. The relation to corresponding verbal or adjectival constructions is not always easy to make. (54) would give the fairly uncontroversial Mr Gorbachev decided to call on the Prime Minister, but (55) is less clear. Perhaps Few would accept being or-

Postmodification

75

dered to buy and sell shares, and in (56) maybe limiting its being free to sell. However, it is not so much the entire constructions that are similar. No suggestion is made here that one should be derived from the other. The point is that these head nouns have got similar subcategorisation features as their verbal and adjectival counterparts. Such a relationship to verbs or adjectives taking infinitive clause complements is entirely absent in the first type of postmodifying infinitive clause discussed above.

4.3.4. Postmodification by prepositional phrases Postmodification by a prepositional phrase is less explicit than any of the above-mentioned types, and it is far more frequent than any other type. It is the entire range of prepositions, including complex prepositions such as outside of, next to or in charge of, that can be used (see Quirk et al. 1985:1274). Some examples with multiple postmodification by prepositional phrases may suffice to show the versatility of this type of modification. (57)

(58) (59)

(60) (61)

(62)

Since his arrival at Anfieldfrom Watford for a fee of £900,000, John Barnes has captured the imagination of supporters ... (G 31-10-87, 18.3) The kick in the teeth for Milton Keynes means that the Corporation will be forced to disband with its job only half done ... (G 17-12-87, 21.2) President Ronald Reagan is expected to scale down sharply his request to Congress this week for more aid to the Nicaraguan Contra rebels (FT 26-1-88, 3.1) Mr Reagan rejected Mr Ortega's renewed call in the letter for direct US-Nicaraguan talks. (FT 26-1-88, 3.1) They believe that most of the help should go to first-time buyers and there should be a phased reduction of relief for established home owners (DMi 8-10-87, 4.2) a dozen young victims in the lottery of life at Birmingham Children 's Hospital. (DMi 27-11-87, 1.3)

In all the above examples, a noun is modified by two or three consecutive prepositional phrases. Occasionally there may be ambiguities as to which head of the noun phrase is modified by a particular prepositional phrase. In a phrase like his request for more aid to the Congress, there would be an ambiguity as to whether the second prepositional phrase modifies the noun, aid, in which case the requested aid would go to the Congress or whether both prepositional phrases modify the same noun request. In this latter case it is the request that goes to the Congress and not the aid. In the actually attested version, however,

76

Structure of noun phrases

there is no such ambiguity because the order of the two modifying phrases is reversed. The phrase in (61) is less clear-cut. Does the phrase for established home owners modify the noun reduction or relief? The context talks about increasing help to first-time house buyers, which apparently necessitates a reduction of help elsewhere. This is both a "reduction of relief and a "reduction to established home owners". It seems to be possible to leave out either of the prepositional phrases, and still a comprehensible and acceptable rest would remain. On the other hand, it does not seem to be possible to reverse the order of the two prepositional phrases, ?*a phased reduction for established home owners of relief. If this judgement is correct, the second prepositional phrase in (61) above, should better be seen as modifying relief rather than reduction. Example (63) illustrates a possible type of ambiguity. The prepositional phrase following the noun activist can both be understood as a clause adverbial or as a postmodifier. (63)

He emerged as a leading activist in the League management committee's revolt against its president Carter over the secret compromise deal with Maxwell (DMa 22-12-87, 31.2)

The fact that the long italicised prepositional phrase is ambiguous can be made more apparent if this sentence is reformulated in such a way that the prepositional phrase is moved to positions in the sentence that can only be occupied by either a postmodifying prepositional phrase, which is part of the noun phrase it modifies, or else to a position that is restricted to adverbials. As it stands, however, it is too long and would produce awkward results in all but sentence final position. For this reason it is being reduced for this test. (63)

a. b. c. d.

In the committee 's revolt, he emerged as a leading activist He emerged in the committee 's revolt as a leading activist *He in the committee's revolt emerged as a leading activist The leading activist in the committee 's revolt said after the meeting ...

In (63a) and (63b), the prepositional phrase in italics can only be understood as an adverbial. As such it can occupy sentence initial position, the position immediately after the verb, or, as in the original (63) above, sentence final position. But as (63c) shows it cannot intervene between the subject and the verb. If the noun phrase a leading activist is made into the subject of a clause, as in (63d), it can be followed by the prepositional phrase because it is now understood as a postmodifier which is a constituent of the subject noun phrase.

Postmodification

77

4.3.5. Postposed adjectives In a limited range of cases, an adjective can follow the noun which it modifies. (64) to (66) are typical examples. (64)

(65)

(66)

the city would surely have been packed with volunteers, art historians and fund-raisers anxious to save a cultural heritage they felt they shared. (G, 31.10.87, 14.4) the biggest UK-owned toy-maker famous for My Yellow Teapot, A la Carte Kitchen, and most recently Mania Force space vehicles, (DE, 23.11.87,30.3) MPs uneasy about the great new privatisation experiment, (DMa, 19.12.87,29.1)

In all these cases the adjective itself is modified, either by a prepositional phrase or by an infinitive clause. In prenominal position these adjective phrases would be ungrammatical (*anxious to save a cultural heritage volunteers) while, on the other hand, the same adjectives without the modifications would only be possible in their normal prenominal but not in postposed position (anxious volunteers but ^volunteers anxious). This structure corresponds to a relative clause in which the head noun functions as subject of a copular clause (volunteers who are/were/would be anxious to save a cultural heritage). Two other types of postposed adjectives are those modifying heads that do not take adjectives in front of them (e.g. something different) and a small number of set phrases (e.g. heir apparent) (Quirk et al. 1985: 1293 f.).

4.3.6. Appositions Only some types of appositions are clear cases of postmodification. Most typically appositions consist of two coreferential noun phrases which fulfil basically the same syntactic role within one single clause. Both parts of an apposition are called appositives. Depending on the type of apposition, the second appositive usually modifies or defines the first appositive, and this is why I treat them under the general heading of postmodification. The other types of apposition, however, are in some instances at least equally important and therefore they must also be mentioned at this stage. The appositive clauses, which were discussed in 4.2 above, are just a special case of apposition as I will argue below. Again using Quirk et al.'s terminology, we can distinguish between full and partial apposition. In a full apposition either of the two appositives can be deleted, and the resultant sentence is still grammatical.

78

(67) (67)

Structure of noun phrases

Mr Simpson, the vicar ofBarmston, said last night:... a. Mr Simpson said last night:... b. The vicar ofBarmston said last night:...

In a partial apposition, on the other hand, only one appositive can be deleted without rendering the resultant sentence ungrammatical. (68) (68)

An unusual present was given to him for his birthday, a book on ethics. a. An unusual present was given to him for his birthday. b. *Was given to him for his birthday, a book on ethics. (Quirk et al. 1985: 1302)

The second dichotomy, which is less important for my purposes, distinguishes between strict and weak apposition. The former refers to appositions in which both appositives belong to the same general syntactic class, as was the case in the two examples above. The latter consequently refers to appositions with two appositives of different syntactic status as in (69): (69)

His only interest in life, playing football, has brought him many friends. (Quirk et al. 1985: 1303)

The apposition in (69) is again full because either of the two appositives could be omitted without affecting the grammaticality of the resultant sentence. The appositive clauses in (70) and (71) are examples of partial, weak appositions: (70)

(71)

the Government received warnings last week - while the siege at Peterhead Prison was going on - that Perth could be the next flash point. (I 6-10-87, 1.3) Another theory, that the fire had started in rubbish under the escalator, was also challenged when ... (T 20-11-87, 2.6)

The two appositives in both cases have a different syntactic status. The first appositives, warnings and another theory respectively, are noun phrases while the second appositives are clauses. In both cases only the second appositive but not the first could be omitted without affecting the grammaticality of the resultant sentence The third dichotomy concerns the difference between restrictive and nonrestrictive apposition. So far all examples - with the exception of (70) - were nonrestrictive appositions, in which both appositives contribute relatively independent information units. In writing this is indicated by bracketing off the second appositive with commas. (72)

a. Mr Simpson, the vicar of Barmston, said last night:... b. The vicar of Barmston, Mr Simpson, said last night:...

Postmodification

79

If the two appositives form one information unit in that one restricts the reference of the other, the apposition is restrictive, as in (73) and (74): (73) (74)

The vicar Mr Simpson Architect Emo Goldfinger

The second appositives, Mr Simpson and Erno Goldfinger, are needed in order to establish the reference of the first appositive the vicar and architect respectively. The three dichotomies are summarised in (75) (Quirk et al. 1985: 1305):

(75)

apposition

strict [same syntactic class] weak [different syntactic class] (/"" [either omissible] I partial [only one omissible] nonrestrictive [different information unit] restrictive [same information unit]

Full, nonrestrictive appositions share some similarities with noun phrases with postposed adjectives in that they, too, can be seen as reductions of copular relative clauses. Thus: (76) (77)

Mr Simpson, who is the vicar ofBarmston, said last night:... His only interest in life, which is playing football, has brought him many friends.

In restrictive appositions, on the other hand, this is not possible: (78) (79)

*The vicar who is Mr Simpson *Architect who is Brno Goldfinger

Sentence (78) would only be possible if the reference of the vicar were established independently. In that case, the apposition - and the corresponding relative clause - would be nonrestrictive. One of the two appositives in all types of apposition is the defined while the other is the defining. In partial appositions, in which only one appositive is omissible, the independent one is the defined and the dependent or omissible one is the defining. In (79) above, Erno Goldfinger is the defined appositive while architect is the defining one. In full appositions, on the other hand, it is generally the first appositive which is the defined, and the second is defining. Thus in (72a) above Mr Simpson is defined by the second appositive, the vicar ofBarmston, and vice versa in (72b). Quirk et al. (1985: 1305) indicate that partial, strict, restrictive appositions of the type architect Erno Goldfinger are typical of journalistic style. How-

80

Structure of noun phrases

ever this is true only for some of the British national newspapers. In fact this type of construction is one of the most clearly stratifying features across the various papers. It is a very popular construction in the down-market papers, and it is almost completely shunned in the up-market papers. For this reason I will devote all of chapter 9 to appositions that combine a personal name and a noun phrase. In that context I will need a further terminological distinction to refer to the two appositives. I will refer to appositives such as Erno Goldfinger or Mr Simpson as name appositives, and to those like architect or the vicar of Barmston as descriptive appositives, irrespective of their positioning in the apposition. The name appositive in most instances identifies a referent unambiguously even if this may not be sufficient for the majority of readers of a newspaper because they do not know that particular person. The descriptive appositive, on the other hand, in some sense describes the referent by indicating his or her role which most probably will be directly related to his or her newsworthiness. This distinction makes it possible to say that in (72a) above the descriptive appositive is the one that defines the name appositive, while in (72b) the name appositive defines the descriptive appositive.

4.4. Multiple modifications As I have outlined above, noun phrases range from the very simple to the very complex. They can consist of a single unmodified pronoun, or they can contain one or several modifiers that may have their own complex internal structure. Theoretically there is no upper boundary to the complexity that is possible within noun phrases. The complexity may be increased by modifying one single head by several independent modifiers. I shall call modifiers that apply independently to one single head concatenated modifiers, whether they are premodifiers or postmodifiers. The complexity of noun phrases may also be increased by modifiers that are modified themselves. In this case I will use the term embedded modifiers. (80)

(81)

Since his arrival at Anfieldfrom Watford for a fee of £900,000, John Barnes has captured the imagination of supporters ... (G 31-10-87, 18.3) to maintain on our behalf collections of modern British and international art that embrace both the best and most significant examples, which are not necessarily coincidental. (FT, 26.01.88: p. 17.1)

Multiple modifications

(82)

81

She also expressed Britain's support for the treaty on the elimination of intermediate nuclear weapons which Mr Gorbachev and President Reagan will sign, (DT, 27.11.87, 1.2)

Sentence (80), quoted as (57) in 4.3.4 above and repeated here for convenience, shows three concatenated prepositional phrases, the last of which is modified itself by a prepositional phrase. In (81) the noun collections is modified by a prepositional phrase whose head is in turn modified by a relative clause. Embedded within this relative clause is yet another relative clause. Quirk et al. (1985: 1298) call such modifying clauses that are embedded within other clauses pushdown elements. In (82) the noun support is modified by a sequence of prepositional phrases that are embedded within each other. It is only the first one which modifies the head noun directly, whereas the two subsequent ones modify in turn the head noun of the preceding prepositional phrase. The last head noun in this sequence, weapons, is followed by a relative clause which, however, does not modify this last noun, and neither does it modify the noun support it clearly modifies treaty. Thus it is not the argument noun phrase which is modified by two concatenated postmodifiers, a prepositional phrase and a relative clause, it is one of the postmodifiers which is further modified by concatenated postmodifiers, by a prepositional phrase with another prepositional phrase embedded within it, and by a relative clause. This interpretation is not directly encoded in the structure of this complex noun phrase. It is only the semantics which makes it clear that the two men will sign a treaty rather than any of the other referents in this noun phrase. In chapters 7 and 8 below, I will analyse premodifiers and postmodifiers as they are attested in my corpus of newspaper English and how they vary across the different categories of newspapers and across the different newspaper sections. There I will illustrate in detail the extent to which the modifying potential of noun phrases is exploited. A summary of all the categories that I distinguish in the analysis of my corpus and some information on the practical details of the analysis are given in Appendix I, Coding Scheme.

5. Noun phrases as style markers

5.1. Introduction In chapter 3 above, I outlined three different approaches to the study of language in its social context. Two of these approaches, correlational sociolinguistics and traditional stylistics, relate the use of relevant linguistic features to their contexts of use. At the end of that chapter, I reviewed a number of studies of media language using a wide range of linguistic features. As this study is devoted to syntactic variation in newspaper language, I went on, in chapter 4, to sketch one area of the grammar of English, the noun phrase, which looks particularly promising as a locus for syntactic variation. In this chapter, I shall tie the two strands together and show in what respect the noun phrase can be used as a variable, and what results are to be expected from such an investigation. Many researchers have recognised the value of the noun phrase as a style marker, both those who work in a correlational sociolinguistic framework and those who work in a traditional stylistic framework. The former concentrate on noun phrase structures that are in a paradigmatic relationship and - at least to some extent - are similar in meaning, such as different realisations of relative pronouns or the opposition between the genitive and the o/-construction. The latter, on the other hand, investigate for instance the complexity of the noun phrases and take this to be one of the stylistically relevant features. In the following I shall discuss the two approaches in some detail because this will provide the basis for my own analysis of the language of the British national daily newspapers.

5.2. Noun phrase variables in paradigmatic relationships Over the last decade or so, many researchers have tried to apply a Labovian framework to variables beyond the immediate limits of phonology, and some of them have investigated the noun phrase as locus of variability. However, few, if any, have tried to adopt the free variation hypothesis to this level of variation. It is just too obvious that in this respect phonology differs substantially from other levels of linguistic description (cf. Bolinger 1977: 3; Lavandera 1978; Jacobson 1980; and Romaine 1984b). The collection of papers in

84

Noun phrases as style markers

Trudgill and Chambers (1991) is a clear indication of the recent increase of interest in syntactic variation. The only way to salvage the Labovian framework for variables beyond phonology is to make a clear methodological distinction between the linguistic and the non-linguistic factors affecting the choice of one variant over the other. If the two can be kept separate, then we are in a position to assess the extent to which social or stylistic factors influence the chosen variable without the necessity to claim that there is free variation between its variants. Two areas within the domain of the noun phrase structure lend themselves quite readily to this type of investigation. The first area is the choice of the relative pronoun, and the second the alternation between the genitive construction and the o/-construction.

5.2.7. Relative clauses In her important study, Romaine (1982) extends the methodology of correlational sociolinguistics in several directions. Not only does she use syntactic variables rather than the more traditional phonological ones, but she also uses a corpus that consists of written texts, and furthermore of texts representing an earlier stage in the development of the language she investigates. She analyses the occurrence of relative clauses and in particular the different types of relative marker in a number of Middle Scots texts, written between 1530 and 1550. Romaine does not use a binary variable, that is to say a variable with two variants, but a three-way opposition. The relative marker in Middle Scots, as in Modern English, could take one of three forms, WH (quhilk - which), TH (thai) or 0 (instances of omitted relative markers). She meticulously distinguishes between linguistic and extralinguistic factors affecting the choice of the relative markers. Among the former she discusses the distinction between restrictive and nonrestrictive relative clauses, the semantic and syntactic features of the modified noun phrase, and the syntactic position or function of the relative marker, whereas the latter are discussed in terms of contextual styles. Croft (1986: 274) is of course correct when he points out about Romaine's linguistic factors that "most if not all of these factors are fundamentally extralinguistic". It is indeed largely an extralinguistic matter whether a given noun modification is restrictive or nonrestrictive, or whether the modified noun refers to an animate or inanimate entity to give just two examples. But Croft really misses the point in this respect. It is an extralinguistic matter what type of antecedent, i.e. head noun phrase, is selected, but Romaine's point is that this choice will then affect, or to some extent restrict, the choice of relative

Noun phrase variables in paradigmatic relationships

85

markers. Hence the relative markers are to some extent dependent on the linguistic structures in which they are embedded and which themselves, naturally, depend on extralinguistic factors. But viewed in Croft's way, everything in the linguistic structure of a text will - ultimately - depend on extralinguistic factors. One of the merits of Romaine's book is the fact that she approaches her data from several quite different angles. In this way she can test out different sociolinguistic methodologies on the same set of data. Her first approach corresponds to what I have called traditional stylistics even though she does not use this term herself. She establishes different measurements relating to her relative clauses, which she then applies to her data as a whole and to its subparts. She finds (1982: 140) that WH forms are the most common markers for non-restrictive relative clauses with some TH forms but very rarely 0, whereas in restrictive relative clauses the most common marker is TH. Both WH and 0 occur also but infrequently. The animacy of the antecedent, in marked contrast to present day English, does not seem to affect the choice of relative marker in Romaine's data of Middle Scots. The type of determiner of the antecedent likewise does not influence the choice in any noticeable way. The syntactic position of the relative marker within the relative clause has some influence on its form. The system differs, however, from modem English in its standard British English form. There are, for instance, a number of cases of 0 relative markers in subject position of their clauses. The two main measurements within the broad framework of traditional stylistics are syntactic complexity and choice of relative marker. First, she discusses the accessibility of the syntactic positions to relativisation, that is to say she investigates the syntactic roles played by the relative markers within their relative clauses. She refers to Keenan and Comrie's (1977, 1979) accessibility hierarchy, which claims that 1) a language must be able to relativise subjects, 2) that if a language can relativise a position low on the accessibility hierarchy, it can also relativise all higher positions, and 3) that languages vary as to how far down on the accessibility hierarchy syntactic positions can be relativised. (Keenan and Comrie 1977: 67). These claims refer to the following accessibility hierarchy (1977: 66): (1)

Subject > direct object > indirect object > oblique > genitive > object of comparison

In support, Keenan and Comrie cite numerous examples from data based on about fifty languages. Many Western Malayo-Polynesian languages, for instance, can only relativise subjects. Welsh can relativise subjects and direct

86

Noun phrases as style markers

objects but not any other positions. Basque and Tamil can both relativise indirect objects besides subjects and direct objects, but again no positions further down the hierarchy. Very few languages allow all positions of the hierarchy to be relativised. English is one of them if (2) is acceptable: (2)

the man who Mary is taller than

This relates directly to a further claim that there is a direct correlation between the position on the accessibility hierarchy and gramm.aticality judgements. Judgements are uncontroversial towards the top end of the hierarchy but for every individual language they become more and more disputable towards their respective cut-off points in the hierarchy. Romaine extends their claim even further. She shows that positions high on the accessibility hierarchy are more often relativised than positions in lower ranks. The second measurement within traditional stylistics is the index of relative marker deletion, that is to say the frequency of the 0 relative marker in relation to all her relative clauses. Romaine then goes on to analyse her data by implicational scaling. The working hypothesis behind this approach is that each text in her corpus represents an isolect which is part of a polylectal grammar but is itself invariant. It varies from a neighbouring isolect with respect to the output of one single rule. The panlectal grid consists of the totality of possible sets of rules for an arbitrarily limited area in space and/or time. The results of the implicational scale are used as a means of checking data against the predictions of a Baileyan wave model .... The assumption is that if the data scale to an acceptable degree, then they can do so by virtue of the fact that a succession (or wave) of rule changes has spread evenly through the grammar of a community. (Romaine 1982: 170) For the purpose of this framework, she distinguishes four syntactic environments and notes for every text in her corpus whether the relative marker is always (1), sometimes (X) or never (0) deleted in that particular environment. However, the terms "always" and "never" have to be taken with more than just a pinch of salt, they refer to more than 25 and less than five per cent deletion respectively, with the variable cases lying between five and 25 per cent. As a result she presents table 5.1.

Noun phrase variables in paradigmatic relationships

87

Table 5.1. Implicational scale for relative deletion in Middle Scots texts

Text The Scottish Correspondence Ane Satyre The Bannatyne Manuscript Acts of the Lords Council Burgh Records Boece Sheriff Court Book

Syntactic position of deleted relative Temp Dir.obj. Prep. 1 1 1 1 1 X 1*

)£ ()

1 1 X X 1* X 0

Subj. X X X 0 0 0 0

Deviations in bold face and asterisked (Romaine 1982: 171, table 6.31).

This framework is designed to plot language change in progress. A rule deleting the relative marker applies "categorically" (i.e. more than 25 per cent) in particular syntactic environments, it then becomes variable and finally becomes categorical again in non-application (i.e. fewer than five per cent). However, the table does not give any information on the direction which the change is taking. It could be from categorical application to categorical nonapplication or the other way round. It is only truly diachronic data that can reveal in which direction a particular change is moving. It is customary to arrange the rows and columns in such a way that a change can be said to enter the grammar at the bottom left-hand comer of the table spreading right across environments and upwards across the isolects usually represented by individual speakers but here by individual texts. It must be stressed, however, that Romaine sets up this table mainly in order to be able to criticise this particular approach (cf. in particular her excursus, 1982: 177-182). Finally, in a third approach, Romaine subjects her data to the CedergrenSankoff variable rule program. This presents some problems because this program had been set up to handle the more traditional type of variables with only two variants, whereas Romaine is dealing with a three-way distinction. What the program, as developed by Cedergren and Sankoff (1974), does, is to estimate mathematically on the basis of observed frequencies the probability with which the relevant rule applies. Frequencies are random variables that cannot be predicted with 100 per cent accuracy, but probabilities of rule application are fixed and part of the competence of speakers (Romaine 1982: 183). Romaine is again highly critical of this approach, and she challenges Labov's (1975: 228) view that a probability theory in sociolinguistics affords a "dramatic increase in power and perception": Probability theories might be considered more powerful with respect to their domain of application because, strictly speaking, they cannot be confirmed

88

Noun phrases as style markers

or disconfirmed by empirical findings. If probability statements are not falsifiable, then they have no empirical content. Thus, they can have no explanatory or predictive power. (Romaine 1982: 183) As I will try to show in detail in the concluding chapter, it is unreasonable (and self-defeating) to set methodological objectives for linguistic and sociolinguistic research which are not even attainable in natural science. My criticism of this approach is based on the fact that it hinges on a highly implausible notion of speech community. It assumes that there are homogeneous speech communities sharing a uniform grammar. The observed variation in surface frequencies of linguistic features is explained as the result of a probability index appended to each rule indicating the likelihood with which it applies for every speaker in every given context. Nevertheless, Romaine's study is important for my present purposes because she discusses many of the fundamental issues of sociolinguistics and because she uses syntactic variables to illustrate her arguments. In the final chapter I shall come back to some of her points on sociolinguistic methodology. As hinted at above I shall suggest that she reaches unduly pessimistic conclusions primarily because of her unrealistic presumptions for sociolinguistics as a scientific method. Her results are moreover important in that she uses one particular type of noun modification as a variable. The choice of the relative marker seems to be quite obvious from a sociolinguistic point of view. Even though she does not talk in terms of free variation - in fact she is fairly sceptical whether it is a useful notion for syntactic variation at all - it nevertheless seems clear that this is one of its attractions. There are linguistic and extralinguistic criteria, some of which have been summarised above, which govern to some extent the choice of one of the three types of relative marker, but given identical contexts, there are no claims that the choice of relative marker would change the meaning or the function of the relative clause. Cheshire (1982 and 1991) is another fairly recent example of a detailed analysis of variation going beyond phonology. Using a Labovian framework, she analyzes a group of adolescents frequenting two playgrounds in Reading, England. By using such a small and comparatively homogeneous group of speakers, she excludes a considerable number of non-linguistic parameters, such as age variation, regional variation and to a large extent social variation, since all her informants come from a basically working class background. Her linguistic variables are also severely restricted. She excludes lexical and phonological variation and concentrates on morphological and syntactic features which are used in non-standard forms by her informants. One of these features is the relative pronoun. The paradigm of relative pronouns in Reading

Noun phrase variables in paradigmatic relationships

89

English has one variant more than Standard English, because what is also used in this function. A further difference to Standard English is the fact that the 0 relative pronoun is also used as a subject of its clause, which is not possible in Standard English.

5.2.2. Genitive versus of-construction Altenberg (1982 see also 1980) offers one of the most comprehensive recent studies of the variation between the genitive and the «/-constructions, as illustrated by the following two examples: (3) (4)

Her father's arrival changed everything The arrival of her father changed everything (Altenberg 1982: 11)

At the outset, Altenberg stresses that the two variants are not in free variation even though they "convey roughly the same meaning" (1982: 11). Like Romaine (1982), he concentrates on historical and written data. In his case the texts were all written in the 17th century. He distinguishes six types of influencing factors, only one of which is concerned with non-linguistic features. Phonetically, a final sibilant in the modifying noun phrase is a significant, albeit not overriding, factor. Especially in formal prose, it tends to increase the percentage of «/-constructions. Morphologically, a regular plural ending in the modifying noun also tends to increase the percentage of «/-constructions. Syntactically, three main types of conditioning factor are distinguished: the premodifier weight of both the modifying and the head noun phrase postmodification and other types of right branching expansions of the modifying noun phrase and the recursive expansion of the modifying noun phrase. The first of these factors has relatively little influence on the choice between a genitive or an «/-construction. A right expanded modifying noun phrase favours the «/construction whereas a right expanded head noun phrase favours the genitive construction, both of which can be seen as strategies to keep modifying and head noun phrase close together. The genitive construction is less likely to have its modifying noun phrase recursively modified more than once, that is to say two genitives in a row seem to be about the limit. The «/-construction, on the other hand, shows slightly more flexibility. Lexically, the choice between the genitive and the «/-construction is conditioned by the semantic nature of the modifying noun phrase. This appears to be one of the strongest conditioning factors. The frequency of the genitive construction is highest with modifying noun phrases denoting human individuals, somewhat lower for animals, very low for human collectives and abstract

90

Noun phrases as style markers

nouns, and practically zero for concrete nouns. The lexical category of the head noun phrase, on the other hand, is marginal and idiosyncratic. On the relational level, the choice is influenced by the kind of relation that obtains between the modifying and the head noun phrase. In Modern English, this allows John's betrayal to be interpreted as subjective and the betrayal of John as objective. Thus subjectiveness favours the genitive construction but the effect of the head-modifier relationship is to a great extent determined by stylistic factors. The last area of conditioning factors deals with communicative aspects. Texts to which the status of private, intimate and personal can be ascribed tend to favour the genitive construction whereas public, distant and impersonal texts favour the o/-construction. In the same way non-specialised texts such as conversations on everyday topics tend to use the genitive construction whereas specialised texts such as religious or political texts or texts with literary topics tend to use the ^/-construction. And finally, non-literary prose genres and strict verse forms prefer genitive constructions whereas literary prose genres prefer o/-constructions. This is only a selection of all the conditioning factors that, according to Altenberg, affect the choice between the two variants of his paradigm, and he repeatedly stresses the problem of quantifying the influence of individual factors, because in every single instance, all the relevant factors work together to determine the choice of one form over the other. But the above outline suffices to sketch a very sophisticated attempt at establishing the linguistic and nonlinguistic factors responsible for one particular example of syntactic variation. He concludes: Roughly speaking, the grammatical and thematic factors account for no less than 85% of the GEN/OF options in the variable prose material. Furthermore, if we take into account that the stylistic choices are also highly patterned, and regard GEN as the unmarked choice in informal contexts and OF as the unmarked choice in most formal contexts, the percentage of "principled" variation rises to 98%. This does not necessarily mean that the remaining 2% are in free variation or that they resist explanation altogether, only that other and more delicate instruments of analysis are needed. (Altenberg 1982: 303) Dahl (1971) writing a decade earlier than Altenberg, investigated the use of the genitive construction in journalistic style. It is generally recognised that the genitive is the preferred construction for those modifying noun phrases that are "highest on the gender scale" (Quirk et al. 1985: 1277). But Dahl points out that in journalistic style, the genitive is used in places that in other varieties would use the o/-construction. According to her, the use of the geni-

Premodifier variation

91

tive for inanimate things is particularly common in newspaper headlines, but it has already extended its influence into the main text of newspapers (1971: 141 f). As the basis of her classification, she uses a large collection of genitives for inanimate things taken out of The Times, The Daily Telegraph, the Daily Mail and The Observer. Her classes include words denoting collective communities, geographical names, the genitive before superlatives, names of animals, means of locomotion and machines, planets, buildings and places, newspapers and periodicals and some groups of less importance. It is unlikely that any of the classes mentioned by Dahl is actually restricted to newspaper language, even though it may be more frequent there. But her data does not allow any comparison with other varieties. She gives the frequency of all the constructions she discusses but they are not set in relation to the length of the texts from which they are taken or alternatively in relation to the frequency of the o/-construction. Because of this failure, it is not surprising that she could not find clear differences between the four newspapers under investigation. Her collection contains more examples from The Times and The Daily Telegraph than from the other two newspapers but this is due to the fact that "their volume of articles and news reports is greater than that of the Daily Mail and The Observer" (1971: 172) and thus it does not provide any relevant information at all.

5.3. Premodifier variation Abberton (1977) presents a very delicate analysis of premodifying structures in noun phrases. She distinguishes 19 different types of premodifiers that can occur between the determiner and the head of the noun phrase. She also sets up twelve form classes according to the different patterns of premodification, but she uses more general terms than those of her 19 types of premodifier. To give some typical examples, pattern one consists of a head noun with neither determiner nor premodifier pattern three of an optional determiner, an adjective and a head and pattern eight of a noun phrase in genitive case as a determinative and an optionally premodified head. The last of her patterns is given as the following formula: (5)

± Det + ((± Adj) + Gen. N) + ((± Adj)+(± N) + N) (Abberton 1977: 58)

The pattern is exemplified by The farmer's butter and The kind farmer's nice buttermilk. In both cases, the bracketing given by Abberton seems implausible. The determiner the is clearly part of the genitive noun phrase acting as a

92

Noun phrases as style markers

determinative for the entire noun phrase. The genitive noun phrase and the determiner can only as a unit be replaced by a name, viz. John Brown's butter but not *The John Brown's butter. Abberton then goes on to compare the use of these patterns in eight texts drawn from the Survey of English Usage, four from the category of novels and four from scientific writing. The corpus consists of some 10,000 words, which yields 6160 noun phrases for her analysis. But several of her twelve patterns are only represented by very few tokens. Hence her conclusions are not very far reaching. Her figures are too small to reveal any clear overall patterns. The general conclusion is that both types of English examined are remarkably similar in nominal group predmodification [sic!] structure: the vast majority of nominal groups are not premodified or are premodified by only a determiner and/or one adjective whose type tends to vary but slightly with the style of English in question. (1977: 63, her emphasis) The most important differences that she is able to locate through her sophisticated classification scheme, that "postmodified nominal groups are far more common in the science texts" (1977: 62). It must be conceded, however, that it was presumably the classification itself that interested Abberton in her research rather than the correlations with the two types of written British English. The corpus seemed to serve more as a data base for the exemplification of her classes. Coates (1977) investigates sequences of premodifiers in 17 texts of the Survey of English Usage. She compares such sequences in different structural positions of the noun phrase. Her main interest lies in the prenominal position between the determiner and the head, which she contrasts to postnominal position after the head to preprenominal position before the determiner and to predicative position outside the noun phrase altogether after the copulative verb to be. Ordering, it seems, is basically a semantic matter. A head is modified to form a new conceptual unit, which can then be modified in its turn. Modification at any level can involve several modifiers or only one. And at the point where the nominal group expresses what the speaker/writer wants to say, modification will cease. (Coates 1977: 19) The prenominal slot turns out to be particularly important in that it very often differs considerably from premodifiers that are further removed from the head. The item in prenominal position very often forms a new head together with its head to form a "temporary compound" (1977: 21), which may itself be further modified.

Complexity as style marker

93

Sichelschmidt (1989), finally, investigates the order of prenominal adjective sequences in German. He is mainly interested in the psycholinguistic processes that govern the production and reception of premodifying adjectives. Is it the first adjective of a sequence which stays most prominent in the memory, or is it the last one, the one nearest to the head of the noun phrase (i.e. is the principle of primacy or the principle of recency more important)? He finds that potential semantic contrasts of adjectives to other adjectives in the context are generally more important than the mere syntactic ordering.

5.4. Complexity as style marker Aarts (1971) is one of the first to notice the stylistic significance of the complexity of noun phrases. His short but important paper is based on extracts totalling 72,000 words from the Survey of English Usage at University College London. His corpus can be divided into four categories: light fiction, scientific writing, informal speech, and formal spoken and written English. It was his working hypothesis that the distribution of noun phrase types within the English clause structure is not random, but that the subject position is associated with structurally "light" noun phrases, whereas non-subject positions are associated with structurally "heavy" noun phrases. He further hypothesised that the distribution pattern would be significantly influenced by the variety of English in which it occurs. Aarts counts those noun phrases as "light" that consist of a simple pronoun, a name or an unmodified noun with or without a determiner. All other noun phrases are understood to be "heavy". After presenting several tables with the detailed figures he summarises his findings in the following table: Table 5.2. Distribution of "light" and "heavy" items after Aarts

Subjects Non-subjects

"Light" items

"Heavy" items

6749 4753

1149 4310

Source: Aarts (1971: 290, table 10).

Table 5.2 provides strong evidence in support of his first hypothesis. "Light" noun phrases are more frequent in subject position. 85.5 per cent of the noun phrases in subject position are classified as "light" as opposed to only 52.4 per cent of the non-subjects. On the basis of the figures for "light" and "heavy" noun phrases in the four text-groups, he can show that there is also some evi-

94

Noun phrases as style markers

dence for his second hypothesis that the distribution of noun phrase types is influenced by the variety of English, but he is very cautious and does not attempt to draw far-reaching conclusions in this respect, even though the data as he presents it shows clearly that the association between "light" items and subject is greatest in informal speech and smallest for scientific writing. Quirk et al. (1972: 933-934, and again 1985: 1350-1352) reprint his figures (without acknowledgement) in a somewhat modified table but with the same distinction between "light" and "heavy", or, as they prefer to call it, between simple and complex noun phrases. Aarts' subsample of "formal spoken and written English" is merely labelled "serious talk" in the 1985 version of their grammar. Unfortunately, the table with which Quirk et al. (1985: 1351) illustrate these findings is incomplete and it contains some blatant mistakes. Therefore I present the entire table with the corrected figures as table 5.3. Quirk et al.'s table lists the number of simple and complex noun phrases in subject and in non-subject position. For the entire sample they add the percentage figures and the totals for the noun phrases both in subject and in non-subject position. For the four subsamples they do not give the percentage figures or the totals. This saves valuable printing space but it makes it difficult to interpret the figures without having recourse to a pocket calculator. The first column of their table gives the total of each row, that is to say the total of both the simple noun phrases and the complex noun phrases either in subject position or in non-subject position. These row totals are also given for the subsamples individually. However, the figures that are given are not the correct sums of the respective rows. The names and pronouns, and the noun phrases with multiple modification are subsets of the simple and the complex noun phrases respectively and should, of course, not be added to these sums. The row totals that are given also do not add up correctly to the grand total of 16961 of the entire corpus. The correct sums of all simple and all complex noun phrases for each row, however, do add up to the stated grand total, and these corrected figures also correspond to those given by Aarts. (1971: 291, table 9).

Complexity as style marker

95

Table 5.3. Noun phrase structure and distribution after Quirk et al. Total

all simple NPs

names & pronouns

all complex NPs

multiple modification

whole sample

subject % not subject % total %

7898 46.57 9063 53.43 16961 100.00

6749 39.79 4753 28.02 11502 67.81

5821 34.32 2193 12.93 8014 47.25

1149 6.77 4310 25.41 5459 32.19

456 2.69 1777 10.48 2233 13.17

informal speech

subject % not subject % total %

2212* 52.77 1980* 47.23 4192 100.00

2064 49.24 1169 27.89 3233 77.12

1941 46.30 677 16.15 2618 62.45

148 3.53 811 19.35 959 22.88

62 1.48 327 7.80 389 9.28

fiction

subject % not subject % total %

2431* 46.45 2803* 53.55 5234 100.00

2220 42.41 1682 32.14 3902 74.55

1943 37.12 754 14.41 2697 51.53

211 4.03 1121 21.42 1332 25.45

92 1.76 434 8.29 526 10.05

serious talk

subject % not subject % total %

2088* 45.40 2511* 54.60 4599 100.00

1745 37.94 1273 27.68 3018 65.62

1478 32.14 599 13.02 2077 45.16

343 7.46 1238 26.92 1581 34.38

127 2.76 492 10.70 619 13.46

scientific subject writing % not subject % total %

1167* 39.75 1769* 60.25 2936 100.00

720 24.52 629 21.42 1349 45.95

459 15.63 163 5.55 622 21.19

447 15.22 1140 38.83 1587 54.05

175 5.96 524 17.85 699 23.81

Source: Quirk ef al. (1985: 1351, table 17.124). Asterisked figures are corrections of obvious faults. Percentage figures relate to the total number of noun phrases per subsample. Percentage figures and column totals for the four different styles are my addition.

On the basis of this table, Quirk et al. observe that less than one third of the noun phrases in the entire sample are modified and less than one-eighth have multiple modification (the exact figure for multiple modification is 13.17 per

96

Noun phrases as style markers

cent, which is, of course, slightly more than one-eighth rather than less). The differences between the four subsamples are very considerable in this respect. In informal speech and in fiction, as few as about one-fourth of all noun phrases are modified and only about one-tenth have multiple modification, whereas more than half of the noun phrases in scientific writing are modified and almost one-fourth have multiple modification. Quirk et al. further observe that nearly one-half of the noun phrases in the sample are pronouns and names, but again there are considerable differences which need to be pointed out. In informal speech almost two-thirds are names and pronouns but in scientific writing they comprise only just over one-fifth. There is a clear association of the complexity of the noun phrases and their syntactic position. In their words "the majority of simple noun phrases - and the overwhelming majority of names and pronouns - are subjects of clauses or sentences, but only rather less than a quarter of complex noun phrases are subjects" (1985: 1351). These facts are less readily apparent from the table above because the relevant percentage figures are not given. For the sample as a whole, 58.7 per cent of all simple noun phrases appear in subject position. The association is strongest in informal speech. In this style 63.8 per cent of the unmodified nouns are in subject position and only 15.4 per cent of the modified nouns. This is a difference of 48.4 percentage points. The same difference is successively reduced across the four styles. It is 41.1 percentage points in fiction, 36.1 in serious talk and 25.2 in scientific writing. On the basis of these overwhelming differences, Quirk et al. (1985: 1352) conclude that even such coarse-grained comparisons as these make clear how sensitive is the noun phrase as an index of style and how responsive it can be to the basic purpose and subject matter in varying types of discourse. The application of a chi-square test to the four values of unmodified and modified noun phrases in subject and nonsubject position respectively in the four subsamples of Quirk et al. 's table reproduced above, reveals that the difference between these samples is statistically highly significant (χ 2 = 1093.64, df = 9, ρ < 0.001). In view of this fact, the validity of the figures given for the sample as a whole are of course somewhat reduced. In strict statistical terms it is in fact inadmissible to combine samples that differ from each other at a level of statistic significance, (cf. Butler 1985: 112-123; Hatch and Farhady 1982: 165-172; Woods et al. 1986: 132-153). And it is easy to see why this is so. The four subsamples vary considerably in size, ranging from 17.3 to 30.1 per cent of the entire corpus, and thus they contribute rather differently to the overall result. However, it is not at all clear

Complexity as style marker

97

whether the four subsamples are really coherent samples, there may be, and presumably are, systematic variations within each of them. This is an example of top down analysis. Generalisations are made about a large corpus as a whole, and subsequently steps are taken to single out subsets within the sample. The subsamples themselves may also turn out to have internal variations at a level of statistic significance. If such variations are found, the validity of the generalisations at the level above are of course reduced in their validity, but not, I believe, rendered invalid, particularly so because they provided a necessary step in establishing the newly found variability. Varantola (1984) deserves some considerable attention because in some respects her aims are similar to mine. She investigates the noun phrase as a style marker in a corpus of engineering English. For her this is an example of a register or a language for special purposes (LSP). In her methodology (1984: 51) she refers explicitly to Enkvist (1964, 1973, 1978). She compares her corpus of engineering English with a "reference corpus, a ' norm' that is used as a yardstick" (1984: 51). Varantola's corpus of engineering texts consists of articles and news items published in professional engineering journals that are addressed to the "general specialist" (1984: 53) rather than to a highly specialised audience, on the one hand, or to lay readers, on the other. The "norm" corpus consists of articles and news items published in The Economist, New Society, The Observer and Newsweek. Both corpora are predominantly but not exclusively British English. It may be regretted that American English texts were included at all because it is unclear if and to what extent this inclusion skews the results. As a general overview of her data she presents the figures that allow a comparison with Quirk et al.'s data. She bases her comparison on the 1972 version because her work of course predates the 1985 version.

Table 5.4. Distribution of NP types in engineering English as compared with a "norm"

Engineering texts General texts

N % N %

Total

all simple NPs

names & pronouns

all complex NPs

multiple modification

2547 100.0 1756 100.0

628 24.7 767 43.7

96 3.8 179 10.2

1919 75.3 989 56.3

1259 49.4 610 34.7

Source: Varantola (1984: 89, table 5, and 90), but transformed into Quirk et a/.'s (1985: 1351) format used above. The subcategory of simple NPs does not include names.

98

Noun phrases as style markers

There are some quite remarkable differences between her data and Quirk et al. 's data. Even her "norm" corpus of journalistic texts has a higher percentage of modified nouns than any of Quirk et al.'s categories. Well over half of all its nouns are modified. In her main corpus of engineering English just over three quarters are modified and almost half of all the nouns have multiple modification. The degree of pronominalisation, on the other hand, is minimal. In the "norm" corpus about ten per cent of all noun phrases are pronouns, whereas in the main corpus that figure is less than four per cent. However, it is difficult to compare this last figure with Quirk et al. 's results because Varantola does not give the figures for names but only for pronouns. Some caution in interpreting her results may be advisable in any case. She concedes that her method of calculation "may be slightly different" (1984: 89), and moreover the category of complex noun phrases given in the table mentioned above does not include the noun phrases with multiple modification. Thus "complex" for her really means "one modifier only". In table 5.4 above I have recalculated the figures in such a way that the category of multiple modification is a subcategory of all the complex noun phrases. But even so, the totals of complex noun phrases for the two corpora yield different figures than those given elsewhere in her book (e.g. 1984: 59 or 91). Nevertheless, the main thrust of her argument is certainly correct when she (1984: 89f) concludes: Modification of nouns is thus a major means of information cramming and space saving in journalism. In engineering journalism its level easily becomes extremely high because of both the need for conciseness and the inherent need to label and specify. As a fundamental difference between her corpus of general journalism and her corpus of engineering journalism, she finds that the former relies more heavily on premodification whereas the latter prefers postmodification. The cause may be found in the projected readership of non-specialist texts. General journalism has to be intelligible to a wide range of readers with varying backgrounds and thus more explicit in expression whereas specialist journals can expect a very high degree of expertise from their readers and thus use language that is not overtly self-explanatory. (Varantola 1984: 91) Varantola discusses premodifiers and postmodifiers in turn. She notes that one of the main functions of premodifiers is to label and to categorise, but they are generally less specific than postmodifiers, in which grammatical relations are often made explicit which in premodifiers are left to be inferred by the reader.

Complexity as style marker

99

She repeatedly stresses the fact that mere length of either premodification or postmodification is not a very reliable indication of complexity. In fact short and compact premodification may be far more complex and require a higher degree of inferencing power on the part of the reader than a long but fairly simple relative clause in postmodification. One notable difference between her two corpora in the sphere of premodification is the fact that in the general language corpus, adjectives are by far the most common premodifier in the slot closest to the head whereas in her main corpus nouns were considerably more frequent in this position. Varantola's approach is in many respects similar to the one chosen for the analysis of newspaper English to be presented in the second part of this study. However, there are also considerable differences. She works with a fairly small corpus of some 2000 noun phrases, which she subjects to a very thorough and detailed analysis. The corpus itself is compared to a " norm" corpus of about half its size. Thus the genre of engineering English is seen as a devia-

Table 5.5. Noun phrase structure and distribution in headlines of The Times and the Daily Mirror Total

all simple NPs

names & pronouns

all complex NPs

Both samples

subject % not subject % total %

1380 50.15 1372 49.85 2752 100.00

832 30.23 666 24.20 1498 54.43

396 14.39 177 6.43 573 20.82

548 19.91 706 25.65 1254 45.57

Times

subject % not subject % total %

1105 49.73 1117 50.27 2222 100.00

636 28.62 503 22.64 1139 51.26

12 + 247 11.66 4+ 117 5.45 16 + 364 17.10

469 21.11 614 27.63 1083 48.74

Daily Mirror

subject % not subject % total %

275 51.89 255 48.11 530 100.00

196 36.98 163 30.75 359 67.74

53 + 84 25.85 30 + 26 10.57 193 36.42

79 14.91 92 17.36 171 32.26

Adapted from Märdh (1980: 109, 110, tables 17, 19).

multiple modification

100

Noun phrases as style markers

tion from, or rather a stylistically marked form of, normal English. In my approach, however, newspaper language is not seen as a coherent variety. On the contrary, its internal variation is at the very centre of my interest. For this reason I do not compare my corpus of newspaper language to any kind of "norm", but I split it up into subsamples, which hopefully will show up the expected variation. The fact that I am comparing not just two but a large number of potentially different styles necessitates a far larger corpus, and it makes it impossible - for practical reasons - to achieve the same level of delicacy in the analysis of the individual noun phrases. Märdh (1980), too, uses the framework initiated by Aarts (1971) and made more generally known by Quirk et al. (1972, 1985) in part of her study on newspaper headlines. This application is particularly pertinent for my purposes because it is also devoted exclusively to British newspapers, and it compares two different categories of newspapers, one represented by the Daily Mirror and the other by The Times. For the comparison with Quirk et al.'s data, she analyses 2222 noun phrases of verbal headlines in The Times, and 530 in the Daily Mirror. She presents the results in two different tables according to the Quirk et al. (1972: 933) format. I have transformed her data into one table corresponding to the more explicit Quirk et al. (1985: 1351) format, which I have used above. From this data (see table 5.5) the following facts emerge: There are fewer simple noun phrases in Märdh's headlines than in Quirk et al.'s aggregate corpus, but in subject position there are actually more complex noun phrases in her data than in theirs. And in The Times there are even more complex noun phrases in subject position than in the Daily Mirror. This, according to Märdh (1980: 111), "indicates that initial position, which is generally the position of the subject, is attractive for the headline writer. ... The sooner the reader's interest is caught the better." In non-subject position, too, there are more complex noun phrases in the headlines, even though in this instance the difference is much smaller. As I have pointed out above, the proportion of pronouns and names in subject position varies very considerably across Quirk et al.'s four styles. They constitute a very large proportion in informal speech and in fiction (87.7 and 79.9 per cent respectively), and a comparatively small proportion in scientific writing (39.3 per cent). In the headlines of The Times, pronouns and names in subject position constitute a smaller proportion than in scientific writing (23.4 per cent) and in the Daily Mirror a larger proportion (49.8 per cent). Not surprisingly the number of pronouns is very much smaller in the headlines than in Quirk et al.'s corpus. Pronouns often depend on a previous mention of the referent. This is not normally possible in headlines. On this basis it may actually

Complexity as style marker

\ 01

surprise that as many as 19.3 per cent of all subject noun phrases of the Daily Mirror are pronouns as compared to 1.1 per cent in The Times. Raumolin-Brunberg (1991) is one of the most recent and most detailed studies of English noun phrase structures. Her corpus consists of Early Modern English, but she does not use it diachronicly. She concentrates on the writings of Sir Thomas More (1478-1535). The first part of her excellent study is devoted to a detailed description of all the noun phrase structures attested in her corpus. In the second part she explores the noun phrase variability across the different text types of her corpus. Her corpus, too, shows some intertextual variation but it is not as dramatic as Quirk et al.'s. It is again the most informal text samples which show the highest percentage of simple nouns. There are 78.4 per cent of simple nouns in a historical text that contains many passages of direct speech. More's private letters are another sample with a high percentage of simple nouns (73.9%). At the other end of the scale there are his polemics (Apology) (58.2%) and his prayers (42.5%) (Raumolin-Brunberg 1991: 123 and 303). But the variation does not always follow the genre distribution. Texts of the same genre are not necessarily at the same end of this scale. She also tests the connection between simple noun phrases and subject position on the one hand, and between complex noun phrases and nonsubject position on the other. Table 5.6 summarises these results for her corpus as a whole. Table 5.6. Noun phrase structure in the writings of Sir Thomas More (1478-1535)

Sir Thomas More

subject % not subject %

total %

Total

all simple NPs

969 36.55 1682 63.45 2651 100.00

823 31.04 892 33.65 1715 64.69

names & pronouns

all complex NPs

multiple modification

146 5.51 790 29.80 936 35.31

Adapted from Raumolin-Brunberg (1991: 118, figure 6.3).

The percentages of simple noun phrases in subject position (31.04) and of complex noun phrases in nonsubject position (29.80) are only marginally different. This is an interesting contrast to Aarts' (1971) results given as Quirk et al.'s table above. However, her figures may not be directly comparable to Quirk et al.'s because it is not quite clear how those noun phrases were count-

102

Noun phrases as style markers

ed that is to say it is not clear whether noun phrases that are themselves modifiers are included in Quirk et α/.'s figures or not (cf. Raumolin-Brunberg 1991:68-70). However, the association of simple noun phrases with subject position and complex with nonsubject position does not show any stratifying power at all in her corpus. With the exception of the prayers all samples have roughly 50 per cent of simple noun phrases in subject position. In spite of the fact that her study deals with Early Modern English data, Raumolin-Brunberg's research has got many things in common with mine. She also uses the noun phrase as a style marker in order to explore the stylistic differences between texts of a fairly coherent and restricted corpus in her case the writings of one single man, and in mine a small segment of British media language. However, there are also considerable differences. Her emphasis lies on a detailed description of the attested noun phrase structures, which she takes to be representative not just for Sir Thomas More's writing but more generally for Early Modern English. The analysis of noun phrase complexity as a style marker, on the other hand, takes second rank.

5.5. Conclusions and hypotheses At this point it seems advisable to survey what has emerged from the review of the approaches undertaken in this and the previous chapter. On this basis it will be possible to formulate precise hypotheses about the variation that can be expected to exist in my data of newspaper language. The review of the studies by Romaine (1982), Cheshire (1982), and Altenberg (1982) has stressed more the shortcomings of their approaches than their undoubted merits. I have suggested that they are modelled too closely on a correlational methodology that had been developed for phonological variables. All three of them are very clear in the difference between phonological variables and their own syntactic variables but nevertheless they chose paradigmatic variables that - in spite of all the denials - stand in some sort of "free" variation. The variation is "free" in as far as the different realisations that are compared share a common core of meaning. It is not "free" to the same extent that phonological variation might well be argued to be free because both linguistic constraints and non-linguistic considerations govern the choice of one variant in favour of the other(s). However, the inventory of variables that suggest themselves for this kind of analysis seems to be severely limited. Beyond the relative pronoun and the

Conclusions and hypotheses

103

genitive versus o/-construction not very many plausible candidates suggest themselves, with the exception of noun phrase name appositions. The noun phrase name appositions appear to be the most promising variable with a systematic variation across my corpus of British English newspaper language. For this reason I shall devote all of chapter 9 to this construction. Chapters 6 to 8 will follow a more traditional stylistic approach. As a starting point I shall apply Quirk et al.'s methodology introduced above to my data, and then, on the basis of a cluster analysis, I will break up my corpus into six different styles, which will be discussed and extensively illustrated in chapters 7 and 8. The following list enumerates the hypotheses that emerged in the course of the above reviews of the relevant literature and in the course of the discussion of the structural properties of noun phrases in the previous chapter. The first two points concern the ordering of the data. They are the conditions that must obtain if claims are to be made about the significance of individual variables. • In all cases of interesting statistical differences, the three newspaper categories will be linearly ordered, that is to say the up-market papers will have the highest or lowest average number, or the highest or lowest proportional share of some category, and vice versa for the down-market papers, while the mid-market papers will have a value lying between the two. • It is implausible to expect the pattern to be perfect for every variable, but generally there will be little overlap between the three categories, that is to say for one variable all the individual values of the up-market papers will be higher, or lower, than all the corresponding values of the mid-market papers, which in turn will be higher, or lower, than the same values of the down-market papers. The first of these two points is an absolute condition. If it does not apply for any given variable, this variable cannot be said to stratify according to the newspaper categories. The second point, on the other hand, is not absolute. Variables can be claimed to stratify significantly even if there is some limited overlap between the categories. The decision whether some variable is relevant or not will ultimately be taken on the basis of statistical tests (chi-square or ANOVA). It is highly unlikely that the statistics indicate relevant variation in cases where the above conditions do not hold in the stated way. The subsequent hypotheses will be stated for the extreme categories only. Thus a claim that the up-market papers have a higher value for some variable χ always implies that the average value of this variable for all the up-market papers is higher than the average value for the mid-markets, which in turn is

104

Noun phrases as style markers

higher than the average value for the down-markets. The references in brackets indicate where the reasoning for these hypotheses has been given • The proportion of premodifiers versus postmodifiers will vary significantly, even though no hypothesis can be made as to the direction the variation will take. Postmodifiers are more explicit than premodifiers. It will depend on the degree to which papers try to be explicit and specific (which favours postmodifications) and the degree to which they can leave relations between head and modifier unexpressed on the assumption that the readers will be able to infer them (which favours premodification). The proportion of premodification versus postmodification will also be influenced by the fact that postmodifiers are more complex than premodifiers (section 4.4 and 4.5). • The up-markets have a higher share of adjectives in prehead position whereas the down-markets have a higher share of nouns in prehead position (section 4.4). • In whatever way the proportion between premodification and postmodification turns out to vary, if in any at all, it can be expected that the up-market papers show a higher degree of complexity. Noun phrase complexity can be measured in various ways leading to the following subhypotheses: • The total number of modifiers will be higher in the up-market papers. • The number of noun phrases with concatenated modifiers will be higher in the up-markets, and among the noun phrases with concatenated modifiers, the proportion of those with more than two modifiers will also be higher in the up-markets (sections 4.4 and 5.3). • The number of modifier embeddings will likewise be higher, and the embeddings will be deeper in the up-market papers (section 4.4). • The proportion of structurally complex modifiers will be higher in the up-market papers (section 4.4). • The association of structurally light noun phrases with subject position and structurally heavy noun phrases with non-subject position will be highest in the down-market papers (section 5.4). • The proportion of all noun phrases that are pronouns and names will be higher in the down-market papers (section 5.4). • The up-market papers have a higher share of classifying, defining and objective premodifiers in prehead position whereas the down-markets have a higher share of describing, subjective and evaluative premodifiers, which are gradable and appear in central position (section 4.4).

Conclusions and hypotheses

105

The last hypothesis does not concern the complexity of the noun phrase as such but a construction that is often said to be typical of journalese (e.g. Quirk etal. 1985: 1305). • The noun phrase name appositions promise to be a particularly significant stratifying feature of newspaper language. The format Opposition Leader Neil Kinnock will have a high share in the down-markets and a very small share in the up-markets (section 2.6.6 this hypothesis will be further elucidated below before it is empirically tested against my data in chapter 9). All the above hypotheses are formulated in terms of variation across the three newspaper categories. In the same way it is expected that there is considerable variation across the five different newspaper sections. However, at this stage it is impossible to hypothesise on the direction which this variation will take. It is unclear which section will turn out to be the most or the least complex. The general procedure will therefore be to establish the relevant variables for the newspaper categories and then test the variation of the same variables across the sections within the newspaper categories and across the entire data. The difference is of course that for the newspaper categories the directionality of the variation is a precondition. If a variable turns out to be highest in the midmarket papers, it will not be considered to be relevant. For the newspaper sections, on the other hand, the ordering of the samples will be considered to be a result rather than a precondition.

6. Distribution and complexity 6.1 Introduction In the preceding five chapters, I have tried to set the scene for the following analysis of the variability of the noun phrase in newspaper language by reviewing the relevant literature on sociolinguistic variability in general and on syntactic variation in particular. I have shown that the British daily newspapers fall into three groups according to the socio-economic profile of their readerships. I have briefly sketched the syntactic complexity of the English noun phrase in general, and I have cited several works which stress the relevance of the noun phrase as a style marker. In this and the following chapters, I shall apply this methodological apparatus to my data base in order to explore the degree to which the structure of noun phrases is a style marker within newspaper English. The present chapter will concentrate on the structural "heaviness" of noun phrases in general. In a first step I will show how the density of modifiers varies across types of newspaper and across sections before I go on to analyse the variation of structurally "light" and "heavy" noun phrases. These two measures are obviously related, but the latter measure increases the delicacy of analysis by including data on the syntactic position of the noun phrases. This part will be a replication of the study reported in Quirk et al. (1985: 1350-1352) and reviewed in chapter 5 above. The following two chapters will further refine the analysis by concentrating on premodifiers and postmodifiers respectively. On the basis of the structural analysis of noun phrases presented in chapter 4 above, I will show which types are preferred in which types of papers and in which sections. Some types of modifier turn out to vary across types of paper and/or sections to an extent that is statistically significant, while others are, perhaps surprisingly, consistent in their relative frequency. It will be necessary to study the semantics of these, structurally defined, types of modifier with considerable care in order to develop with some degree of plausibility hypotheses about the reasons for these differences. Finally, chapter 9 will be devoted to one particularly salient construction, that is to say noun phrase name appositions. They are interesting because they stratify the three types of paper more clearly than any other structural property of the noun phrase. They are also the feature that is most obviously influenced by explicit editorial policies adopted by the papers and enforced by the editors and sub-editors.

108

Distribution and complexity

6.2. The density of modifiers My data base consists of 43 samples of 1000 noun phrases each. In a first approximation, I compare the number of modifiers in each sample. The coding scheme used for the analysis of this corpus is given in appendix I. At this stage no distinction is made between simple modifiers like adjectives or premodifying nouns and complex modifiers like prepositional phrases or relative clauses. Table 6.1 gives the relevant figures for all 43 samples plus the average numbers for the same section in all papers and the same paper across all sections. Table 6.1. Number of modifiers per 1000 NPs per section and newspaper arts

business

foreign news

home news

sports

all sections

420 411 381 404.00

344 344 374 354.00

382.00 377.50 410.33 392.86

DMi Star Sun all down

476 476.00

DE DMa Ίο all mid

511 485 460 485.33

468 506 440 471.33

468 444 466 459.33

413 409 389 403.67

465.00 461.00 438.75 454.92

482

DT FT G I T all up

471 573 488 652 576 552.00

491 566 634 561 640 578.40

577 648 579 607 549 592.00

579 619 614 448 527 557.40

407 474 470 458.25

520.00 601.50 544.40 548.40 552.40 551.33

all papers

552.00

536.00

546.75

488.82

410.60

498.63

Average figures for individual papers and individual sections across types of papers and for all papers.

There are obvious differences between the three types of paper. The last column of table 6.1 gives the average number of modifiers for each individual paper across all the sections available for that particular paper. The averages for the down-market papers are about 400, for the mid-market papers about 450 and for the up-market papers about 550. The three groups are quite clearly separated, and there is no overlap between the ranges of the three groups if only the averages across all sections are considered. On the other hand, there is some overlap if the figures for the individual sections are compared. The

The density of modifiers

109

figure for the home news sample of The Independent, for instance, is considerably lower than two of the mid-market samples. A 1 -way ANOVA test shows whether the variation within groups is significantly smaller than between groups (for an outline of the statistical tests used in this analysis cf. any of the increasing number of textbooks on statistics in linguistics, e.g. Butler 1985; Hatch and Farhady 1982; or Woods et al. 1986). In the above example, it establishes whether the variation across all the samples within the individual newspaper types is smaller than across all samples of all newspaper types. The null-hypothesis says that the difference between the within group variation and the between group variation is accidental and must be attributed to chance. The ANOVA gives the probability for the nullhypothesis being true. If it is smaller than 5%, i.e. p < 0.05, the nullhypothesis is sufficiently implausible and has to be rejected, even though there is of course still a residual chance that it is actually true. It can never be excluded with absolute certainty. In the above example this probability is actually smaller than 0.1% (p < 0.001), and thus the newspaper category appears to be a significant factor for the number of modifiers that are used for a given number of noun phrases (here 1000). If the newspaper section is taken as the independent factor, the 1-way ANOVA again reveals that the null-hypothesis must be rejected (p < 0.001). It appears that different subject matters require different densities of modifiers. The density is smallest in the sports section. This is true not only of the corpus as a whole but also of each of the three newspaper categories. The arts section, which only exists in the up-market papers, has the highest average both for the entire sample and for the up-markets alone. The former fact might of course be concomitant with the fact that only up-market papers, which have got higher modifier densities anyway, contribute to its overall average. But the second fact, that it has the highest average among the averages for the up-markets only, suggests that it is indeed the subject matter of the arts pages which seems to require a particularly high density of modifiers. In chapter 4 above, I pointed out that postmodifiers tend to be more explicit than premodifiers because they indicate explicitly the kind of relationship that holds between the head noun and the modifiers, either by means of prepositions or by means of entire finite or infinite clauses. It therefore seems plausible to expect that there will be a significant difference in the ratio of premodifiers versus postmodifiers that are used in the different categories of paper and in the different sections. Table 6.2 shows the percentage of all modifiers for each sample that are premodifiers. Table 6.2 shows that there is surprisingly little variation both across types of newspaper and across sections. The percentages of premodifiers for most

110

Distribution and complexity

Table 6.2. Percentage of modifiers that are premodifiers

% premodifiers

arts

business

foreign news

all sections

home news

sports

54.76 47.69 55.91 52.72

53.78 54.36 57.49 55.27

54.32 50.73 56.78 54.44

DMi Star Sun all down

56.93 56.93

DE DMa To all mid

56.36 57.11 56.96 56.80

52.35 53.16 48.64 51.49

56.20 54.73 54.29 55.08

46.97 49.63 53.47 49.96

53.23 53.80 53.39 53.47

53.73

DT FT G I T all up

51.38 54.80 50.61 5460 57.29 53.95

57.43 48.41 59.62 56.68 55.47 55.57

53.03 56.33 52.68 53.38 51.00 53.38

55.61 52.83 53.26 48.66 50.47 52.39

48.65 49.58 48.51 50.19

54.27 53.20 53.45 52.92 52.82 53.33

all papers

53.95

56.07

52.77

53.15

51.44

53.50

For the absolute figures of modifiers cf. table 6.1, above.

individual samples and for all aggregate categories are slightly higher than 50. In this case the null-hypothesis cannot be rejected for either factor (newspaper category or section). Hence, on this level of delicacy, there is no discernible difference in the use of premodifiers versus postmodifiers. However, I shall have a closer look at individual types of premodifier and postmodifier in chapters 7 and 8 respectively.

6.3. Multiple modifications and modifier embedding The next step is an analysis of the complexity of the modifiers that are used in the various sections of the different newspapers. There are at least two ways of measuring this complexity. Noun phrases can be modified by one or several individual modifiers, and it can easily be established how many noun phrases in each sample have just one modifier, two modifiers, or three modifiers and so on. Modifiers which modify the same head noun, whether they are premodifiers or postmodifiers, I call concatenated modifiers. But noun phrases are also regularly modified by modifiers containing noun phrases which are themselves modified. These I call embedded modifiers.

Multiple modifications and modifier embedding

111

Table 6.3 gives a summary of the numbers of noun phrases with varying numbers of concatenated modifiers in the individual sections across the three newspaper categories. Table 6.3. Average number of noun phrases with one, two, three or four and more concatenated modifiers newspaper category

section

average 1

number of 2

NPs with 3

4 + modifiers

down

bu hn sp total

233.0 234.0 185.7 213.1

97.0 64.7 70.3 71.7

18.0 15.3 10.0 13.4

1.0 2.3 1.7 1.9

mid

bu fn hn sp total

253.3 264.7 243.0 219.7 245.2

81.0 86.0 86.3 72.7 81.5

22.3 13.7 16.3 12.7 16.3

3.7 2.3 3.0 2.7 2.9

up

ar bu fn hn sp total

257.8 272.8 285.8 297.2 257.0 274.8

96.8 106.2 104.4 100.2 77.5 97.8

27.6 28.2 29.2 18.4 17.0 24.4

7.8 5.4 4.2 3.4 1.3 4.5

256.5

89.0

20.3

3.7

all papers

Figures given for all sections across newspaper categories and for the aggregate sample of all newspapers per category (detailed figures for all samples are given in Appendix III, table A6.1).

Table 6.3 reinforces the findings from the previous section in that the upmarket papers generally use more modifiers than the mid-markets, and the mid-markets use more than the down-markets. This is true not just of the noun phrases with one single modifier but also of noun phrases with two, three or more concatenated modifiers. Within the newspaper categories, there is considerable variation between the sections, but in all instances the sports section has the smallest number of modifiers. Indeed the comparison holds good not just for the newspaper categories as a whole but also for the individual sections, that is to say the sports section for instance, in spite of the fact that it has fewer modifiers than any other section within the same category, has got the highest number in the up-markets and the lowest number in the down-markets. There are very few exceptions that do not fit this pattern. For instance, the

112

Distribution and complexity

business section in the down-markets has more noun phrases with two modifiers than the business section in the mid-markets, and the home news section of the down-markets has got more noun phrases with three modifiers than the home news sections of the mid-markets. The sports section is quite remarkable in this respect. In the up-market papers, the number of noun phrases with one modifier is considerably lower than in the foreign news section of the mid-markets, and the numbers of noun phrases with more than one modifier are lower than several corresponding figures in the mid-markets with the exception of the sports section. A very similar picture obtains when one compares the sports section in the mid-market papers and the down-market papers. A second way of analysing the complexity of noun phrase modifications is to establish the level of embedding that occurs for each sample. Table 6.4 gives the number of noun phrases that occur not as the heads of arguments in a clause structure but as a modifier. If a noun phrase modifies a head of a clause argument, it is considered to be embedded at level one. The noun phrases modifying level one embedded noun phrases are at level two, and so on. This coding structure unfortunately fails to take into account embeddings within finite verbal postmodifiers because they have an argument structure of their own. Noun phrases occurring in such clauses are considered to be heads of arguments unless of course there are further embeddings within such a clause (cf. appendix I for the details of the coding scheme). Chapter 8 below, which looks in detail at postmodifiers, contains numerous examples that demonstrate the great complexity possible in this respect. This perspective offers a fairly similar picture to the one given in table 6.3 above. The up-markets have the highest number of noun phrases that are embedded within other noun phrase modifiers, and they have more of them at deeper levels of embedding. The down-markets have the smallest number, and the mid-markets again hold the middle ground. If individual sections are compared across the three newspaper categories, the pattern is still perfect in most cases. The business section in the down-markets and in the mid-markets is again an exception in that the values for the former are higher than for the latter, but the values for the business section of the down-markets is not an average because there is only one down-market newspaper with a regular business section, viz. The Sun. In all three categories it is again the sports section which has the smallest number of embedded noun phrases. Indeed in the sports section of the upmarket papers the values for the first two levels of embedding are even smaller than the corresponding values of all sections of the mid-market papers with the exception of the sports section. The values for level three, and for level

Multiple modifications and modifier embedding

113

Table 6.4. Average number of noun phrases embedded within modifiers newspaper newspaper category

section

down

bu hn sp total

242.0 175.0 159.7 178.0

55.0 38.0 24.7 34.7

6. 8.3 4.3 6.3

mid

bu fn hn sp total

233.3 205.3 203.0 179.3 205.3

49.7 44.3 48.7 40.0 45.7

9.3 7.7 7.7 6.3 7.8

2.5 1.0 2.0 1.5 1.8

up

ar bu fn hn sp total

202.4 258.8 230.4 235.6 183.0 223.7

50.2 77.6 63.8 74.6 39.8 62.1

11.8 15.2 16.6 14.0 7.8 13.3

3.6 4.0 3.6 3.3 3.3 3.6

211.1

53.1

10.6

3.3

all papers

average number of NPs at level 1 2 3

4 + embedding

Figures indicate average number of noun phrases embedded within modifiers (excluding finite verbal postmodifiers) at the indicated level of embedding for all sections across newspaper categories and for the aggregate sample of all newspapers per category (detailed figures for all samples are given in Appendix HI, table A6.2).

four and deeper, however, are slightly higher than the corresponding values for the other sections in the mid-markets. This means that the sports section of the up-market papers, in spite of the low number of embedded modifiers, is slightly more likely to have deep embeddings than the mid-market papers. The down-market papers are, of course, most reluctant to use deep embeddings. Anything beyond level three appears to be categorically avoided. In my entire down-market corpus, amounting to 7,000 noun phrases, there is not one single instance of a level four embedding. In fact only 19.6 per cent of the noun phrases embedded at level one are further modified by another noun phrase, i.e. a level two embedding. In the mid-markets this percentage rises to 22.3 and in the up-markets to 27.8. Hence the higher figures for deeper embeddings in the mid- and up-markets are not just a result of the higher overall number of modifiers. They reflect a genuinely higher tendency to use embedded modifiers.

114

Distribution and complexity

This tendency also works for individual sections across the newspaper categories. For the home news section, for instance, the percentage of level one embeddings that are further modified by a noun phrase is 21.7 for the downmarkets, 24.0 for the mid-markets and 31.7 for the up-markets. For the foreign news section it is 21.5 in the mid-markets and 27.6 in the up-markets. But again the pattern is not perfect for all the sections. The business section of The Sun is higher than the business section of the mid-markets, 22.7 as compared to 21.3. The business section of the up-markets is considerably higher than both of them with 30.0 per cent. The sports section, too, deviates slightly from the perfect pattern. The percentage of level one embeddings with further embeddings is as low as 15.5 in the down-markets. In the mid-markets it is 22.3, but in the up-markets it is slightly lower again with 21.7. To summarise this section, there is a steady increase in the number of concatenated modifiers and embedded modifiers from the down-market papers through the mid-markets to the up-markets. The up-market papers use more modified noun phrases and they are more likely than the other two categories to use more than one modifier. Similarly the up-markets use more modifiers that have further noun phrases embedded within them than the other two categories, and they are more likely to use multiple embeddings. The different sections may vary considerably if they are compared within one newspaper category, but if one section is compared across the newspaper categories, it generally follows the pattern of the newspaper categories as a whole. The sports section stands out from all the others because in all newspaper categories it is less complex than any of the other sections. It has fewer noun phrases with concatenated modifiers, and it has fewer embedded modifiers and even fewer multiple embeddings.

6.4. Noun phrase complexity in the three newspaper categories In the following I shall try to refine Quirk et al.'s analysis of the noun phrase as a style marker, that is to say to provide a slightly less "coarse grained" analysis. The size of my sample is about two and a half times as large as theirs but it covers a far smaller stylistic range. It is my hypothesis that the same variability that Quirk et al. demonstrate across their disparate range of styles also exists across the very narrow range of styles exhibited by different newspapers in their different sections. As my working hypothesis I postulate that the directional variability that exists from informal speech through fiction and serious talk to scientific writing will be paralleled as directional variability from the down-market papers through the mid-market papers to the up-market papers,

Noun phrase complexity in the three newspaper categories

115

that is to say that features that occur with ever increasing (or decreasing) frequency in the four styles of Quirk et al.'s sample will also occur with ever increasing (or decreasing) frequency in the three categories of newspapers, where the down-markets correspond to the less formal and the up-markets to the more formal end of the scale. Table 6.5. Noun phrase structure and distribution in the British daily newspapers Total

all simple NPs

names & pronouns

all complex NPs

multiple modification

all papers

subject % not subject % total %

16131 37.51 26869 62.49 43000 100.00

11767 27.37 15369 35.74 27136 63.11

8465 19.69 6622 15.40 15087 35.09

4364 10.15 11500 26.74 15864 36.89

1317 3.06 3523 8.19 4840 11.26

downmarket

subject % not subject % total %

2757 39.39 4243 60.61 7000 100.00

2203 31.47 2696 38.51 4899 69.99

1753 25.04 1188 16.97 2941 42.01

554 7.91 1547 22.10 2101 30.01

170 2.43 439 6.27 609 8.70

midmarket

subject % not subject % total %

4672 38.93 7328 61.07 12000 100.00

3508 29.23 4342 36.18 7850 65.42

2580 21.50 1891 15.76 4471 37.26

1164 9.70 2986 24.88 4150 34.58

339 2.83 869 7.24 1208 10.07

upmarket

subject % not subject % total %

8702 36.26 15298 63.74 24000 100.00

6056 25.23 8331 34.71 14387 59.95

4132 17.22 3543 14.76 7675 31.98

2646 11.03 6967 29.03 9613 40.05

808 3.37 2215 9.23 3023 12.60

Samples are aggregates of all the sections analysed.

Table 6.5 gives an overview of the entire sample 43,000 noun phrases drawn from five sections of eleven British daily newspapers. There are 1000 nouns per section per newspaper. But the down-market papers and the mid-market papers do not have as many sections as the up-markets, and therefore, strictly speaking, the up-market papers are over- represendted because they contribute

116

Distribution and complexity

24 sections, the mid-markets 12 and the down-markets seven (for details see chapter 1). It is of course unclear in any case, what distribution would give an unbiased picture of newspaper language. Would it suffice to use the same number of noun phrases from every paper, or should the fact that there are five upmarkets but only three down-markets and three mid-markets, be taken into consideration? It may even be argued that the samples should be represented in proportion to the circulation figures of the individual papers. This would certainly give the down-market papers a far stronger influence on the aggregate results. If table 6.5 is compared to Quirk et al.'s results reproduced as table 5.3 in the preceding chapter, the following points emerge. The percentage of modified nouns for the whole sample is somewhat higher than the corresponding percentage in Quirk et α/.'s whole sample. It lies between the percentages for serious talk and scientific writing, that is to say towards the more formal end of their stylistic range but still well within it. The percentage of nouns that are modified more than just once, on the other hand, is slightly lower than the corresponding average for Quirk et al.'s whole sample. It lies between the two styles of fiction and serious talk. However, it is not clear whether these figures can be compared directly. I have spelt out my coding scheme in appendix I. It may well be that Aarts (1971), on whom Quirk et al. (1985) based their section on the stylistic impact of noun phrase complexity, used a different scheme. It is unclear, for instance, whether he included noun phrases that are embedded in other noun phrases in his analysis (as I do), or whether he restricted his analysis to the topmost level (as Raumolin-Brunberg (1991) does in her analysis of the noun phrase complexity in the writings of Sir Thomas More). The three categories of newspapers pattern very much in the way one would expect. The down-market papers have the smallest percentage of modified nouns and the smallest percentage of nouns modified more than once, whereas the up-market papers have the highest percentage for both these counts. The difference for the modified nouns is quite considerable. It is roughly five percentage points both between the down-markets and the midmarkets, and between the mid-markets and the up-markets. Just over one-third of the noun phrases in the whole sample of newspaper language are names and pronouns. This is considerably less than in Quirk et al.'s whole sample, in which the names and pronouns comprise almost onehalf of all noun phrases, but it is still within the range of styles contained in their sample. It lies again between the styles of serious talk and scientific writing.

Noun phrase complexity in the three newspaper categories

117

Across the three categories of newspapers, the predicted pattern emerges. The ratio of names and pronouns is highest for the down-market papers and smallest for the up-market papers. The difference between the down-markets and the mid-markets, and the difference between the mid-markets and the upmarkets are both roughly five percentage points. The situation is less clear-cut for the association of simple noun phrases with subjects and complex noun phrases with non-subjects. Quirk et al. report that this association is greatest for the more informal styles and smallest for the more formal styles. I suggested above that this association may be quantified for comparative purposes by juxtaposing the percentage of all simple nouns that appear in subject position to the percentage of all complex nouns that appear in subject position. For informal speech in Quirk et al. 's sample the difference turned out to be 48.4 percentage points and in scientific writing 25.2 percentage points. In my sample of newspaper language this association is considerably smaller. 43.4 per cent of all the simple nouns phrases and 27.5 per cent of all the complex noun phrases appear in subject position (as compared to 58.7 and 21.1 per cent respectively in Quirk et al.'s sample as a whole). The difference in my sample is thus only 15.9 percentage points. However, Quirk et al.'s observation that the association is smaller in more formal contexts is exactly paralleled in my data. The difference between unmodified and modified noun phrases in subject position is 18.6 percentage points for the down-markets, 16.6 percentage points for the mid-markets, and 14.6 percentage points for the up-markets. It is, however, unclear why the association between simple noun phrases with subject position and complex noun phrases with nonsubject positions should be so much smaller than in Quirk et al. 's sample. This value is clearly outside of the range of the four styles analysed by Quirk et al. Almost three quarters of the names and pronouns in all four styles appear in subject position. In my data, this value, too, is lower. For the sample as a whole, 56.1 per cent of all names and pronouns are in subject position. The values for the three newspaper categories are 59.6, 57.7 and 53.8 per cent. As expected the percentage is highest for the down-markets and smallest for the up-markets, even though it is considerably lower in all three cases than for any of the styles in Quirk et al.'s sample. A chi-square test which compares the four basic values of table 6.5 above in the three newspaper categories (i.e. modified versus unmodified noun phrases in subject or nonsubject position) reveals that the differences are statistically significant (χ2 = 293.84, df = 6, ρ < 0.001). This means of course that the figures for the entire sample cannot be taken to be valid for newspaper language in general. The up-market papers have a far greater share of the

118

Distribution and complexity

whole sample than either the down-markets or the mid-markets. The figures for the whole sample therefore are skewed towards the up-market papers. To summarise the results so far, the prediction is clearly borne out that the variability across the four styles of Quirk et /.'s sample is paralleled within the far smaller range of styles in the data of newspaper language. The sample as a whole, albeit not truly representative of newspaper language in general because of its bias towards up-market papers resembles most closely Quirk et ίζ/.'s style of serious talk. Features that increase in their frequency from informal speech to scientific writing in Quirk et al. 's corpus increase their frequency from the down-market papers across the mid-markets to the up-markets. This supports both the claim that the down-markets represent a less formal vaTable 6.6. Noun phrase structure and distribution in the three down-market papers Total

all simple NPs

names & pronouns

all complex NPs

multiple modification

downmarket

subject % not subject % total %

2757 39.39 4243 60.61 7000 100.00

2203 31.47 2696 38.51 4899 69.99

1753 25.04 1188 16.97 2941 42.01

554 7.91 1547 22.10 2101 30.01

170 2.43 439 6.27 609 8.70

Daily Mirror

subject % not subject % total %

783 39.15 1217 60.85 2000 100.00

631 31.55 786 39.30 1417 70.85

519 25.95 394 19.70 913 45.65

152 7.60 431 21.55 583 29.15

38 1.90 128 6.40 166 8.30

The Star

subject % not subject % total %

822 41.10 1178 58.90 2000 100.00

659 32.95 741 37.05 1400 70.00

506 25.30 330 16.50 836 41.80

163 8.15 437 21.85 600 30.00

51 2.55 106 5.30 157 7.85

The Sun

subject % not subject % total %

1152 38.40 1848 61.60 3000 100.00

913 30.43 1169 38.97 2082 69.40

728 24.27 464 15.47 1192 39.73

239 7.97 679 22.63 918 30.60

81 2.70 205 6.83 286 9.53

Samples are aggregates of all the sections analysed.

Noun phrase complexity in the three newspaper categories

119

riety than the up-markets, and the claim that there is indeed a discernible difference between the down-market papers and the mid-market papers. The next step to be taken is to ascertain how much variability there is within the three categories and how much overlap there is between the categories. Without this step, any claims that the three categories are distinct would remain somewhat dubious. For this purpose it is necessary to present the tables for all the individual papers. Table 6.6 shows the results for the three down-market papers, the Daily Mirror, The Star and The Sun. In this and the two subsequent tables, the papers are ordered according to the percentage of modified nouns. This is the feature that stratifies the three categories of papers most clearly, and therefore it seems reasonable to expect that it is also the feature that might show up any variation within each category. The figures for the individual papers consist of samples from the home news section and the sports section for all three papers, and for The Sun they also include the business section. Most values are almost identical for the three papers. The variation in the percentage of modified noun phrases is less than one and a half percentage points across the three papers, and for the noun phrases that are modified more than once, it is less than two percentage points. There is some discernible variation for the number of names and pronouns. The Daily Mirror clearly has the highest percentage, while The Star and The Sun have got somewhat lower figures, but the overall variation is still not greater than six percentage points. There is also some slight variation for the association of simple noun phrases with subject position and complex noun phrases with nonsubject position. The percentage of unmodified noun phrases in subject position varies from 43.9 per cent to 47.1 per cent, and the percentage of modified noun phrases in nonsubject position from 29.2 to 30.6. To quantify the strength or weakness of the association, the differences in percentage points were established. According to the prediction made on the basis of Quirk et al.'s data and the overall variation across the three categories of papers, the association should be strongest for the most informal style and weakest for the most formal style. However, the difference between the three down-market papers is less than two percentage points. The respective figures are 18.5 for the Daily Mirror, 19.9 for The Star and 17.8 for The Sun. A chi-square test, applied to the four key figures of unmodified versus modified noun phrases in either subject or nonsubject position, reveals that the three samples are indeed so similar that the null-hypothesis cannot be rejected. There is no variation at a level of statistic significance (χ2 = 4.85, df = 6, p: not significant). Thus the three down-market papers can quite confidently be

120

Distribution and complexity

Table 6.7. Noun phrase structure and distribution in the three mid-market papers Total

all simple NPs

names & pronouns

all complex NPs

multiple modification

midmarket

subject % not subject % total %

4672 38.93 7328 61.07 12000 100.00

3508 29.23 4342 36.18 7850 65.42

2580 21.50 1891 15.76 4471 37.26

1164 9.70 2986 24.88 4150 34.58

339 2.83 869 7.24 1208 10.07

Today

subject % not subject % total %

1565 39.13 2435 60.88 4000 100.00

1182 29.55 1469 36.73 2651 66.28

840 21.00 615 15.38 1455 36.38

383 9.58 966 24.15 1349 33.73

113 2.83 274 6.85 387 9.68

Daily Express

subject % not subject % total %

1583 39.58 2417 60.43 4000 100.00

1177 29.43 1436 35.90 2613 65.33

881 22.03 652 16.30 1533 38.33

406 10.15 981 24.53 1387 34.68

127 3.18 310 7.75 437 10.93

Daily Mail

subject % not subject % total %

1524 38.10 2476 61.90 4000 100.00

1149 28.73 1437 35.93 2586 64.65

859 21.48 624 15.60 1483 37.08

375 9.38 1039 25.98 1414 35.35

99 2.48 285 7.13 384 9.60

Samples are aggregates of all the sections analysed.

treated as one single style, even though there still may be some systematic variation across the different sections of these papers. Table 6.7 gives the results for the three mid-market papers, Daily Express, Daily Mail and Today. They are again ordered according to the percentage of their modified noun phrases. For all the papers, four sections were analysed: business, foreign news, home news, and sports. The variation among these three papers is, if anything, even smaller. They do not vary for more than two percentage points either in the percentage of modified noun phrases or the percentage of noun phrases that are modified more than once. The percentages of names and pronouns, too, are almost identical for the three papers.

Noun phrase complexity in the three newspaper categories

121

The association between simple noun phrases and subject position and between complex noun phrases and nonsubject position is also almost identical. The percentage of unmodified noun phrases in subject position varies from 44.4 to 45.0 and the percentage of modified noun phrases in subject position from 26.5 to 29.3. The difference between the two values is 16.2 for Today, 15.8 for the Daily Express and 17.9 for the Daily Mail. The chi-square test confirms that there is no difference at a level of statistic significance between the papers as far as the four key values are concerned (χ2 = 5.11, df = 6, p: not significant). This means that again the figures for the mid-market papers as a whole are representative of all three papers. The difference between the three down-market papers and the three midmarket papers, on the other hand, is fairly clear-cut. Today, which has the lowest percentage of modified noun phrases among the mid-markets, has still got three percentage points more than The Sun, the down-market paper with the highest percentage of modified noun phrases. There is also no overlap as far as the percentage of multiple modification is concerned. The down-market paper with the highest percentage is again The Sun with 9.53 per cent of noun phrases that are modified more than once. The mid-market with the lowest percentage is the Daily Mail with 9.60. This difference is admittedly trivial but it suffices to make the point that there is no overlap between the downmarket and the mid-market papers. For the names and pronouns it is the down-markets that have the higher scores. Again, there is no overlap. The Sun with the smallest percentage among the down-markets has still got roughly one and a half percentage points more than the Daily Express, which has got the highest score among the mid-markets. So far, then, the pattern is absolutely perfect. There is very little variation among the papers of one category, and none at all that is statistically significant, but there is quite a considerable difference between the two categories as a whole, and moreover there is no overlap of the relevant values between the papers across the two categories. It is again the association of simple noun-phrases with subject-position and complex noun phrases with nonsubject position that patterns in a less clear-cut way. According to the prediction, the difference between simple noun phrases in subject position and complex noun phrases in subject position should be greatest in the down-market papers and smaller in the mid-market papers. This prediction is borne out for the categories as a whole. It is 18.6 percentage points in the down-markets and 16.6 in the mid-markets, but there is some small overlap between the individual papers. The Sun has the weakest association of simple noun phrases with subject position and complex noun phrases

122

Distribution and complexity

Table 6.8. Noun phrase structure and distribution in the five up-market papers Total

all simple NPs

names & pronouns

all complex NPs

multiple modification

up market

subject % not subject % total %

8702 36.26 15298 63.74 24000 100.00

6056 25.23 8331 34.71 14387 59.95

4132 17.22 3543 14.76 7675 31.98

2646 11.03 6967 29.03 9613 40.05

808 3.37 2215 9.23 3023 12.60

The Daily Telegraph

subject % not subject % total %

1697 33.94 3303 66.06 5000 100.00

1150 23.00 1896 37.92 3046 60.92

811 16.22 780 15.60 1591 31.82

547 10.94 1407 28.14 1954 39.08

158 3.16 420 8.40 578 11.56

The Guardian

subject % not subject % total %

1813 36.26 3187 63.74 5000 100.00

1322 26.44 1713 34.26 3035 60.70

952 19.04 769 15.38 1721 34.42

491 9.82 1474 29.48 1965 39.30

158 3.16 483 9.66 641 12.82

The Times

subject % not subject % total % .

1840 36.80 3160 63.20 5000 100.00

1263 25.26 1727 34.54 2990 59.80

819 16.38 722 14.44 1541 30.82

577 11.54 1433 28.66 2010 40.20

167 3.34 444 8.88 611 12.22

The Independent

subject % not subject % total %

1932 38.64 3068 61.36 5000 100.00

1344 26.88 1645 32.90 2989 59.78

908 18.16 730 14.60 1638 32.76

588 11.76 1423 28.46 2011 40.22

176 3.52 464 9.28 640 12.80

Financial Times

subject % not subject % total %

1420 35.50 2580 64.50 4000 100.00

977 24.43 1350 33.75 2327 58.18

642 16.05 542 13.55 1184 29.60

443 11.08 1230 30.75 1673 41.83

149 3.73 404 10.10 553 13.83

Samples are aggregates of all the sections analysed.

Noun phrase complexity in the three newspaper categories

123

with object position among the down-markets. Its difference between simple noun phrases in subject position and complex noun phrases in subject position is 17.8 percentage points. The corresponding value for the Daily Mail, for which this association is the strongest among the mid-markets, is 17.91 and thus slightly higher than for The Sun. In table 6.8, a distinctly less coherent picture emerges for the five upmarket papers, The Daily Telegraph, the Financial Times, The Guardian, The Independent and The Times. For these papers five sections were analysed, arts, business, foreign news, home news and sports. The only exception is the Financial Times, which does not have a sports section. There is very little variation in the percentages of modified noun phrases between the five up-market papers. It ranges from 39.1 per cent for The Daily Telegraph to 41.9 per cent for the Financial Times. In contrast, the midmarket with the highest percentage of modified noun phrases, the Daily Mail, has only got 35.4 per cent. For multiple modification, too, there is not too much difference among the up-markets. It ranges from 11.6 per cent in The Daily Telegraph to 13.8 per cent in the Financial Times. There is again no overlap with the mid-markets, the one with the highest percentage, the Daily Express, reaching only 10.9 per cent. The percentages for names and pronouns vary slightly more. They range from 34.4 per cent in The Guardian to 29.6 per cent in the Financial Times. This compares with the lowest percentage in a mid-market paper of 36.4 per cent. Thus there is again no overlap between the up-markets and the midmarkets. However, if the association of simple noun phrases with subject position and complex noun phrases with nonsubject position is considered, a far less coherent picture emerges. There is some considerable variation between the percentages of unmodified noun phrases in subject position (from 37.8 per cent to 45.0 per cent) as well as between the percentages of modified nouns in subject position (from 25.0 to 29.2). The difference between the percentages of unmodified noun phrases in subject position and modified noun phrases in subject position, is very considerable indeed. According to the prediction it should be greatest in the least formal style and smallest in the most formal style. For the combined sample of all up-market papers it is indeed lower than for the mid-market papers (14.6 percentage points as compared to 16.6 percentage points), but the individual papers do not pattern according to the prediction. The Daily Telegraph, which appeared on the top of table 6.8 because it has the smallest percentage of modified noun phrases among the up-markets and thus is nearest to the mid-markets, has got by far the weakest association of

124

Distribution and complexity

simple noun phrases with subject position and complex noun phrases with nonsubject position. The difference between the unmodified noun phrases in subject position and the modified noun phrases in subject position is only 9.8 percentage points. For The Times it is 13.5 percentage points, which is near the average for this newspaper category as a whole. The Financial Times and The Independent have got differences of 15.5 and 15.7 percentage points respectively. These values are very close to the lowest mid-market value of 15.8 in the Daily Express. The highest up-market value, however, is considerably higher than that. The Guardian has a difference of 18.6 percentage points, which is more than the highest mid-market value and indeed more than two of the down-market values. Measured in this way, it is only The Star that has got a stronger association of simple noun phrases with subject position and complex noun phrases with nonsubject position. If a chi-square test is applied to the four key figures of the five up-market papers (unmodified and modified noun phrases in subject and in nonsubject position), it turns out that the variation is statistically significant (χ2 = 56.21, df = 12, ρ < 0.001). However, as I have shown above, there is not very much variation in the percentages of modified versus unmodified noun phrases. And indeed if the chi-square test is applied to these two figures only (that is to say the total of unmodified nouns compared with the total of modified nouns) without considering the syntactic position, the variation is not statistically significant (χ2 = 8.33, df = 4, p: not significant). The variation is thus due solely to the differing degrees of association of simple noun phrases with subject position and complex noun phrases with nonsubject position. The Daily Telegraph and The Guardian are the two papers with exceptional differences between simple noun phrases in subject position and complex noun phrases in subject position, the former being particularly low and the latter particularly high. If the chi-square test is applied to the remaining three papers, but again taking into account the syntactic position, there is no difference of statistic significance, (χ2 = 12.92, df = 6, p: not significant) To summarise and conclude the present section, the complexity of noun phrases stratifies not only widely disparate styles such as those used in Quirk et /.'s (1985: 1350-1352) data, but also the very narrow range of newspaper language. The complexity was measured in very simple terms, distinguishing modified noun phrases, with a subgroup of those that have multiple modification, i.e. more than one modifier, from unmodified noun phrases. Among the unmodified noun phrases names and pronouns were singled out because of their numerical and distributional significance.

Noun phrase complexity in the five newspaper sections

125

Quirk et al. show that the percentage of modified noun phrases as well as the percentage of noun phrases that are modified more than once is increasingly higher through their four styles, informal speech, fiction, serious talk and scientific writing, and at the same time fewer names and pronouns are used. My data of newspaper language as a whole most closely resembles Quirk et al.'s style of serious talk, but if it is broken up into the three newspaper categories down-market, mid-market and up-market, the same stratification can be observed, with the down-markets having the lowest percentage of modified nouns and the highest percentage of names and pronouns. Quirk et al. also point out a strong tendency of simple noun phrases to occur in subject position and complex noun phrases to occur in nonsubject position. This tendency is strongest in informal speech and weakest in scientific writing. Likewise it is strongest in the down-market papers and weakest in the up-market papers. The variation across the three newspaper categories is statistically significant. The down-markets and the mid-markets, comprising three papers each, are very coherent categories with no internal variation of statistical significance. The up-market papers are also very coherent as far as the ratio between modified and unmodified noun phrases is concerned, but two papers differ from the remaining up-market papers either by having a stronger association of simple noun phrases with subject position and complex noun phrases with nonsubject position (The Guardian) or by having a weaker association (The Daily Telegraph). The three categories are so clearly distinct that there is virtually no overlap between them. It is only the association of simple noun phrases with subject position and complex noun phrases with nonsubject position that gives a slightly less coherent picture because of a vary small overlap between the down-markets with the mid-markets and because of The Guardian whose association of simple noun phrases with subject position and complex noun phrases with nonsubject position is only surpassed by one single down-market paper, The Star.

6.5. Noun phrase complexity in the five newspaper sections In the preceding section, the entire sample of newspaper language was broken down into the three categories, down-market, mid-market and up-market, and further into the individual papers, without, however, taking into consideration the various sections of the newspapers from which the data was drawn. It is the purpose of this section to establish whether the noun phrase complexity

126

Distribution and complexity

stratifies the individual sections in the same way that it stratifies the different newspaper categories. If the entire sample is broken down into the five sections, arts, business and finance, foreign news, home news and sports, without taking into consideration the different newspaper categories, there is again variation at a level of statistic significance. As above a chi-square test was applied to the four key figures of unmodified and modified noun phrases in subject and nonsubject position, in all the five sections (χ2 = 318.36, df = 12, ρ < 0.001). However, the variation is not very revealing. Not surprisingly, it is the home news section and the sports section that have the smallest percentages of modified nouns, and the sports section that has the highest percentage of names and pronouns, because it is only in these two sections that data from the down-market papers is included. The samples for the individual sections are therefore very unequal and make it necessary to search for variation across sections within the individual categories. In table 6.9, the figures are given for the aggregates of the samples taken from the home news section and the sports section of the three down-market papers. The business section of The Sun is not included in this table because of the lack of corresponding data from the other two down-markets papers. Table 6.9. Noun phrase structure and distribution in two sections of the down-market papers Total

all simple NPs

names & pronouns

all complex

complex modification

home news

subject % not subject % total %

1238 41.27 1762 58.73 3000 100.00

953 31.77 1098 36.60 2051 68.37

111 24.07 439 14.63 1161 38.70

285 9.50 664 22.13 949 31.63

68 2.27 179 5.97 247 8.23

sports

subject % not subject % total %

1160 38.67 1840 61.33 3000 100.00

981 32.70 1216 40.53 2197 73.23

847 28.23 619 20.63 1466 48.87

179 5.97 624 20.80 803 26.77

70 2.33 176 5.87 246 8.20

Samples are aggregates of all three down-market papers.

Noun phrase complexity in the five newspaper sections

127

There is a considerable difference between the two samples. All the features that distinguish the down-market papers from the other two categories are even more prominent in the sports section than they are in the home news section. Very few noun phrases are modified at all in both sections but in the sports section it is almost five percentage points fewer than in the home news section. More noun phrases are names or pronouns in the sports section, and the association of simple noun phrases with subject position and complex noun phrases with nonsubject position is stronger. In the home news section, 46.5 per cent of all unmodified noun phrases are in subject position and 30.0 per cent of all modified noun phrases (a difference of 16.5 percentage points), whereas the respective figures for the sports section are 44.7 and 22.3 per cent (difference of 22.4 percentage points). A chi-square test confirms that the difference between the two sections is statistically significant (χ2 = 31.23, df = 3, ρ < 0.001). The final test, then, has to ascertain whether there is still some variation within the samples of table 6.9, that is to say between the three home news samples and between the three sports samples. The result is that there is no variation of statistical significance (for the home news sections: χ2 = 6.34, df = 6, p: not significant, and for the sports section: χ2 = 1.96, df = 6, p: not significant). Thus we have isolated two distinct styles. The down-market home news style and the down-market sports style. These two styles are used without any significant variation in all the three papers of this category. In table 6.10, the figures are given for four sections as aggregates of the three mid-market papers. Among the four sections of the mid-market papers, the sports section clearly stands out. It has again a higher percentage of names and pronouns and it has the smallest percentages of modified noun phrases and of noun phrases with multiple modification. The other three sections, on the other hand, do not vary very much. The percentages of unmodified noun phrases are very similar and the association of simple noun phrases with subject position and complex noun phrases with nonsubject position is about equally strong. The difference between the percentages of unmodified noun phrases in subject position and modified noun phrases in subject position for business, foreign news and home news section is 14.1, 16.0 and 16.2 percentage points, respectively. This compares with a corresponding figure of 21.5 percentage points for the sports section. A chi-square test over all four sections reveals that there is a variation of statistic significance (χ2 = 45.64, df = 9, p < 0.001), but if the sports section is left out, there is no statistically significant variation (χ2 = 5.29, df = 6, p: not significant). The variations across the three papers in all four samples of table

128

Distribution and complexity

Table 6.10. Noun phrase structure and distribution in the four sections of the mid-market papers Total

all simple NPs

names & pronouns

all complex NPs

multiple modification

business

subject % not subject % total %

1170 39.00 1830 61.00 3000 100.00

846 28.20 1073 35.77 1919 63.97

562 18.73 352 11.73 914 30.47

324 10.80 757 25.23 1081 36.03

93 3.10 228 7.60 321 10.70

foreign news

subject % not subject % total %

1157 38.57 1843 61.43 3000 100.00

844 28.13 1056 35.20 1900 63.33

571 19.03 457 15.23 1028 34.27

313 10.43 787 26.23 1100 36.67

91 3.03 215 7.17 306 10.20

home news

subject % not subject % total %

1222 40.73 1778 59.27 3000 100.00

906 30.20 1048 34.93 1954 65.13

664 22.13 468 15.60 1132 37.73

316 10.53 730 24.33 1046 34.87

90 3.00 227 7.57 317 10.57

sports

subject % not subject % total %

1123 37.43 1877 62.57 3000 100.00

912 30.40 1165 38.83 2077 69.23

783 26.10 614 20.47 1397 46.57

211 7.03 712 23.73 923 30.77

65 2.17 199 6.63 264 8.80

Samples are aggregates of all three mid-market papers

6.10 above are not significant (business: χ2 = 6.15, df = 6, p: not significant foreign news: χ2 = 9.97, df = 6, p: not significant home news: χ2 = 2.57, df = 6, p: not significant sports: χ2 = 7.46, df = 6, p: not significant). As for the down-markets, there are two clearly distinct styles for the midmarkets, which are adhered to by all three papers of this category. There is the style of sport reporting which is in contrast to the style used in all other sections. For the up-markets, finally, the situation is far less coherent. Table 6.11 gives the figures for five sections, all of which comprise data from the five upmarket papers with the one exception of the sports news which does not in-

Noun phrase complexity in the five newspaper sections

129

Table 6.11. Noun phrase structure and distribution in the five sections of the up-market papers Total

all simple NPs

names & pronouns

all complex NPs

multiple modification

arts

subject % not subject % total %

1680 33.60 3320 66.40 5000 100.00

1269 25.38 1781 35.62 3050 61.00

910 18.20 843 16.86 1753 35.06

411 8.22 1539 30.78 1950 39.00

125 2.50 536 10.72 661 13.22

business

subject % not subject % total %

1782 35.64 3218 64.36 5000 100.00

1219 24.38 1743 34.86 2962 59.24

752 15.04 618 12.36 1370 27.40

563 11.26 1475 29.50 2038 40.76

188 3.76 492 9.84 680 13.60

foreign news

subject % not subject % total %

1866 37.32 3134 62.68 5000 100.00

1219 24.38 1663 33.26 2882 57.64

806 16.12 744 14.88 1550 31.00

647 12.94 1471 29.42 2118 42.36

203 4.06 486 9.72 689 13.78

home news

subject % not subject % total %

1983 39.66 3017 60.34 5000 100.00

1300 26.00 1604 32.08 2904 58.08

827 16.54 605 12.10 1432 28.64

683 13.66 1413 28.26 2096 41.92

200 4.00 410 8.20 610 12.20

sports

subject % not subject % total %

1391 34.78 2609 65.23 4000 100.00

1049 26.23 1540 38.50 2589 64.73

837 20.93 733 18.33 1570 39.25

342 8.55 1069 26.73 1411 35.28

92 2.30 291 7.28 383 9.58

Samples are aggregates of all five up-market papers. The sports section does not include data from the Financial Times.

elude data from the Financial Times. The five sections are the same as for the mid-markets (business, foreign news, home news and sports) plus one additional section (arts).

130

Distribution and complexity

As for the down-markets and the mid-markets, it is again the sports section that has the highest percentage of names and pronouns and the lowest percentage of modified noun phrases and of noun phrases with multiple modification. This style is followed immediately by the arts section, which has got a slightly higher percentage of modification and multiple modification and a smaller percentage of names and pronouns. The remaining three sections have got figures that are fairly similar. No clear pattern emerges. The foreign news section has got the highest percentage of modifications, whereas the business section has got the lowest percentage of names and pronouns. A comparison of the individual sections across the three newspaper categories shows that they all pattern in the predicted way. The percentages of modification and of multiple modification are always highest in the up-markets, lower in the mid-markets, and lowest in the down-markets for those sections that exist in this category at all. The percentage of names and pronouns are highest for the down-markets (or in the mid-markets for those sections that do not occur in the down-markets) and lowest for the up-markets. And the association of simple noun phrases with subject position and complex noun phrases with nonsubject position is strongest in the down-markets (or mid-markets for those sections that do not occur in the down-markets) and weakest in the upmarkets. The variation across the five sections proves to be too great to be attributable to chance. It is statistically significant (χ2 = 155.45, df = 12, ρ < 0.001). In the two sections that seem to stand out from the others, the sports section and the arts section, the variation is still statistically significant, even though the probability is now only just smaller than 0.1 per cent (χ2 = 22.57, df = 6, ρ < 0.001). If all five sections are compared with each other in a pairwise fashion, only two pairings show a variation that is not statistically significant. These are the business section and the foreign news section on the one hand (χ2 = 7.53, df = 3, p: not significant), and the foreign news and the home news section on the other (χ2 = 5.62, df = 3, p: not significant). Thus the foreign news section seems to be close enough to both the business section and the home news section so that there is no significant variation, but the business section and the home news section in turn show enough variation for it to be statistically significant (χ2 = 20.88, df = 3, p < 0.001). The last step again has to ascertain the degree of variation within the five samples given in table 6.11. The question to be asked is thus whether there is any significant variation between the samples that have been drawn from five papers for one particular section.

Noun phrase complexity in the five newspaper sections

131

The sports and the arts sections fall into two clearly distinct groups each. The sports section comprises three samples, from The Guardian, The Independent and The Times, that do not vary significantly among each other (for all three samples: χ2 = 5.38, df = 6, p: not significant). The sports sample taken from The Daily Telegraph, on the other hand, differs from each of the other samples at a level of statistic significance. In the arts section, it is The Independent that stands out from the rest. The other four papers do not vary significantly (χ2 = 15.94, df = 9, p: not significant). For both the home news section and the foreign news section the three papers, The Daily Telegraph, the Financial Times and The Guardian, form one style with no significant variation (foreign news: χ2 = 12.94, df = 6, p < 0.05, not significant at the 1 per cent level home news: χ2 = 6.52, df = 6, p: not significant), but in the foreign news sections there is an overlap with another group because The Guardian can also be grouped together with The Times (χ2 = 8.54, df = 3, p < 0.05 not significant at the 1 per cent level). In the business section, finally, there are again two overlapping groups that show no significant internal variation. The Daily Telegraph and the Financial Times form one group (χ2 = 6.56, df = 3, p: not significant). And the other group is formed by the Financial Times, The Guardian, The Independent, and The Times (χ2 = 21.43, df = 9, p < 0.02, not significant at the 1 per cent level). Thus there is a very strong contrast between the down-market and the midmarket papers on the one hand, and the up-market papers on the other. The former only distinguish two styles within their own category. For both the downmarkets and the mid-markets it is the sports section that differs from the other sections. But there is no significant variation between the individual papers of the same category. For the up-market papers, on the other hand, a more complex picture emerges. There is a considerable amount of statistically significant variation between the individual sections of the various papers. For all the sections some papers form groups using a style with no significant internal variation, but these groups vary from section to section and they sometimes overlap. To conclude, the complexity of noun phrases increases from the downmarkets through the mid-markets to the up-markets. The percentage of noun phrases that are names or pronouns at the same time decreases. And the association of simple noun phrases with subject position and complex noun phrases with nonsubject position is strongest in the down-markets and weakest in the up-markets. In all three categories, the sports section uses a style that is clearly distinct from the other sections in that the percentage of complex noun phrases is always lower than for that particular newspaper category as a whole, and the percentage of names and pronouns is higher than for the category as a whole.

132

Distribution and complexity

Furthermore, it is evident that there is not just a two-way category distinction between the broadsheet papers and the tabloids. The tabloids are made up of the two clearly distinct categories the down-markets and the mid-markets. The two tabloid categories both show a considerable amount of internal coherence. There is no significant variation between the papers of the same category. For the up-market papers, on the other hand, there is a lot of internal variation. They show a lot more stylistic diversity than the down-markets or the mid-markets.

6.6. Cluster analysis On the basis of a replication of Quirk et al.'s (1985) brief study of noun phrase complexity, I have tried to arrange the different samples of my corpus into groups of similar styles. This attempt, which proved more successful for the down-market papers and the mid-market papers than for the up-market papers, was based on a very small number of variables. In this section, I shall try to refine this grouping by using a more comprehensive range of variables. In the analysis of my data, a fairly large number of variables emerges. For every sample, i.e for every section analysed in one particular newspaper, there are counts for 21 categories of modifiers. These categories fall into seven superordinate categories, which in their turn fall into the two categories, premodifiers and postmodifiers. The categories and subcategories were expounded in some detail in chapter 4 above. They are summarised in appendix I. With the help of a cluster analysis, I want to show how these counts can be used to compare the individual samples, that is to say I want to show how the samples fall into clusters according to their overall similarity. A cluster analysis is a multivariate technique because it compares more than one variable per case. It must be recognised, however, that this technique is merely exploratory and descriptive. It may reveal interesting clusters in the data but it does not test hypotheses about possible independent factors. This technique is very sensitive to many fairly arbitrary choices that have to be made by the investigator. The selection of variables that are used is of course crucial. In the case of this investigation, it has to be decided, for instance, whether the 21 individual categories of modifiers are to be used or the seven major categories, or indeed a mixture of the two. I have decided for the last of these possibilities by using the counts for the premodifier categories precentral, central and postcentral, ignoring the subcategories of the latter two, the counts for the different categories of prehead modifiers, i.e. prehead adjectives, prehead nouns and prehead names, and the counts for the three catego-

Cluster analysis

133

ries of postmodifiers, finite verbal, infinite verbal and nonverbal postmodifiers, again without considering the subcategories. It seems reasonable to use the subcategories for the prehead modifiers because this is by far the most important group anyway and even its subcategories have higher counts than many of the other counts. Once it has been decided which variables are to be used, there are still a number of decisions to be taken which influence the result (cf. Woods et al. 1986: 260). It has to be decided how the distance or dissimilarity between cases is calculated. It is usual for cluster analysis to measure the distance or dissimilarity between cases rather than the similarity. I have used the squared Euclidean distance, according to which the distance between two cases (x and y) equals the sum of the squared differences between the values of the variables that are used for the clustering (Distance (x,y) = Zj(Xj-yj)2). The second decision that has to be made concerns the clustering algorithm, that is to say the way in which individual cases are integrated into clusters. It may be required that a case is sufficiently similar to all the members of a group (so called complete linkage or furthest neighbour), or it may suffice that a case is sufficiently like just one member of a group (single linkage or nearest neighbour). In this case the average linkage between groups was used, which is an intermediate requirement between single linkage and complete linkage. Both the measure for dissimilarity and the clustering algorithm that were used to produce the dendrogram in figure 6.1 are the default parameters in the standard SPSS-X computer package that was used for its calculation. Some skilful manipulation of the parameters will very often provide the desired result but for this reason I deliberately relied on the default parameters. One of the possibilities of representing the result of a cluster analysis is the form of a dendrogram as shown in figure 6.1. On the vertical axis, all the language samples are listed, and on the horizontal axis, the value of their dissimilarity is plotted using a scale from 0 to 25. In this way it is possible to see which samples appear to be "closest together". The business sections of Today, Daily Express and The Daily Telegraph appear to be closer to each other than to any other sample (dissimilarity = 1). Equally the business section and the home news section of the Financial Times are closer to each other than they are to any other sample even though in this instance the dissimilarity is greater (d = 3). The dendrogram shows how the samples fall into clusters of relative similarity. A cluster is defined by setting an arbitrary value for d and plotting a vertical line at this level. Every intersection with a horizontal line then defines as members of a cluster all those samples that can be reached by moving leftwards from the point of intersection.

134

Distribution and complexity

Rescaled Distance Cluster 5 +

0 1LdUCl ahpl

h

τχ - -

Qiin h i i

__

To hu

J

DF hu DTbu

Tn hn

. . —1 _

FT hu



10 •+ ·

15 •-+

20 +

2 •—+

FThn IV

Thii Π hn

J

Ibu a DTfn DThn _l G hn Tfn T hn DMi on

,

Stsp

,— ...

1

DMi hn Q i t n hn

—i ,.

ζιιη on

DMasp —ι -l— —ι—-

Gu >>μ "n Το sp Ill

—1

1 SD

-. ..

Tsp

—1

DF in DFfn

DMafn 1 St hn Ihn 1 — DMa hn DTsp —1 — DEhn DMa bu —1 Tn fn

II

,-

DTar Gar b FT nr

Tar a 1 ar 1

R fn 1 fn

FTfn Figure 6.1. Dendrogram of the 43 samples, based on squared Euclidean distance and average linkage between groups.

Cluster analysis

135

If the dissimilarity value d is set at 15, there are four clusters. They have been indicated in the left margin and numbered from bottom to top. If the value d is set at ten there are two more clusters, as indicated by the letters a and b in the clusters II and IV respectively. Cluster I is made up of three upmarket foreign news sections. This appears to be a fairly coherent cluster with no dissimilarity value greater than 3. Cluster II contains all five arts sections, which appear to be more like each other than like any other sample even though the dissimilarity values in this cluster are as high as 14. For this reason, this cluster is split into two clusters if the value d for cluster definition is set at 10. In that case the arts section of The Independent forms a cluster of its own and is set off against the four other arts sections. Cluster III is less coherent. It contains all down-market samples, with the exception of the business section of The Sun; it contains most of the midmarket sections with the exception of two business sections and it contains all the sports sections. Cluster IV, finally, contains the up-market home news, business and foreign news sections two mid-market business sections and the Sun business section. There are a few exceptions, the home news section of Today appears in cluster IV rather than III, whereas the home news section of The Independent appears in III rather than in IV. A dissimilarity value of d = 10 for cluster definition again splits this cluster into two clusters, the first of which contains only up-market sections, whereas the second contains the Sun business section and all three mid-market sections of cluster IV besides one single up-market section, the business section of The Daily Telegraph. Table 6.12 summarises the pattern across these clusters. Table 6.12. Five styles according to the cluster analysis Cluster

Classes of sections

Exceptions

Cluster IV b Cluster IV a Cluster III

mid bu, down bu up bu, fn, hn down sp, hn mid fn, hn, sp upsp up ar up (FT, G, I) fn

+ To hn, DT bu

Cluster II a + b Cluster I

+ DMa bu + Ihn

Summary of the cluster analysis in figure 6.1 for a cluster definition value of d = 10.

As table 6.12 reveals, the clustering is not perfect in that not all the sections of the same newspaper category fall into the same cluster. The arts sections of the up-market papers, even though they form one single cluster at d = 15, fall into two clusters at d = 10. The foreign news sections of the up-markets are

136

Distribution and complexity

split up between cluster I and IVa. The more idiosyncratic exceptions are noted in the right-hand column of table 6.12. It must be stressed again that a cluster analysis is merely exploratory and descriptive. It is very sensitive to many fairly arbitrary decisions as outlined above. Different measures for distance, different cluster algorithms, and different linkage requirements produce what may at first sight appear to be very different dendrograms. In the dendrogram reproduced above, I have used the default settings both for the measure of distance (squared Euclidean measure) and for the clustering algorithm (average linkage between groups). However, several features that are apparent in the dendrogram reproduced as figure 6.1 recur in all the other possible dendrograms I have checked through. The following points appear to be most salient. The arts sections form a cluster of their own. These sections are closer to each other than any of them is to any other sample. The up-market sports sections cluster together with the midmarkets and the down-markets. The down-markets tend to be concentrated in one cluster with the exception of the business section of The Sun, whereas the mid-markets and the up-markets are more likely to be spread out over several clusters. The one section that always seems to end up in the "wrong" cluster is the home news section of The Independent. These observations cannot be taken to be results in themselves. But they suggest, among other things, that the newspaper section is just as important, if not more important, than the socio-economic categories of the papers to account for the stylistic variation across newspaper language, and also that a simple matrix of socio-economic category and newspaper section will not be enough to account for the interaction. The following two chapters will be devoted to a comprehensive illustration of the variables that were used for the above cluster analysis. I will have to assess in detail which variables are significant in the stratification of my data. I will have to show to what extent the styles suggested by the above cluster analysis help to explain the variation across the data, and I will have to provide the relevant data which either strengthens or refutes the hypotheses set up at the end of the previous chapter.

7. Premodifications

The level of delicacy aimed for in the analysis of modifiers is a very thorny problem. It is clear that any decision in this area will ultimately influence the results. If important distinctions are not made in the classification of the data, they can obviously not be used in the second stage of the analysis, the correlation of the linguistic structures with contexts of use. On the other hand, if the individual classes are too small, they may in fact obscure the relevant generalisations because the number of instances exemplifying each individual structure decreases as the number of categories increases. A structure illustrated by very few tokens might be interesting in itself if the point of interest is to show that such structures are actually used at a particular time in a particular context. But in a correlational study, in which generalisations are made about particular structures in particular contexts, it is essential that enough tokens are available for each structure, at least in some contexts, in order that its presence in these contexts and possible absence in others can be shown to be statistically significant. In the previous chapter, I have shown that the number of premodifiers in relation to all modifiers does not vary significantly across my corpus. Among the premodifiers, however, there are some important and statistically significant variations. The following figures for individual types of premodifiers are in each case given in relation to the total number of premodifiers or in relation to one of its major subcategories. This will ensure that differences in the overall number of modifiers in various samples do not falsify the results. A relatively rare type of premodifier in a sample with a high modifier density might turn out to be just as frequent as in a sample with low modifier density in which its relative share is considerably higher. The number of premodifiers in my samples can be deduced from tables 6.1 and 6.2 in the previous chapter, which give the overall numbers of modifiers and the relative share of premodifiers. For convenience the absolute numbers of the various types of premodifier are given in appendix III (table A7.1). In the following sections, I shall present the statistical results in some detail. The following tables will contain the percentage figures for all the major premodifier categories for each individual sample, and for all the relevant aggregate samples. Readers who are not interested in the statistical details are advised to skip section 7.1. The relevant findings will be summarised in section 7.2 before a stylistic interpretation of the statistical facts is attempted.

138

Premodifications

7.1. Types of premodifiers I have reviewed several attempts to correlate noun phrase structure with types of English. Not very many studies, however, have tried to do this for premodifiers in media English. One notable exception is Märdh (1980), who compares headlines in The Times and in the Daily Mirror. However, her results, as given in table 7.1 are not particularly revealing. Table 7.1. Frequency of open class premodifiers in nominal headlines of The Times and the Daily Mirror Type of open class premodifier

Daily Mirror

The Times

(a) Common noun (b) Proper noun (c) Noun with genitive 5 (d) Noun with plural s (e) Adjective (f) Participle (present or past) (g) Minor types

N 81 43 18 5 51 8 2

Totals

208

% 38.9 20.7 8.7 2.4 24.5 3.8 1.0

N 76 24 20 7 54 10 7

% 38.4 12.1 10.1 3.5 27.3 5.1 3.5

198

Source: Märdh (1980: 68, table 3).

The only difference of any significance seems to be in the number of proper nouns as premodifying items. They are more frequent in The Times than they are in the Daily Mirror. However, it must be stressed that these results refer exclusively to nominal headlines and thus it is unclear how they would carry over to the body of the text. Märdh unfortunately fails to distinguish different types of adjectives in spite of the enormous differences between central adjectives, which are gradable and intensifiable, and the classifying prehead adjectives. In chapter 4 above, I have outlined Quirk et al.'s (1985) distinction between the four premodifier slots, precentral, central, postcentral and prehead (for a summary of all the categories distinguished in the analysis cf. appendix I). In the following, I am going to review these positions in detail. The precentral and the postcentral slot turn out to be less important than the other two.

7.1.1. Precentral In precentral position, there are the nongradable, peripheral adjectives like amplifiers, intensifiers and downtoners. On the whole they are not very numer-

Types of premodifiers

139

ous. Over the entire sample not even five per cent of all premodifiers fall into this category. For this reason a further subclassification is not feasible because the number of tokens would be too small for any significant conclusions. Table 7.2 gives the number of adjectives in precentral position in per cent of all premodifiers for every sample, for the aggregate samples of all sections for each paper, for the aggregate samples of all papers per newspaper type for each section, and over the entire sample. The absolute number of premodifiers for each sample is given in appendix III, as table A7.1. Table 7.2. Premodifiers in precentral position in per cent of all premodifiers arts

business

foreign news

home news

sports

Total

3.91 4.08 9.39 5.79

3.24 4.81 4.65 4.26

3.61 4.44 5.87 4.88

DM/ Star Sun Down

4.06 4.06

DE DMa To Mid

3.47 3.97 3.05 3.51

2.45 2.97 2.34 2.61

2.28 4.53 3.95 3.56

3.61 4.43 5.29 4.46

2.93 3.93 3.63 3.49

3.86

DT FT G I T Up

4.55 2.55 3.64 3.09 4.24 3.56

3.19 6.20 2.91 2.52 2.82 3.42

5.88 4.66 3.93 3.09 4.29 4.37

7.76 7.95 6.42 8.26 3.76 6.85

8.59 3.83 2.63 4.57

5.17 5.31 4.81 3.86 3.56 4.52

Total

3.56

3.51

3.81

5.74

4.45

4.31

A 1 -way ANOVA test across all percentage figures of the individual samples (excluding of course the aggregate samples) reveals that the newspaper category as independent factor is not significant but the section is statistically significant at the 5% threshold (p = 0.037). This means that we cannot reject the null hypothesis that the variation within and between groups according to the socio-economic differences of the papers is merely random. For the sections, on the other hand, the null hypothesis does not seem to hold. The subject matter therefore appears to have some effect on this variable. Adjectives in this position are slightly more frequent in the sports section and particularly in the home news section than they are in the other three sections. However, it seems impossible to hypothesise any reasons why this should be so. Similar reason may be involved as for the categories of modified central

140

Premodifications

adjectives. If it can be shown that the same sections use graded, intensified, comparative and superlative central adjectives more often than the other sections, it might be hypothesised that these sections more than others need to specify degrees of application.

7.7.2. Central In central position, the most typically adjectival elements appear. The adjectives that appear in this position are gradable and can be intensified. Together with the prehead position this is the most important position for premodifiers. Table 7.3 shows that in some samples almost half of all the premodifiers are central adjectives (e.g in the arts section of The Daily Telegraph). However, a fourth of all premodifiers is a more typical value. Table 7.3. Premodifiers in central position in per cent of all premodifiers arts

business

foreign news

home news

sports

Total

26.09 18.37 26.76 23.94

30.27 32.62 22.79 28.28

27.95 25.33 22.89 24.92

DMi Star Sun Down

19.93 19.93

DE DMa To Mid

21.88 25.27 17.94 21.77

20.41 18.22 14.02 17.72

23.95 22.22 19.37 21.87

29.90 24.63 28.37 27.60

23.64 22.48 19.74 21.99

23.17

DT FT G I T Up

47.11 37.58 41.30 44.10 36.06 40.97

17.38 20.80 25.93 23.90 18.31 21.47

15.69 14.52 16.07 12.96 20.00 15.70

18.63 22.94 17.43 18.81 11.28 18.01

31.31 28.94 30.26 28.15

23.46 23.67 25.29 26.46 23.24 24.45

Total

40.97

21.40

16.33

20.36

28.03

23.88

There is indeed a great amount of variation that ranges from less than 12 per cent to almost 50 per cent. The null hypothesis again states that the variation is entirely random, whereas the alternative hypothesis is that either the newspaper category or the section influences the variation. In the former case, the null hypothesis cannot be rejected. The newspaper category (as for the precentral position above) does not affect the above variation. The sections, on the other hand, appear to be highly significant (p < 0.001).

Types of premodifiers

141

There are several points that are worth pointing out in table 7.3. The range of percentages is far greater for the up-market papers than it is for the midmarkets or for the down-markets, even within the two sections that are represented for all the papers, i.e. sports and home news. The difference between these two sections is over ten percentage points for the up-markets, about six percentage points for the mid-markets and less than five percentage points for the down-markets. It seems that the up-markets use a greater range of styles or rather more clearly distinct styles than the mid-markets or the down-markets do. For all three paper categories in aggregate as well as for all but one of the individual papers, it is the sports section that has the higher proportion of central adjectives. It is only for The Sun that the two percentages are reversed. The foreign news section has a very low percentage of central adjectives both for the mid-markets and for the up-markets, whereas the arts section in the up-markets has consistently high percentages for all individual papers. In all the arts sections between a third and half of all the premodifiers are in central position. The business sections and the home news section are fairly similar both in the mid-markets and in the up-markets. They have a rather low percentage of central adjectives, albeit not quite as low as the foreign news section. These results have some considerable plausibility in functional terms. It seems intuitively plausible that the arts section depends more heavily on central adjectives, which describe rather than classify. Evaluation on the part of the journalist is not only admissible but actually called for in this section. This is probably also true of the sports section, in which journalists try to convey something of the excitement of the competitive nature of sports activities. In the foreign news sections, evaluations and the expression of excitement are not appropriate to the same extent as in the arts and sports section. Why the business section should differ from the foreign news section in this respect is unclear, but the difference seems to be quite systematic. There is only one paper, The Times, which has got a higher share of central adjectives in the foreign news section than in the business section. In all the other papers the percentage is higher in the business section. It is somewhat easier to postulate a reason for the middle position of the home news section. This section is presumably the one with the greatest range of topics, encompassing reports on politics and parliamentary proceedings, the latest news about the British royalty, reports on crime and accidents, reports on social institutions, and many other topics. The list of articles used as a data base for this study, which is given in appendix II, gives an indication of the great variety of topics covered in this section. This variety might well level

142

Premodifications

out some of the tendencies either to use a high proportion of descriptive, i.e. central, adjectives, or else to use more classifiers, which appear in prehead position. It is further noteworthy that the range of percentages for central adjectives is greatest for the home news section, where it is more than fifteen percentage points, and it is smallest for the business and the foreign news sections, where it is between seven and eight percentage points.

7.7.5. Postcentral The postcentral position is less important numerically than either the central position or the prehead position. The adjectives that appear here are largely descriptive like the central adjectives, though presumably less evaluative. They differ from the central adjectives mainly in the fact that they cannot be graded or intensified. Table 7.4. Premodifiers in postcentral position in per cent of all premodifiers arts

business

foreign news

home news

sports

Total

18.26 19.39 14.08 17.21

13.51 20.86 16.28 16.87

16.14 20.10 13.45 15.90

DMi Star Sun Down

10.70 10.70

DE DMa To Mid

11.46 13.36 11.07 11.97

8.98 14.13 21.50 14.56

15.59 16.87 15.42 15.94

15.46 24.63 21.63 20.66

12.73 16.73 16.97 15.45

DT FT G I T Up

16.12 18.15 12.55 14.61 15.76 15.51

10.99 13.87 9.52 12.89 12.68 11.89

11.76 12.05 9.84 11.11 15.36 11.96

12.73 12.84 13.15 12.84 10.53 12.47

20.08 22.73 18.72 21.05 20.54

14.10 14.14 12.71 13.85 14.80 13.92

Total

15.51

11.79

12.78

14.45

19.55

14.57

The picture that emerges from table 7.4 is largely similar to that of the central premodifier position. The arts section and the sports section have got the highest percentages whereas the other three sections have got fairly low percentages. The section is a statistically significant factor for the variation of these percentage values (p < 0.000).

Types of premodifiers

143

The paper category, on the other hand, is not statistically significant. It does not appear to be worthwhile to split up this category into its subcategories, which include colour adjectives as well as past and present participles, because the absolute figures involved for each cell would be too small, and the differences are not statistically significant.

7.1.4. P re head The prehead position is by far the most important premodifier position in my corpus. This is true numerically - with the exception of the arts section in the up-market papers this position has got a higher share than any of the other positions - and it is also true because of the variation of its subcategories that prove to be statistically significant not just for the factor section but also for the factor newspaper category. Table 7.5 gives the prehead modifiers again in per cent of all premodifiers for each individual section and for the relevant aggregates of samples. Table 7.5. Premodifiers in prehead position in per cent of all premodifiers arts DMi Star Sun Down

business

foreign news

65.31 65.31

DE DMa To

Mid

home news

sports

Total

51.74 58.16 49.77 53.05

52.97 41.71 56.28 50.60

52.29 50.13 57.80 54.31

63.19 57.40 67.94 62.76

68.16 64.68 62.15 65.11

58.17 56.38 61.26 58.63

51.03 46.31 44.71 47.27

60.71 56.85 59.66 59.06

52.90

DT FT G I T Up

32.23 41.72 42.51 38.20 43.94 39.96

68.44 59.12 61.64 60.69 66.20 63.22

66.67 68.77 70.16 72.84 60.36 67.97

60.87 56.27 63.00 60.09 74.44 62.67

37.37 48.51 46.05 46.74

57.26 56.88 57.18 55.82 58.40 57.11

Total

39.96

63.29

67.07

59.45

47.96

57.24

The 1-way ANOVA test shows (as for all the other major premodifier positions) that the newspaper category is not a relevant factor. For this factor the null hypothesis cannot be rejected. Once again, however, the section is a relevant factor (p < 0.001) and in this instance the null hypothesis has to be rejected.

144

Premodifications

This position is clearly the complementary category for both the central position and the postcentral position. Whereas the adjectives in those positions are largely descriptive, the premodifiers that appear in prehead position are classifying. The sections that have high shares of premodifiers in central and postcentral position, i.e. the arts section and sports section, have correspondingly low shares of premodifiers in prehead position. In the foreign news section of every individual paper it is almost two thirds of all premodifiers that appear in this position. In the business section, the share of this position is only slightly lower. Thus there seems to be a great need for classifying nouns rather than for adding descriptive detail in these sections. The differences are made in all newspaper categories. But again the variation within the up-market papers is greater than among the mid-markets or even the down-markets. For the up-markets the difference between the home news section and the sports section is almost 16 percentage points, for the mid-markets it is just over ten percentage points and for the down-markets it is about 6.5 percentage points. Again the up-market papers appear to use fairly distinct styles in their different sections whereas the down-markets use a fairly uniform style, with the mid-markets appropriately holding the middle ground between the two. What may nevertheless be surprising is the fact that all three categories of newspaper show the differences in the same direction. All of them have higher shares of descriptive adjectives (i.e. premodifiers in central and postcentral position) in the sports section than they have in the home news section and vice versa for classifying premodifiers in prehead position. Classifying premodifiers can be adjectives, nouns, names or rank-shifted elements with an internal clause syntax but functioning as a prehead noun. The constituents of such rank-shifted premodifiers are usually hyphenated. They are however comparatively rare. In no sample are there more than one or two examples among the 1000 noun phrases analysed. They are therefore excluded from further consideration in the following subsections. These will be devoted to the three main prehead items, adjectives, nouns and names. The following tables give an idea of the relative importance of the prehead, i.e. classifying, premodifiers. The percentages given here therefore do not relate to the total number of premodifiers as in the previous tables but only to the total number of premodifiers in prehead position. The absolute figures in these instances do not provide an adequate basis for comparison because of the considerable differences in the overall number of modifiers in the different samples. The corresponding tables with the absolute figures are given in appendix III. In this case there are clear differences not just from one section to the other but also across the three different newspaper categories. In the down-market

Types of premodifiers

145

Table 7.6. Prehead adjectives in per cent of all preheads arts

business

foreign news

home news

sports

Total

18.49 19.30 25.47 20.94

22.45 12.82 16.53 17.51

20.28 16.67 18.32 18.45

DM/ Star Sun Down

15.25 15.25

DE DMa To Mid

12.09 25.79 20.22 19.08

34.73 33.33 42.11 36.29

35.95 24.09 19.35 26.52

25.25 24.47 19.35 23.08

26.62 27.48 25.04 26.39

24.82

DT FT G I T Up

55.13 35.11 48.57 31.62 57.93 44.87

17.10 16.67 30.04 24.87 25.53 23.43

38.24 50.60 47.66 53.39 44.38 47.30

35.20 29.89 28.16 25.19 27.78 29.51

22.97 29.82 40.95 29.77

31.81 35.03 35.82 35.06 37.21 35.01

Total

44.87

21.26

43.93

27.02

24.28

30.70

papers, less than a fifth of all preheads are adjectives in the mid-markets it is about a fourth and in the up-markets it is as much as a third. There is no overlap between the categories if the aggregate samples of all the papers are considered. All the down-markets have percentages that are lower than any of the mid or up-markets, and all the up-markets have percentages that are considerably higher than any of the others. The 1-way ANOVA test indicates that in this case we have to reject the null hypothesis (p < 0.002). The section also proves to be a relevant factor (p < 0.001). The foreign news and the arts section both have got shares of prehead adjectives among their preheads that are very considerably above the averages for the other sections. The business section, on the other hand, has a fairly low share of prehead adjectives. This is true even for the one single business section in the down-market papers. The percentage of prehead adjectives in the business section in The Sun is lower than in any other down-market sample. In both the down-markets and in the mid-markets, the percentage of prehead adjectives is higher in the home news section than it is in the sports section. For the up-markets, the two values are almost equal in spite of considerable differences across the individual papers. On the face of it, this seems to be counter-evidence to the claim that the upmarket papers use a less uniform style than the mid-markets or the downmarkets, but the close similarity in the two values for the aggregates of all

146

Premodifications

home news and sports sections in the up-markets respectively conceals the variation of the individual papers. Table 7.7 gives the share of all prehead modifiers that are nouns for all individual samples. Table 7.7. Prehead nouns in per cent of all preheads arts

business

foreign news

home news

sports

Total

65.55 63.16 59.43 62.83

44.90 51.28 57.85 51.85

56.22 58.33 63.37 60.27

DMi Star Sun Down

69.49 69.49

DE DMa To Mid

67.03 44.65 63.48 58.96

46.11 55.17 41.35 48.10

44.44 54.74 61.94 53.71

51.52 46.81 45.16 47.90

52.91 50.71 54.74 52.78

51.09

DT FT G I T Up

38.46 56.49 42.86 51.47 36.55 45.71

67.36 64.81 54.08 58.55 60.85 60.73

45.10 35.86 32.24 29.24 41.42 36.31

51.53 58.15 54.37 54.96 54.04 54.54

54.05 47.37 45.71 49.30

52.35 51.65 47.12 46.67 49.41 49.38

Total

45.71

61.10

39.92

55.97

49.65

51.62

The 1-way ANOVA shows the section to be a relevant factor (p < 0.001), but the newspaper category in this case fails to be statistically significant (p = 0.081). The null hypothesis cannot be rejected. Nevertheless, in spite of the fairly small differences, the values for the aggregate samples for the individual newspapers do not overlap. The percentage of prehead nouns is highest in the down-markets, somewhat lower in the mid-markets and lowest in the up-markets. There is, however, considerable variation across the sections of individual papers, and some samples, e.g. the business sections of the upmarket papers, have got percentages of prehead nouns that are as high as or even higher than many of the individual samples in the down-markets. Among the sections it is the business sections for all three newspaper categories in aggregate that have the highest percentage of prehead nouns. Only the Daily Mail and The Guardian have other sections with higher or equally high percentages for prehead nouns. In most cases the home news section has got a percentage of prehead nouns that is slightly higher than the sports section. The only exception in this case is the Daily Express.

Types of premodifiers

147

Table 7.8, finally, shows the percentage of all premodifiers in prehead position that are names. Table 7.8. Prehead names in per cent of all preheads arts

business

foreign news

home news

sports

Total

14.29 16.67 12.26 14.45

32.65 35.90 25.62 30.64

22.58 24.48 17.57 20.54

DMi Star Sun Down

15.25 15.25

DE DMa To Mid

20.33 29.56 16.29 21.77

19.16 11.49 16.54 15.61

18.95 21.17 18.71 19.55

23.23 28.72 35.48 29.02

20.13 21.81 20.21 20.71

DT FT G I T Up

6.41 8.40 7.62 16.91 5.52 9.24

15.54 18.52 15.88 16.58 13.62 15.85

16.67 13.15 20.09 17.37 14.20 16.29

13.27 11.96 17.48 19.85 18.18 15.96

24.09 22.97 21.93 13.33 20.70

15.84 13.19 16.95 18.15 13.38 15.53

Total

9.24

17.58

16.09

16.60

25.96

17.51

Between roughly a fifth and a sixth of all preheads turn out to be names. For this variable, both the factor section and the factor newspaper category are statistically significant at the 1 per cent level, (p = 0.009 for the factor paper category p < 0.001 for the factor section). The share of prehead names is smallest for the up-market papers and highest for the down-markets. The Sun, however, has got a share of prehead names that is considerably lower than the other two down-markets, and it therefore reduces the aggregate value to such an extent that it is actually even marginally lower than the aggregate value for the mid-market papers. The value for The Sun is in fact lower than even one of the up-market papers, i.e. The Independent. But otherwise there is again no overlap between the three different newspaper categories. The percentage of prehead names in all cases but one is higher in the sports section than it is in the home news section. The one exception is The Times. This variable is very unusual in that the difference between the home news section and the sports section is considerably more pronounced in the downmarket papers than it is in the mid-markets or the up-markets. The difference for the down-markets is over 15 percentage points, for the mid-markets it is

148

Premodifications

roughly ten percentage points, and for the up-markets it is a mere five percentage points. The arts section in the up-market papers has the smallest percentage of prehead names. Only in one paper, The Independent, does it have over ten per cent. The difference between the business section, the foreign news section and the home news section is not very great. Moreover it does not pattern in the same way in the up-markets and in the down-markets. Indeed in the midmarkets and in the up-markets there are no two papers which rank all their sections in the same order. The only generalisation that can be made is that most up-market papers have the lowest percentage in the arts section, and that most papers across all newspaper categories have got the highest percentage in the sports section.

7.2. Stylistic significance of premodifier variation In this section I shall attempt an interpretation of the statistical facts that have been reviewed in the previous section. Such an attempt must necessarily be speculative to a considerable degree. The premodifier categories set up in the previous section were based almost exclusively on syntactic criteria. As I have repeatedly implied, these categories generally coincide with semantic differences, but the relationship is not in all cases very clear and uncontroversial. Furthermore, even if this relationship can be established, it is a further and very complex step to reach any conclusions about the stylistic aims of specific newspapers and newspaper sections on the basis of numerical differences in the use of such categories even if these differences can be shown to be statistically highly significant. It must be stressed again that statistics, as used above, can only show whether it is likely that differences in distribution are random or not. Technically, randomness can never be entirely excluded. And even if the probability of randomness is extremely small, the statistics never provide any reasons for the variation. Before I go on to analyse the semantic correlates of the premodifier categories set up above, I will briefly summarise the salient points that emerged in the previous section. The relative share of the four major premodifier categories in individual samples varies significantly according to the newspaper section, but there is no significant variation according to the newspaper category. High proportions of premodifiers in central and in postcentral position tend to coincide with low proportions of premodifiers in prehead position and vice versa. The proportions of premodifiers in precentral position are in all cases small.

Stylistic significance of premodifier variation

149

The sports section in all newspaper categories and the arts section which only exists in the up-market papers have high proportions of premodifiers in central and postcentral position and low shares of premodifiers in prehead position, whereas in the foreign news section, the home news section and the business section, the proportion of premodifiers in prehead position is high. The prehead position is particularly important, both numerically - all newspapers have more than half of all their premodifiers in this position - and because of its internal variation. For the subcategories of premodifiers in prehead position - adjectives, nouns and names - both the newspaper category and the section are important for the internal variation. The up-market papers have the highest proportion of adjectives and the lowest share of nouns and names in prehead position, whereas the downmarket papers have a high proportion of nouns and names in prehead position but fewer adjectives. The mid-market papers hold the middle ground for all these values. The arts section and the foreign news section have got a high share of adjectives and few nouns, whereas the business section and the home news section have many nouns but a comparatively small proportion of adjectives. The sports section, finally, has got a higher proportion of names in prehead position than any of the other sections. In chapter 4 above, I briefly reviewed some of the work that has been done on the ordering of premodifiers. I will now have to address this question again from the perspective of the interpretations that suggest themselves for the above variations on the basis of semantic differences of the different subcategories of premodifiers. Researchers generally seem to agree that the premodifiers nearest to the head are those that are most objective they include the "least adjectival and most nominal premodifiers" (Quirk et al. 1985: 1339) they are non-emotive and classifying. Relevant examples are given in (1) to (9) (1) (2) (3) (4) (5) (6) (7) (8) (9)

Scottish prisons Iraqi planes ministerial ranks local authority prison authorities daylight strike Perth jail British Telecom flotation Iranian oil tanker (all examples I 6.10.87, 1)

150

Premodifications

In all these examples the premodifiers are in prehead position. They specify a subset of the possible referents of the head noun. This is done by means of an adjective, a noun or a name. In (1) and (2) the premodifiers are adjectives that are based on nationality, and in (3) and (4) they are denominal adjectives. These adjectives typically cannot take an intensifier like very or extremely and they do not allow comparisons (*very Scottish prisons, *extremely Iraqi planes, *more ministerial ranks, *the most local authority). In (5) and (6) the premodifier is a noun which in both cases serves the same purpose as the adjectives in (1) to (4). It specifies the reference of the head noun. (5) refers to a particular type of authorities, and (6) refers to a particular type of strike. (7) and (8) have got a name as premodifier. In the case of (7) it is one single proper noun, whereas in (8) it is a complex name consisting of a proper noun preceded by an adjective. This adjective clearly is part of the name British Telecom. It does not modify the head noun flotation. This analysis depends on background knowledge that is assumed on the part of the reader. The reader is expected to know that the firm referred to is called British Telecom and given the high profile of this particular firm it is a very reasonable assumption. Otherwise (8) is actually structurally ambiguous. It could also refer to the British flotation of Telecom. Example (9) finally has got two modifiers in prehead position, one an adjective and one a noun. Premodifiers that are further away from the head are the more typically adjectival elements which are more subjective they are descriptive and characterising rather than classifying. These adjectives can be intensified, and they allow comparison. They are in central position. (10) (11) (12) (13) (14) (15) (16) (17) (18)

a succulent encore the one obvious fault their grim and gloomy mien no intense personal vision (DT 18.12.87, 10.1) a splendidly gloomy Vladimir (DT 27.11.87, 20.4) an intelligent, coherent, if stubbornly unmoving adaptation of the play (DT 27.11.87, 20.1) a much more able and imaginative musician than supposed a most subtly shaded account of the Prokofiev symphony the most hardened cynic (unless otherwise marked all I 6.10.87, 16)

The adjectives in the examples (10) to (18) do not specify a subset of the referents of the head noun. They rather provide some additional description of the head noun. The adjectives are gradable, and it is usually a matter of personal

Stylistic significance of premodifier variation

\ 51

opinion whether they apply to the head noun at all or not. The mien of the musicians of the Leningrad Philharmonic referred to in (12) or the encore of example (10) might have been perceived quite differently by a different art critic. In the examples (1) to (9), on the other hand, it is not a matter of opinion whether, and if so to what degree, the premodifiers apply. Thus, the premodifiers in (10) to (18) are not just descriptive, they are also to a large extent evaluative and subjective. In (12) and (13) two central adjectives are conjoined. (14) and (15) illustrate intensified central adjectives, and (16) to (18) show them in comparisons. The distinction between the premodifiers in (1) to (9), on the one hand, and in (10) to (18), on the other, relates directly to what Quirk et al. (1985: 1341) call the "subjective/objective polarity", which applies to all premodifiers and which accounts for the ordering of premodifiers in sequences. On this basis an interpretation can be postulated for some of the facts reported in the preceding section. The polarity between objective assessibility and subjective description and evaluation is not an important stylistic feature in order to distinguish the language as it is used by the three different types of newspapers. It is, however, an important stylistic feature for the individual newspaper sections. That the arts section relies more heavily on subjective description and evaluation than any other section is perhaps not surprising. But it is less predictable that it should be the foreign news section which relies the least on descriptive and evaluative premodifiers. On the other hand, the foreign news section, both in the mid-markets and in the up-markets, has got a very high share of modifiers in prehead position, which relate to objectively assessible properties. The arts section is correspondingly low on this count. It may also be surprising that the sports section in this respect is stylistically close to the arts section and rather distinct from the remaining three sections. For the sports section, too, much importance seems to lie in evaluation and subjective description. The second important variable in noun phrase premodification is the distribution of the subtypes of preheads. In chapter 4 above, I reported the research by Levi (1978) on the different stylistic values of noun premodifiers and denominal adjectives. Her data show that the use of denominal adjectives increases with the educational level of the speaker and addressee. As examples she mentions the fact that medical personnel is more likely than lay persons to use the term renal disease instead of kidney disease, and engineers or geologists are more likely to talk about fluvial currents instead of river currents. This difference is indeed important, and it applies in the way suggested by Levi. In the up-markets more than a third of all modifiers in prehead position are adjectives, while in the mid-markets the proportion is about one fourth,

152

Premodifications

and in the down-markets less than a fifth. Correspondingly a much higher proportion of prehead modifiers are nouns in the down-markets than in the upmarkets. Hence the hypothesis that the higher educational profile of the readership of the up-market papers tends to go together with a higher incidence of prehead adjectives can be supported by the available data. This has interesting corroborations because the relative proportions of prehead adjectives and prehead nouns also vary significantly across the different sections. In this respect it is more difficult to predict which section would have the highest incidence of prehead adjectives. In fact, as pointed out in the previous section, in the up-market papers it is the foreign news section, closely followed by the arts section, while the business section has the lowest proportion of prehead adjectives and correspondingly the highest proportion of prehead nouns.

7.3. The individual styles In chapter 6 above, I presented a cluster analysis of all 43 samples using the number of individual types of premodifiers and postmodifiers as variables. The results, presented in the form of a dendrogram in figure 6.1, suggested that the arts sections, for instance, form a cluster of their own, that is to say the five samples of language from the arts sections of the five up-market papers are - on the basis of these variables - more similar to each other than they are to any other sample. The sports sections of all types of papers fall into one cluster together with most sections of the down-markets and mid-markets. Separate clusters are formed by most business, foreign news and home news sections of the up-markets on the one hand, and by the business sections of the mid-markets and The Sun, which is the only down-market with a regular business section, on the other. Three up-market foreign news sections form one of the most coherent independent small clusters. As I stressed above, these clusters can only be suggestive of what kind of variation may be relevant. In section 7.1 I analysed the data in greater detail and with statistical tests that allowed the evaluation of explicit hypotheses. In the following I shall exemplify the individual styles as they suggested themselves in the cluster analysis, without following the clusters obtained there in every detail. The first style comprises all the sports sections from all three types of papers. The second style comprises the foreign news section of the mid-markets and the home news section of the down-markets and mid-markets. In the cluster analysis presented above these styles formed one single cluster. Here I treat

The individual styles

153

them separately for ease of presentation. Style three comprises the business sections of the mid-market papers and of the one down-market paper which has got a regular business section, i.e. The Sun. Style four is formed by the arts sections, which are confined to the upmarket papers. The fifth style comprises the business, the home news and the foreign news sections of the up-market papers with the exception of the foreign news sections of the Financial Times, The Guardian, and The Independent, because they form a cluster of their own, according to the cluster analysis presented in chapter 6. Hence these three sections make up the last style. These styles are ordered according to the average number of premodifiers per section. Table 7.9 gives a survey of the most important variables in averages per section and - in order to show the distribution of different types of premodifiers - in per cent of the total number of premodifiers per section. Table 7.9. Average number of types of premodifiers in six styles style

N

all premodifiers

precentral

central

postcentral

all preheads

prehead adj

prehead nouns

prehead names

1 (III)

10

204.2

2 (III)

9

236.2

3 (IVb) 4

274.5

4(11)

297.8

2.4 1.18 9.2 3.90 10.0 3.64 10.6 3.56 15.3 4.96 13.0 3.92

59.2 28.99 49.8 21.07 58.5 21.31 122.0 40.97 59.8 19.32 48.0 14.49

41.3 20.23 37.4 15.85 32.0 11.66 46.2 15.51 38.3 12.39 36.7 11.07

101.3 49.61 139.8 59.17 174.0 63.39 119.0 39.96 195.9 63.34 233.7 70.52

24.6 12.05 40.1 16.98 31.5 11.48 53.4 17.93 59.2 19.13 118.3 35.71

50.3 24.63 75.6 31.98 107.3 39.07 54.4 18.27 106.3 34.38 76.00 22.94

26.3 12.88 23.3 9.88 35.0 12.75 11.0 3.69 30.3 9.81 39.0 11.77

5

5 (IVa) 12

309.3

6(1)

331.3

3

Average number in the first line of each row and in per cent of all premodifiers in the second line. Roman figures in brackets indicate the cluster as given in the dendrogram of figure 6.1. Style 1: sports style 2: mid and down, home news and foreign news style 3: mid and down, business style 4: arts style 5: up, business, home news and foreign news (excluding foreign news of FT, G and /) style 6: foreign news of the Financial Times, The Guardian and The Independent.

These six styles have to a considerable extent been suggested by the cluster analysis presented in the previous chapter. They deviate from the statistically computed analysis in as far as cluster III of that analysis is now split up into two styles, style 1 and style 2. The second deviation is the home news section

154

Premodifications

of The Independent, which falls into cluster III, but here it is taken together with all the other home news sections of the up-markets in style 5. Hence, the six styles are set up on the basis both of the suggestions derived from the cluster analysis and of internal plausibility. This grouping, moreover can also be shown to be statistically significant. A 1-way ANOVA test for each relevant variable shows whether the suggested grouping into six styles is statistically relevant as an independent factor, that is to say it establishes the probability for each variable that the difference in the variation within groups and between groups may be the result of random sampling. In fact, this can be excluded for most variables. The variation within the groups, i.e. the individual styles, is significantly smaller than between the groups. Even if the requirement for statistic significance is set at the 0.1% level, the following variables prove to be significant: number of premodifiers, central adjectives, total number of prehead modifiers, prehead adjectives, prehead nouns, prehead names, unmodified central adjectives, and intensified central adjectives. The precentral adjectives are only significant at the 5% level whereas the postcentral modifiers are not significant even at the 5% level. However, it must be recognised that other groupings may also produce significant variation across some or even most variables that have been taken into account. As I have shown above the newspaper category as an independent factor produces significant variation for several variables, and the newspaper section turns out to be a significant factor for almost as many variables as the six styles suggested above.

7.3.1. The sports section The sports section in all categories of papers is the one with the smallest number of modifiers and also the smallest number of premodifiers. Among the premodifiers, it has a very high proportion of descriptive, i.e. central and postcentral, premodifiers and a correspondingly low proportion of prehead modifiers. As in the majority of styles, there is a higher proportion of prehead nouns than prehead adjectives. In this style there are about twice as many prehead nouns. The following extracts are relevant examples: (19) (20) (21) (22) (23)

Furious Chris Nicholl, Star, 18.12.87, 25.5 a grim-faced Frankham, Star, 17.12.87, 32.5 Pakistan's loudmouth skipper, Star, 17.12.87, 32.1 Embury, whose invaluable innings of 70 rescued England from a nightmare 85 for 6 to a respectable 222 for 7 by the close, Star, 17.12.87, 32.4 Allen collected a phenomenal 49 goals last season, DMi, 28.10.87, 35,4

The individual styles

(24) (25) (26) (27) (28) (29) (30) (31)

155

It was magical Davie Cooper, the man singled out by Souness as a possible match winner, DE, 05.11.87, 46.5 Flamboyant trainer Rod Simpson, renowned for the outstanding placing of his runners, DE, 05.11.87, 44.1 the central figures in the drama at a disbelieving Gaddafi Stadium were the captains, DE, 05.11.87, 42.2 For sheer unadulterated watchability, the match yesterday .., G, 31.10.87, 17.8 he had made a scintillating 80, G, 31.10.87, 17.6 the young left-handed Gurusinghe, G, 31.10.87, 17.6 a one-under-par 70, DT, 18.12.87, 24.6 the dashing 'up and at 'em' tone, I, 16.10.87, 32.3

These examples reveal quite a lot which is not immediately apparent in the mere figures of the statistics given above. As in the sports sections as a whole, there are many central adjectives in these examples, but what is striking is the high proportion of central adjectives that modify personal names. The adjectives furious and grim-faced in (19) and (20) modify the heads of their noun phrases non-restrictively by adding some evaluative description. Example (21) is very similar in its effect. The modifier loudmouth adds some descriptive and evaluative information to the noun skipper. But the classification of loudmouth is not simple. It is in itself a compound consisting of an adjective and a noun. As a compound it is a noun itself because it could take the appropriate position as head of a noun phrase. As a premodifier it is unusual. It is clearly non-restrictive. The correct reference assignment of the head noun skipper does not depend on this modifier. It does not restrict the set of possible referents to the one intended referent. The genitive noun in determinative position, Pakistan's, unambiguously identifies the referent. On the other hand, loudmouth cannot act as a central adjective because it fails all the relevant tests. It cannot be used in a comparison (*the more loudmouth skipper than ...), it cannot easily be modified (?*the very loudmouth skipper), and it cannot be used predicatively (*the skipper was loudmouth). It might in fact be argued that the construction loudmouth skipper is a dvandva construction, in which the two elements stand in a coordinate relationship both of which could act alone as the head of the noun phrase (Pakistan 's loudmouth and Pakistan 's skipper) or the order of the two could be reversed (?Pakistan's skipper loudmouth). None of these interpretations is entirely satisfying, but the example serves as an illustration of the difficulties that occur in the classification of language as it is actually used in contrast to the examples invented by linguists in order to illustrate some classificatory scheme.

156

Premodifications

Examples (22) and (23) illustrate similar points as examples (19) to (21) though with noun phrases not referring to people but to cricket scores and to goals respectively. The three premodifiers in (22) are invaluable, nightmare and respectable, the first and last of which - as phenomenal in (23) - are clearly central adjectives, which could potentially be used in a comparison, they could be intensified or used predicatively. The modifier nightmare, on the other hand, fails all these tests and therefore cannot be a central adjective, but it is clearly non-restrictive. The score 85 for 6 does not need any modification to be unambiguous. So far all examples have been taken from the down-market papers. Examples (24) to (26) are taken from mid-market papers and (27) to (31) from upmarkets. The statistical analysis above has shown that - as far as the different categories of modifiers are concerned - the differences across the three types of papers within the sports section are not significant. And indeed, the examples for the mid-market papers are very similar to the ones discussed above. (24) and (25) are exactly parallel to examples (19) to (21). The adjectives magical and flamboyant are central adjectives which non-restrictively modify a name or a noun referring to a person. Both adjectives are evaluative and reflect the personal opinion of the journalist. The adjective disbelieving in (26), as a present participle, is in postcentral position, but again it is descriptive and non-restrictive. The entire phrase is also remarkable because it is a metonymy, that is to say an expression which literally refers to a place is used to refer figuratively to the people gathered there. The examples from the up-market papers are slightly different. (27), for instance, illustrates an adjective in precentral position, i.e. sheer. It is a peripheral adjective because it fails the relevant tests of comparability, intensifiability and predicateability. The adjective unadulterated in the same noun phrase is again a participle but in contrast to disbelieving in (26), it is a past participle. The modifiers young and left-handed in (29) are clearly non-restrictive they modify a name, but they are not evaluative as the modifiers furious, grimfaced, flamboyant, or magical above. The last two examples in (30) and (31) are not typical, but they are nevertheless interesting. It is not quite clear how the two modifiers one-under-par and "up and at 'em' have to be classified. The former is non-restrictive, descriptive and therefore not a classifier in prehead position, but it also is not an adjective. The hyphens between the three words suggest that it is a rankshifted modifier, but it lacks the internal clause structure. The modifier in (31) has a better claim to be a rank-shifted item. In this case there are no hyphens, but the construction is bracketed off by inverted commas.

The individual styles

157

7.3.2. The home-news and foreign news sections in the down- and mid-markets The home news sections of the down-market and mid-market papers and the foreign news sections of the mid-market papers again form a fairly coherent style. There are of course no foreign news sections in the down-market papers. Articles that would go into this category are rare in these papers, and if they appear, they are not assigned to a separate section. These sections also have relatively small numbers of premodifiers, even though they are slightly higher than the sports sections. Just over one third of all premodifiers in these sections are descriptive. This is an average value for all the six styles as set up for this chapter. In the foreign news section of the Financial Times, The Guardian and The Independent, it is as small as a fourth, and in the arts sections of the up-market papers more than half. Correspondingly there is also a fairly average value for the proportion of prehead modifiers, i.e. about 60 per cent. The ratio of nouns and adjectives in prehead position is about the same as for the sports section there are almost twice as many nouns. The following examples again combine fairly typical and rather untypical but therefore all the more remarkable instances of premodifiers. (32) (33) (34) (35) (36) (37) (38) (39)

Boozy bully Steven Mullins was banned from EVERY pub in Britain yesterday, Star, 18.12.87, 1.1 Alluring Alexandra Peacock believes in free enterprise. Star, 19.12.87, 3.1 And the curvy cutie really is made to measure for success, Star, 19.12.87,3.1 Attractive Tracey Knowles has been hunted by police, Star, 17.12.87, 9.4 Tearful Tracey, a North London Council worker, Star, 17.12.87, 9.4 Gorgeous Ruth Gordon studied the human form ..., Star, 18.12.87, 3.1 Glamour girls are fleeing from the smutty Star - and flocking to the marvellous Mirror, DMi, 08.10.87, 3.2 controversial consultant Dr Marietta Higgs, DE, 20.11.87, 17.2

All the examples (32) to (39) contain premodified names or premodified nouns referring to persons. These adjectives are non-restrictive, descriptive and in most cases evaluative and subjective rather than objective, and they all alliterate with the name or noun they modify. In one instances, i.e. in (32), the alliteration is kept up even outside the noun phrase. The initial b of boozy and bully is taken up in banned, Britain and in the voiceless version in pub. Such examples of alliterations are of course not restricted to the down-market pa-

158

Premodifications

pers as example (39) indicates but they are particularly prevalent in the downmarkets. (40) (41) (42)

petite Priscilla Welch, DMi, 02.11.87, 3.1 Curvy Karen Brennan, DMi, 17.12.87, 3.1 the flamboyant Hatten, DE, 23.11.87, 18.1

Examples (40) to (42) show some adjectives that are often - and stereotypically - used in connection with people in the down-market and to a lesser extent in the mid-market papers. The two down-market examples, (40) and (41), again have an alliteration of the adjective and of the name. Waterhouse comments on such stereotypical adjectives in the down-market papers. His main concern is the language of the Daily Mirror but his remarks often claim a wider applicability than just this paper. For a profession that is supposed to be hard-boiled, journalism is remarkably chivalrous with its adjectives. Models are attractive or stunning, or they are given the accolade of top ..., or they are curvy - which seems to have come into vogue when someone realised that curvaceous was getting long in the tooth. (So is curvy, by now.) Small young women are petite. Small old women are tiny. Young men are often dashing - an anyone-for-tennis word from the same era as vivacious. There are few ordinary housewives, but model housewives abound (1981: 14-15). It is of course one of the major differences between the three socio-economic types of newspapers that the down-markets seem to rely on a small number of short, simple and stereotypical adjectives, whereas the mid-markets and to a far greater extent the up-markets use a greater range of adjectives many of which are neither simple nor short as the illustrations in the subsequent sections will show. The premodifiers in the down-market and the mid-market papers are very often hyperbolic, as the following examples demonstrate. (43) (44) (45) (46)

Police also broke the tragic news to Caroline's boyfriend, Star, 17.12.87,2.2 He faced a devastating onslaught from the Conservative Education Association, DMi, 08.10.87, 4.3 Their anguished parents have discovered the shocking truth, DMi, 27.11.87, 1.2 A euphoric Lord King, DMa, 22.12.87, 2.1

For the tabloid papers there seems to be a need to describe everything in particularly dramatic terms. Things are seen to be at the extremes of human expe-

The individual styles

159

rience. However, this is not expressed, as might have been expected, in terms of modified or graded adjectives. The statistics revealed that these sections do not tend to use more adjectives that are intensified by very or some other adverb or more adjectives that are in the comparative or superlative. Both the arts sections in the up-market papers and the combined sample of sports sections have got a higher percentage of intensified and graded central adjectives. The examples (43) to (46) illustrate that the difference is again not structural but lexical. The onslaught in (44), for example, is not very damaging, it is devastating. Thus it is not a neutral gradable adjective that is used together with an indication of the degree to which it applies, but an adjective is used which has "extreme" as part of its lexical meaning. (47) (48)

It was in a state of utter despair that he committed these offences, To, 18.12.87, 10.1 hero fireman Colin Townsley, who died in the King's Cross tragedy after sending his men off to fetch breathing apparatus, DE, 20.11.87, 1.4

Example (47) makes the same point. Here a peripheral adjective - in precentral position - is used as an amplifier (Quirk et al. 1985: 429-430) in order to express an extreme degree to which the head noun applies. (48), on the other hand, is interesting because the modifier hero is in prehead position it classifies the head nounßreman. As in example (21) in section 7.3.1 above it could also be analysed as a dvandva compound because both the head noun and the modifier refer independently to the intended referent. However, in order to be a true dvandva it should be possible to invert the order of the two nouns, i.e. fireman hero, which seems possible but in the given context far less likely than the order in which it actually appears. The following two examples are illustrations of fairly complex premodification structures. They are not typical for the down-market or mid-market papers in that such long and multiply premodified noun phrases are quite rare, but they are also typical in that they show possibilities that do exist and are used regularly if not in such concentrated form. (49) (50)

he runs a 130 mph silver X-registered BMW coupe, valued at more than £ 12,000, DE, 23.11.87, 18.1 Canadian-born Roman Catholic divorcee Sylvana Tomaselli, DMa, 18.12.87,9.4

In (49) there are four separate premodifiers and a postmodifying participle phrase. All the premodifiers are restrictive because they successively, from

160

Premodifications

right to left, further narrow down the range of possible referents. The premodifier nearest to the head, BMW is a name and stands in prehead position. It is preceded by the past participle X-registered, which indicates the age of the car, in postcentral position. Also in postcentral position is the colour adjective silver and the indication of the maximum speed 130 mph. The postmodifying participle phrase, on the other hand, is less restrictive. It adds some more information about the car. This may of course further restrict the range of referents. There may be cars answering to the description given by the premodifiers that are more expensive or less expensive. But it is ultimately not important whether any or all of the modifiers in this noun phrase are restrictive or non-restrictive. They certainly do not restrict the reference of the head noun to one single individual car even if they drastically reduce the possible set of referents. These details are added because they are felt to be interesting in themselves. They serve the purpose of describing in some detail one particular car, and thereby characterise its owner. Example (50) is very similar. There are two premodifiers which are neither clearly restrictive nor clearly non-restrictive. The head noun itself stands in apposition to a name of a person. This construction will be dealt with in detail in chapter 9. Here it suffices to say that this is an example of what Waterhouse (1981: 91) terms "the practice of prefacing names with potted biographies." The following two examples, (51) and (52), are taken from the foreign news section of two of the mid-market papers. In both cases there is a noun phrase name apposition in which the noun is preceded by two premodifiers. Both examples are the first part of the opening sentence of an article. (51) (52)

Sacked Moscow Communist Party boss Boris Yeltsin, DE, 20.11.87, 13.4 Crooked Wall Street financier Ivan Boesky, who made a fortune from "insider dealing", DMa, 19.12.87, 10.1

The first premodifier in both cases is a past participle in postcentral position. This is followed by a complex name consisting of two or three proper nouns in prehead position. The prehead elements are restrictive. It makes clear what type of boss or financier we are talking about, whereas the past participles in postcentral position merely provide some additional information. In these specific cases, however, the past participles may be crucial for the reader because they encapsulate the reason why these two men had been in the news for some time previous to the issues from which these excerpts were taken. Boris Yeltsin had lost his job as leader of the Moscow Communist Party, and Ivan Boesky had been accused of insider trading in a recent takeover scandal. Both

The individual styles

161

men would not normally be known to the great majority of readers of these papers. They are not public figures in a British context. But they had been in the news for several days, and the "potted biography" prefacing their names immediately recalls their newsworthiness. The last example is also from the foreign news section of the mid-market paper, the Daily Mail. (53)

in Kuro, a bleak concrete industrial suburb of Seoul, DMa, 18.12.87, 10.4

In (53) a central adjective bleak precedes two prehead modifiers, the first of which is a noun, concrete, and the second an adjective, industrial. Only the central adjective would be gradable or intensifiable. Again it is clear that it is this adjective which introduces an evaluative, subjective element into the description of this suburb. The two prehead modifiers, on the other hand, are more objective descriptions which could be checked. But it is also clear that this distinction is not absolute. There is a reading for this noun phrase in which both concrete and industrial are gradable descriptions, which may apply to some extent only. After all there might be another suburb with an even higher concentration of concrete and industry.

7.3.3. The business section in the down- and mid-markets This style comprises only four sections the business sections of the three midmarket papers and the business section of The Sun. Its average number of premodifiers, however, is considerably higher than that of the preceding two styles, and the proportion of descriptive premodifiers, that is to say premodifiers in central and postcentral position, is further reduced. Only just under 33 per cent of all premodifiers fall into this category as compared to almost 37 and 49 per cent in the home news and foreign news sections of the down- and mid-markets and in all the sports sections respectively. The proportion of preheads is correspondingly higher. It is about 63 per cent. In this style the difference in the proportions of nouns and adjectives in prehead position is at its most extreme. As in the previous two styles, there are more prehead nouns than prehead adjectives, but in this style there are almost three and a half times as many nouns as there are adjectives. There is a comparatively high percentage of prehead names. About 13 per cent of all premodifiers fall into this category. It is only in the sports sections that this ratio is equally high, or indeed marginally higher. Excerpts (54) and (55) show that the urge to stress the extreme size and importance of reported events exists in the business section, too.

162

Premodifications

(54) (55)

the mammoth fall in share prices, DE, 05.11.87, 41.3 And by last night, it stood at $ 1.7810 against sterling, a mammoth drop of almost 4 cents in just 24 hours. DE, 06.11.87, 41.5

The premodifier mammoth is here used as a noun in prehead position. It might also be argued that it is an adjective in prehead position, even though it clearly is not a central adjective because it fails the relevant tests. It cannot be intensified (*a very mammoth drop), and it cannot be graded (*a more mammoth fall than last week). The fact that it cannot even be used predicatively (*the fall was mammoth) indicates that it is a prehead noun rather than a prehead adjective because only the latter but not the former can usually be used predicatively (cf. the plane was American versus *the driver was train). The following two examples illustrate the preference in this style of using prehead nouns. (56) (57)

Heath, a leading specialist share dealer, Sun 08.10.87, 23.2 Money men were strangely absent from the windowsills of their highrise office blocks after the London stock market followed the lead of the Wall Street crash. Sun, 20.10.87, 21.2

In (56) there are three premodifiers in sequence. The first is a present participle in postcentral position, the other two are nouns in prehead position. The head noun dealer stands in apposition to the name Heath. There are again two possible interpretations for the two prehead nouns. Whereas share is clearly a restrictive modifier of dealer, it could be argued that specialist and dealer together form a dvandva. Both nouns can refer directly to the same referent, who is also referred to by the personal name Heath. Moreover the order of the two nouns could be inverted (a leading share dealer specialist). However, it is more plausible to take specialist to be a premodifier not of dealer but of share. In that case Heath would be a leading dealer for specialist shares. This interpretation is borne out by the rest of the article which makes clear that this particular banker is both a specialist himself and deals with specialist shares, i.e. Japanese shares. The noun phrase Money men in (57) is of course structurally equivalent to, for instance, stock brokers or merchant bankers. They all consist of a noun premodified by another noun. But lexically it is unspecific and vague, and full of associations even for readers who might not be aware what a stock broker or a merchant banker is. Again it is evident that some of the major differences between the different types of newspapers are not structural but lexical. It seems quite unlikely that any of the up-market papers or even the mid-market

The individual styles

163

papers would use the expression money men as a cognate of stock brokers or more generally to refer to all the people working in one of the financial institutions in the City of London. The other three premodified noun phrases of this excerpt show again the predominance of premodifiers in prehead position. All the premodifiers in high-rise office blocks, London stock market and Wall Street crash are restrictive modifiers in prehead position. London and Wall Street are of course prehead names, high-rise is a prehead adjective, and office a prehead noun. The following example from a mid-market paper is both lexically and structurally recognisable as non-up-market usage. (58)

Aussie entrepreneur Alan Bond, DMa, 22.12.87, 25.5

The premodifier Aussie is what dictionaries call an informal word for Australian, which is less likely to be used in an up-market paper. Structurally it is a prehead adjective, which modifies the head noun restrict!vely. It is, however, the absence of a definite article which makes this example even structurally more likely to be from a mid-market or down-market paper. Noun phrase name appositions will be treated in more detail in chapter 9. (59)

his income - a cool £2.5 million a year, Sun, 08.10.87, 23.2

The adjective cool in (59) is an interesting example, which is again structurally unexceptional. In its literal sense, cool is a non-restrictive central adjective that would be gradable and intensifiable, and could be used predicatively. However, the point of interest is the fact that it is here used figuratively, meaning "without exaggeration". Thus it is an intensifying adjective, and more specifically an emphasizer (cf. Quirk et al. 1985: 429). It seems reasonable to suggest that cool here refers to the state of mind of the person reporting this vast sum of money and makes clear that he or she therefore does not exaggerate the sum because of sheer excitement about it. The last example shows a rank-shifted premodifier. They are exceedingly rare not just in this style, but they do occur occasionally and therefore deserve to be mentioned. (60)

The bank is behaving towards small shareholders with all the disdain and snootiness of the mother-knows-best variety. DE, 06.11.87, 41.2

The premodifier mother-knows-best occupies a slot in prehead position, but it has got its own internal clause structure with a subject and a verb. The term

164

Premodifications

"rank-shift" normally indicates that a clause is used at a different level of syntactic description. I use it more specifically for cases such as (60), in which an element with an internal clause structure occupies a premodifier position, which is, as I argued in chapter 4 above, a structural position that does not allow a fully fledged noun phrase but it allows simple nouns with a small range of modifiers. (60)

a. b. c. d. e.

the snootiness of the [bank] variety the snootiness of the [City] variety the snootiness of the [City of London] variety *the snootiness of the [a/the bank] variety *the snootiness of the [the City of London] variety

The clause mother-knows-best cannot be replaced by a full noun phrase, as shown in (60d) and (60e), but it can be replaced by the entire range of possible premodifiers such as a simple noun in (60a), a proper noun in (60b) or a complex name in (60c). It is considered to be non-standard or rather informal to use a clause in such a position. Quirk et al. (1985: 1336-7) compare this type of premodification with premodification by adverbs and other phrases, and they claim for both of them that the flexibility of this type of premodification tends to be exploited only colloquially, and most examples have (and seem deliberately to have) a flavour of originality, convention-flouting, and provisional or nonce awkwardness. (Quirk etal. 1985: 1336).

7.3.4. The arts section in the up-markets The arts sections represent in many respects an outstanding style. The cluster analysis in the preceding chapter suggested that the arts sections of the five up-market papers are more like each other than any of them is to any other section, in spite of the fact that the distance coefficient between some of the arts section was relatively high. Indeed if the value d for the cluster definition is set at 10, the arts sections form two independent clusters, one consisting of The Independent arts section, and the other of the remaining four arts sections. They only form one single cluster if d is set at 15. The number of premodifiers in itself is not exceptional. The average for all five samples is just under 300, which is considerably more than the previous three styles, but it is below the two styles made up by the home news, foreign news and business sections of the up-market papers. This style is exceptional, however, in that it has a higher proportion of premodifiers in central position than any other style. Over 40 per cent of all its

The individual styles

165

premodifiers are in this position as compared with 28 per cent in the sports sections, which is the style with the second highest proportion of these premodifiers. Correspondingly it has a very low proportion of modifiers in prehead position. Only 40 per cent of its premodifiers are in this position as compared to the sports sections with almost 50 per cent. The remaining four styles have proportions of prehead premodifiers ranging from some 60 per cent to over 70 per cent. In all three styles discussed above, there were substantially more prehead nouns than prehead adjectives. The ratio varied from about twice as many to almost three and a half times as many prehead nouns than prehead adjectives. In the arts section, however, the ratio is almost exactly one to one. There are on average 53.4 prehead adjectives per section and 54.4 prehead nouns per section. This style, then, is quite distinct from the others on syntactic grounds alone. As suggested above, the main reason for this difference lies in the evaluative statements that are made in the arts sections. The journalist's opinion is important, and the descriptions will necessarily be more personal and more subjective using gradable and intensifiable central adjectives rather than the classifying prehead modifiers. The following examples all show how gradable and intensifiable adjectives are used with increasing complexities to describe their head nouns subjectively. Example (61) is relatively simple, it uses ungraded and unintensified central adjectives, even though both could be intensified or graded. (61) (62) (63)

the beautiful Gregory Corso is shaggy and toothless but withered William Burroughs never changes. T, 08.10.87, 20.5 Max Wall was a splendidly gloomy Vladimir, DT, 27.11.87, 20.4 ... was admirably ostentatious in the steely-fingered passagework with which he dispatched Prokofiev's First Piano Concerto. T, 08.10.87, 20.4

Both adjectives in (61), beautiful and withered, modify a personal name. They are clearly non-restrictive, but they add descriptive detail. They are subjective in that a different critic might not agree with these particular epithets. In (62) the central adjective gloomy is intensified by a preceding adverb, which gives a nice semantic contrast. The adverb splendidly as well as the corresponding adjective splendid are usually associated with brilliance, magnificence, or a brightly gleaming expression, whereas gloomy is in some respects almost the opposite, being associated with despair, sadness and dismay. The intention of the author here seems to be to stress that one particular

166

Premodifications

actor performed the gloom and sadness of the character Vladimir so well that it was a pleasure to watch. But premodifiers, as has been pointed out several times, are not very explicit, and this contrast is only hinted at without being made explicit. In example (63) an adjective modified by an adverb appears in predicate position. In this position adjectives and adverbs have not been considered for the present investigation, but similar points can be made about them in this particular example. Both the adjective ostentatious and the adverb admirably are extremely subjective epithets reflecting the journalist's personal opinion. The main interest in this excerpt, however, is the premodifier steely-fingered. In form, it is a past participle of the not very common verb "to finger", which in this case may refer to the fingering on the piano. This participle is modified by steely, which is a denominal adjective by virtue of its ending -y. (cf. Quirk et al. 1985: 1553). Steel itself is often used as a premodifier, as in steel engraving, steel band, steel guitar, or steelwork. In these examples the modification suggests that the material steel is an essential characteristic of the referent. Steel engraving refers to the method of engraving drawings or letters on plates of steel. A steel band plays on percussion instruments made of steel. A steel guitar is steel strung and usually has a body made of steel. And steelwork refers to the frame or foundation of a building made of steel. In steelyfingered, however, such a direct relationship does not hold. Neither the fingers of the pianist nor the keys of the piano are made of steel, and the modifier must be understood metaphorically, describing the pianist's playing as - presumably - very clear-cut, sharp, precise, and perhaps also cold. The possible negative connotations are partly set off by the preceding description of the pianist as admirably ostentatious, but this phrase is somewhat ambiguous in itself. The following verb dispatch, however, suggests that the pianist's main aim was - according to the critic - to get the concert over and done with as soon as possible. Whether the analyses offered for the above examples correspond exactly to the critics' intention or not, they make clear that the arts section of the upmarket papers at least sometimes reach a subtlety that remains unmatched in the down- and mid-market papers, and even in the other sections of the upmarket papers. In (62) and (63), this subtlety is achieved by the clever juxtaposition of lexical items that express opposite meanings. In the statistics presented above, this is reflected by the high proportion of central adjectives, which can be graded and intensified. Surprisingly there is only a marginally higher proportion of central adjectives that are actually intensified or graded in the arts section. The following extracts are relevant examples.

The individual styles

(64) (65) (66)

167

this most celebrated of Soviet orchestras, T, 08.10.87, 20.3 the most hugely successful failed human being, G, 31.10.87, 14.3 a most subtly shaded account of the Prokofiev symphony, I, 06.10.87, 16.2

All three examples, (64) to (66), include a superlative, but in (65) and (66) it is not actually the adjective, successful or shaded, that is graded, it is an adverb which modifies this adjective. In (65), there is an indication of an extreme degree, which I described as typical of the down-market papers. The syntactic means employed here are of course different, but the main difference is the fact that an adjective which is intensified to an extreme degree is immediately relativised by the following adjective failed. The hall-mark of the arts sections, however, as far as their premodification structures are concerned, is their sometimes astonishing complexity which can be found. The following examples give some idea of the possibilities and how they are exploited. (67) (68)

the row of lurid paint-by-numbers sunsets that mark the low point in this show, G,31.10.87, 14.7 its half-baked tunes and cafe orchestration, I, 16.11.87, 15.5

Excerpts (67) and (68) give an indication of how journalists in the arts sections search for unusual and innovative modifiers. The premodifier paint-bynumbers suggests the prolific nature of the artist's work. The term moreover indicates that in the critic's opinion the sunsets painted by the artist are so similar that they can only be distinguished by their numbers. These sunsets are described as being the low point of the exhibition reviewed in this particular article. But the critic's objections to them are only given - in less than explicit form - in the premodification of the noun sunsets. Excerpt (68) refers to an early symphony by Tchaikovsky, of which the writer is clearly critical. The premodifier half-baked metaphorically suggests that the tunes are not mature enough. They presumably were not given long enough to develop. In the same way that bread which is not thoroughly baked is unpalatable, these tunes are unpalatable to the writer. The noun orchestration is modified by cafe, which keeps up the food metaphor albeit in a rather extended sense. It presumably suggests that this symphony is orchestrated in a fashion similar to the popular tunes that are played in such public places as cafes. The following examples show a range of increasingly complex premodifications.

168

Premodifications

(69)

his massive two-courtyard wide, three-story high masterpiece, G, 31.10.87, 14.4 With uneven, often unsupported tone and slithery intonation, T, 20.11.87, 19.1 in Faustus, an intelligent, coherent, if stubbornly unmoving adaptation of the play by the Actors Touring Company at... DT, 27.11.87, 20.1 Rivera was an intoxicating (and by all accounts often intoxicated) colourist,G, 31.10.87, 14.6

(70) (71) (72)

In (69) there are three premodifiers, two of which are modified themselves by a noun and a numeral. As they indicate the physical size of the masterpiece, they are not to the same extent subjective as many of the other central adjectives discussed in the previous excerpts, but - as in this example - they can be modified and made more precise. Examples (70) and (71) illustrate attempts to arrive at a very precise description through premodifications that are themselves modified. The tone in (70) is uneven and often - but perhaps not all the time - unsupported. The adaptation in (71) is intelligent and coherent but nevertheless unmoving. The adverb stubbornly modifying this last adjective describes the adaptation as obstinate in refusing to be moving. Thus this modification is not only syntactically complex but it is also semantically complex by attaching conflicting epithets to the head noun. The syntactic complexity of example (69), on the other hand, is not matched in a similar fashion by a semantic complexity. The three modifiers there all add to the same notion of great physical size. Example (72), finally, plays on the two distinct although related meanings of the verb to intoxicate. It is first used as a present participle and is thus understood to have the painter as subject. It is Rivera who intoxicates. Therefore it must mean "stimulating" or "exciting". The second use of the verb, however, is in the form of a past participle, and here Rivera is the passive subject of the verb, that is to say he does not cause the intoxication but he is the one who is intoxicated. And in this context it can only be understood in the meaning "inebriate" or "drunk". To summarise the main features of this style, there is a far larger proportion of gradable and modifiable central adjectives in this style than in any of the others. The modifications that are used are more personal and subjective. Whereas the styles reviewed above seemed to aim at descriptions that stress the extreme size, importance or unusualness of the events under description, this style seems to aim at multifaceted, and above all semantically complex descriptions. Events are not seen as unidimensional but as conflicting.

The individual styles

169

7.3.5. The business, home news and foreign news sections in the up-markets In this section I am going to review the two styles comprising all business, home news and foreign news sections of the up-market papers. The cluster analysis presented in the previous chapter suggested that the three foreign news sections of the Financial Times, The Guardian and The Independent form an independent cluster in that these sections are more like each other than any of them is to any other section. The average number of premodifiers is highest in these two styles. There are almost 310 in every 1000 noun phrases in the larger style and more than 330 in the foreign news sections of the Financial Times, The Guardian and The Independent. In both of them there is a very high proportion of prehead modifiers (63 per cent and 70 per cent respectively) and a correspondingly low proportion of descriptive premodifiers (31 per cent and 26 per cent). The proportion of nouns versus adjectives in prehead position is another important indicator of these styles. The majority of styles have a far larger proportion of prehead nouns than adjectives. The only exception so far were the arts sections in which the proportion is almost exactly one to one. In the larger of the two styles under review here, the proportion of prehead nouns is also greater than the proportion of prehead adjectives. There are nearly twice as many adjectives. In the foreign news sections of the Financial Times, The Guardian and The Independent, however, the proportion of prehead nouns is considerably lower than the proportion of prehead adjectives. There are on average 118 prehead adjectives per 1000 noun phrases but only 76 prehead nouns. The first series of examples are all taken from the business sections of the up-market papers. (73) (74) (75) (76)

prime lending rate, T, 08.10.87, 25.2 substantial new oil and gas reserves, T, 20.11.87, 25.5 higher [money market] interest rates, T, 08.10.87, 25.2 [foreign exchange] dealers, T, 08.10.87, 25.2

Examples (73) to (76) give some indication of the need in this style to classify rather than to evaluate as in the preceding style. This does not exclude central adjectives, but they are numerically far less important than the prehead modifiers or classifiers. In (73) both adjectives prime and lending are classifying adjectives in prehead position. They are restrictive in that they specify exactly what type of

170

Premodifications

rate the writer wants to refer to. The adjective lending cannot be taken as a descriptive present participle because the head noun rate cannot be understood as the active subject of the verb to lend. As in the majority of styles, a higher proportion of prehead modifiers tend to be nouns rather than adjectives. The nouns oil and gas in (74) are used in such a way. They modify the head noun reserves. The preceding two adjectives are, however, central adjectives. They both modify the entire phrase oil and gas reserves. In (75) and (76), on the other hand, the modification structure is less straightforward. In (75) there are three premodifiers, a graded central adjective and two prehead nouns, the first of which, market, is modified itself by another prehead noun, money. That this is the correct syntactic interpretation can be established by means of elimination tests: (75)

a. b. c. d. e.

money market interest rates higher interest rates ?higher money market rates ?higher market interest rates *higher money interest rates

As long as complete modifiers are dropped, the result is still grammatical, as shown in (75a) and (75b). The premodifier nearest to the head combines most closely with it, and therefore it is questionable whether it could be dropped without affecting the grammaticality of the remaining phrase. If the embedded prehead noun money is dropped the result is also questionable, but if the head of this complex premodifiers, market, is omitted the result is unquestionably ungrammatical. The foreign exchange dealers in (76) are likewise premodified by a complex prehead noun. The adjective foreign does not modify dealers but exchange. This can be shown by the fact that only (76a) but not (76b) corresponds to the intended meaning of the phrase in (76): (76)

a. dealers in foreign exchange b. foreign dealers in exchange

Both foreign and exchange, however, are prehead modifiers. They classify their respective heads, exchange and dealers, in an objective way. (77)

to sell 40 per cent of Northern Telecom pic, its wholly-owned, London-based subsidiary, to STC, T, 08.10.87, 25.7

The individual styles

(78)

(79) (80) (81)

171

by a rights issue of 5 1/2 per cent convertible redeemable preference shares on the basis of five, at 100p each, for every six ordinary shares held, 7,08.10.87,26.2 a balance of social, residential, commercial and retail developments, G, 17.12.87,21.3 Further consideration of Jaguar American earnings prospects prompted bargain hunting in the market. DT, 18.12.87, 18.4 he would face substantial, though not necessarily unsuperable, barriers, FT, 22.01.88, 18.3

The excerpts (77) to (81) illustrate some more complex examples taken from the business sections of the up-market papers. In (77) there is an apposition of a name and a noun phrase. The noun phrase its wholly-owned, London-based subsidiary has two premodifiers both of which are past participles that are modified themselves. This once again makes the point that the premodifiers are - strictly speaking - not individual words but phrases at a level that is intermediate between full phrases and words. The first of these is modified by an adverb whereas the second is modified by the proper noun London. As in noun phrases, premodification of adjective phrases is restricted in that the premodifying item cannot be a fully fledged noun phrase. The simple proper noun London is possible but a noun phrase with a determiner would be ungrammatical (*its wholly-owned, the capital of Britain-based subsidiary). Both example (78) and (79) show how premodification is able to take on a heavy load of additional information. They also show that it is up to the reader to interpret the relationship that is supposed to exist between the premodifier and the head it modifies. In contrast to postmodifications, premodifiers are usually vague and allow - at least in theory - several interpretations. Hence they require the reader to have more background knowledge in order to interpret these noun phrases correctly. The noun shares in (78) is modified by three preheads, two of which are adjectives: convertible and redeemable. All three are technical terms in business and finance and refer to the status of shares for the purchaser. In postmodifying position, as prepositional phrases or relative clauses for instance, more information would have to be given such as what the shares are convertible into, and how and when they are redeemable (as in shares that are convertible into cash and redeemable on or after ...). However, these details are cumbersome and superfluous for readers who are well versed in these details and do not need the extra explicitness, whether it is what I have suggested or something else. Example (79) is fairly unusual because there are four premodifiers in a row. All of them are in prehead position. They restrictively classify the head noun

172

Premodifications

developments. Three of them are adjectives social, residential and commercial. They could potentially all appear in predicate position (the development was social, residential and commercial). But the fourth retail is a noun in prehead position (*the development was retail). A further example of a fairly complex premodification structure with lacking explicitness is given in (80). Here it is the noun prospects that is modified by three preheads, a proper noun, Jaguar, an adjective, American and a nominalisation, earnings. The intended meaning is spelled out in (80a):

(80)

a. the prospects of the earnings that Jaguar is likely to make in America

Excerpt (81), finally, illustrates a type of premodification that I argued was typical for the arts section. The noun barriers is modified by the central adjective substantial and by the modified central adjective unsuperable, with the conjunction though indicating the possible contrast between the two. In contrast to the examples given for the arts section, the premodifiers in (81) are not evaluative. Even though the exact degrees might be a matter of opinion, they are less subjective than the majority of premodifiers quoted for the arts section. However, none of these examples should in any way be taken to be exclusive. Hardly anything that has been illustrated for one particular style does not also exist in all the other styles. What differs is the frequency with which these constructions are used in the proposed styles. The following examples illustrate the home news sections of the up-market papers. Much of what has been said above of the business section applies here too. As the tables with the statistics above suggest, there are numerical differences between the business sections and the home news sections in the upmarket papers, but they are not very great. (82) (83) (84) (85) (86) (87) (88) (89) (90) (91) (92)

the historic Washington summit, DT, 27.11.87, 1.2 international environmental protection, T, 08.10.87, 2.2 at the prestigious international display, DT, 18.12.87, 2.4 A leading neurological paediatrician, T, 08.10.87, 3.1 the occasional older, noisier jets, T, 08.10.87, 7.1 company health and pension schemes, DT, 27.11.87, 16.2 at chairman and managing director level, DT, 27.11.87, 16.2 a most constructive, imaginative and useful input, DT, 27.11.87, 16.2 during the early Thatcher years, DT, 18.12.87, 4.7 light aircraft accidents, T, 08.10.87, 7.2 Chip shop fruit machines, G, 17.12.87, 5.3

The individual styles

173

These examples are not entirely typical to the extent that all of them have multiple premodification, but they are indicative of the range of premodifiers that do occur. In examples (82) and (83) there are only prehead modifiers, whereas in (84) and (85) the prehead modifiers are preceded by the central adjective prestigious and the postcentral present participle leading respectively. In (88) and (89) the premodifiers are paratactically linked by and. In (88) two prehead nouns are linked, the second of which is itself premodified by managing, whereas in (89) and links the last in a row of three central adjectives, which are all to be understood as superlatives. To avoid this interpretation the writer of (89) would have had to rearrange the order of the premodifiers in order to take the other central adjective out of the scope of most (a useful, imaginative and most constructive input). Examples (90) to (92) demonstrate again how much background knowledge is required on the part of the reader for the correct interpretation of sequences of premodifiers. In (90) the reference of the head noun years is restricted by the proper noun in prehead position Thatcher. The adjective in precentral position, early, restricts the reference still further. It thus modifies the entire phrase Thatcher years. In (91), on the other hand, a different structure obtains. The adjective light does not modify the phrase aircraft accidents, it classifies only the noun aircraft. The combined phrase light aircraft in turn modifies the head noun accidents. In this instance light is not a central adjective but a classifier in prehead position. It defines a type of aircraft and is in this usage not gradable (*lighter aircraft accidents). In (92), finally, the head noun machines forms a compound with the premodifier fruit, and as such is further modified by the compound chip shop. (9la) and (92a) indicate the correct interpretation by means of bracketing. (91) (92)

a. [[light aircraft] accidents] a. [[Chip shop] [fruit machines]]

The following examples will illustrate some of the premodifiers that are used in the last style, that is to say in the foreign news sections of the Financial Times, The Guardian and The Independent. Again there is nothing in this style which does not also exist in other styles. It is merely the frequency of occurrence of particular types of premodifiers and above all the very high overall number of premodifiers in the three sections that belong to this style. (93) (94)

intellectual liberal urban vote, FT, 26.01.88, 3.8 in spite of enormous organisational, financial and credibility gaps, G, 17.12.87,8.1

174

Premodifications

(95) (96)

The Chernobyl nuclear power plant disaster, FT, 22.01.88, 2.5 his principal campaign theme: authenticity, consistency, the what-yousee-is-what-you-get Democrat, FT, 26.01.88, 3.5

Examples (93) and (94) show rather extreme examples with three and four prehead adjectives respectively. As I have pointed out above, this style is the only one that has a higher proportion of prehead adjectives than prehead nouns. (95), however, again shows a fairly complex premodification structure in that there is a double embedding of premodifiers. Structurally it can be interpreted in two ways as indicated by Figure 7.1. In (a), the head noun disaster is premodified by the prehead name Chernobyl and by the prehead noun plant, which is itself premodified by power, which is in turn modified by the prehead adjective nuclear. In (b), on the other hand, disaster is modified by one single but complex premodifier. In this interpretation Chernobyl modifies plant rather than disaster. Semantically both interpretations make sense. It was of course a nuclear power plant in Chernobyl which was affected by the disaster, and the disaster itself took place in Chernobyl. Thus there is nothing to choose between the two interpretations.

Chernobyl nuclear

power

plant

disaster Chernobyl nuclear

power plant

disaster

Figure 7.1. Ambiguous noun phrase structure

Example (96), finally, shows that clausal premodifiers also occur in this style in spite of their extreme rarity. As I have shown above, this type occurs with some sort of regularity only in the down-market and mid-market home news and foreign news sections. However, the example that is attested in this style is a particularly complex one because this clausal premodifier consists of a main clause containing two subordinate clauses in subject and predicate position respectively. This is of course a catch phrase which probably stems from the sales patter of market traders but which is very fashionable at the moment in the relatively modern world of personal computers and in this context refers to computer systems that display computer documents on the screen in the same form as they appear on paper once they are printed out. Thus there are several connotations attached to this epithet that are not made explicit. They may be positive, negative or fairly neutral, depending presumably to some extent on what the reader's opinion is of computers or market traders.

Summary

175

7.4. Summary The classification of the premodifiers in the previous chapter was based on syntactic and in part morphological criteria. The main classification distinguished between the positions they occupied in relation to each other within the noun phrase, either actually - if there were more than one premodifier present - or potentially. These classes were further subcategorised according to morphological criteria, as in the case of postcentral position, which can be occupied by present or past participles. And in the case of the prehead position, it was word class criteria that distinguished between subclasses. These criteria were correlated systematically with semantic differences in that premodifiers nearest to the head were found to be most objective, restrictive and classifying, whereas premodifiers further away tended to be more subjective, very often non-restrictive, and descriptive rather than classifying. The most important opposition turned out to be the descriptive, that is to say the central and postcentral, premodifiers versus the classifiers in prehead position. The arts section and the sports section rely heavily on descriptive premodifiers, whereas the others and the foreign news section in particular rely on modifiers in prehead position, i.e. on classifiers. Moreover there are significant differences in the proportion of prehead modifiers that are either nouns or adjectives. These were the main results as revealed by the statistics. Thus the statistics provided a very important first step in the analysis of the data. It suggested a way of clustering the data which might not have been obvious otherwise. It gave clear patterns of preferences that are followed by individual sections in individual papers, but it did not produce any patterns which are restricted to one section or to one type of paper only. In a subsequent step, an interpretation of the facts as they were established by the statistics was offered on the basis of the semantic correlates of the preferred and dispreferred types of premodifiers. Thus I suggested that the arts section favours descriptive premodifiers because they are evaluative and subjective, whereas the foreign news section favours the objective classifiers in prehead position. Finally, however, many important characteristics were uncovered in a somewhat less systematic way by discussing a large range of actual examples. The down-market papers, for instance, are characterised by many nonrestrictively premodified names and by a large number of alliterative patterns that combine head nouns and their premodifiers. The examples also showed that the down-market papers rely on a restricted set of premodifiers. Especially

176

Premodifications

words referring to persons are often modified by a small range of stereotypical premodifiers. The up-market papers, in particular in the arts sections, use a large range of premodifiers including unusual and innovative ones. In many of the cited examples, the premodifiers qualified and described the head nouns in a complex way by hinting at inherent incongruities in the referent. Thus what in some cases appear to be only slight statistical differences turn out to be fairly important if looked at in detail. It has repeatedly been stressed that premodifiers are less specific than postmodifiers. They leave many semantic relationships to be inferred rather than stating them explicitly. This of course magnifies potential difficulties of interpreting the more complex types of premodifiers and their more complex structures in the up-market papers. In many instances, finally, more than one syntactic analysis could plausibly be offered for one single construction. This reiterates one of the major advantages of a linguistic analysis taking a descriptive corpus study as its starting point what in many grammatical models appear to be clear-cut choices turn out to be fluctuating decisions if they are applied to real data as opposed to data invented for illustrative purposes. This may produce "messier" results than the neat models which account only for the invented examples, but it is very likely to produce more adequate results. However, it also makes clear that statistics can be no more than a first step in the analysis because inevitably the figures used for the statistics are based on decisions that had to be taken one way or the other in spite of possible indeterminacy.

8. Postmodifications In this chapter I shall present a survey of the different types of postmodifiers. The criteria used to classify postmodifiers are again mainly syntactic, but they do not depend on the ordering in potential or actual strings of several postmodifiers. As a general rule, postmodifications are syntactically more complex and semantically more explicit than premodifiers. Whereas the individual premodifiers usually consist of single words, even though they can be and sometimes but not frequently are modified themselves, the postmodifiers are more often than not fully fledged phrases. The syntactic details of all the most common postmodifiers were reviewed in chapter 4 above (for a summary of all the categories used in the analysis see appendix I). Postmodifiers are more specific and restrictive than the premodifiers, which serve a categorising and class-forming function. Varantola (1984: 127f) points out that this contrast is often used in single texts. Referents may be introduced by explicit postmodifications and subsequently referred to by potentially ambiguous premodifiers that are disambiguated by the previous mention. She suggests that this phenomenon can be seen as an "instance of ad hoc lexicalization within a text" (1984: 128). This chapter will roughly follow the structure of the previous one. In the first section, I shall present the statistical data showing the frequency of the various constructions and their variation across the individual newspaper categories and sections of my sample. Readers not interested in the statistical details are again advised to skip this part and turn to the following section, in which the relevant findings will be summarised. The remaining part of this chapter will then be devoted to a close analysis of relevant examples in order to see whether, and if so in what way, they are stylistically significant.

8.1. Types of postmodifiers In chapter 6 above, I showed that there are considerable differences in the average numbers of modifiers between the three newspaper categories. The down-market papers have on average just under 400 modifiers per 1000 noun phrases, the mid-markets have roughly 450 and the up-markets 550. In the same chapter I also pointed out that the relative proportion of premodifiers and postmodifiers - perhaps surprisingly - does not vary significantly either across the three newspaper categories or across the five sections. In most indi-

178

Postmodifications

vidual sections just over half of all modifiers are premodifiers. The relative share of some of the subtypes of the premodifiers, on the other hand, does vary significantly, as I showed in chapter 7. Considering the considerable differences in explicitness between the individual types of postmodifier (cf. chapter 4 above) it seems reasonable to expect that the different newspaper categories and the different newspaper sections will vary significantly in the use of individual types of postmodifiers. However, there is some evidence that this might not be the case. Varantola (1984), who compared a corpus drawn from professional engineering journals and a reference corpus consisting of general journalism, found that premodifiers are more stylistically salient than postmodifiers. The overall rate of postmodification per noun phrase turned out to be almost identical in both corpora (1984: 132). Her data unfortunately is not directly comparable to mine because she distinguishes between "slots" in the sequence of postmodifiers and gives the relative share of individual types of postmodifiers for the individual slots rather than for all postmodifiers, i.e. she gives the relevant figures for all postmodifiers that appear immediately after the head noun, and independently for all postmodifiers that appear in the second slot after a first postmodifier. The differences are only slight but they seem to be fairly consistent. Varantola (1984: 140) concludes that "the main tendency is that the general texts often opt for a semantically more explicit type." This points to a potential paradox. The syntactically most complex postmodifiers (i.e. relative clauses and appositional clauses) are also the semantically most explicit modifiers, whereas the syntactically simple postmodifiers such as prepositional phrases are semantically less explicit. Thus there might be a conflict between the wish of a newspaper to be semantically explicit and the wish to be structurally simple. I will show below how this conflict is solved. As in the previous chapter, the tables to be presented here will contain percentage figures rather than absolute figures. They will generally show the relative importance of one particular postmodifier type among all the postmodifiers of one particular sample. The absolute figures, which are given in appendix III, are of course dependent on the varying numbers of modifiers, that is to say two sections with an identical absolute number of relative clauses, for instance, would still be very different in their relative proportions of relative clauses if there was a difference in the overall number of modifiers. Märdh (1980: 73), in a study on newspaper headlines, found the following frequencies of postmodification in nominal headlines. These results are not directly comparable to mine because of the differences in the headlines, as investigated by Märdh, and the copy text analysed for this study.

Types of postmodifiers

179

Table 8.1. Frequency of different types of postmodification in nominal headlines of The Times and the Daily Mirror

Type of postmodification (I) Finite verbal (II) Non-finite verbal (III) Non-verbal

(i) Restrictive relative clause (ii) Appositive clause (i) Present participle clause (ii) Past participle clause (iii) Infinitive clause (i) Prepositional phrase (ii) Adverb

Totals

Times N 4 1 2 20 271 13 311

% 1.3 0.3 0.6 6.4 87.1 4.2

Daily Mirror N % 17 3 1 6 127 3 157

10.8 1.9 0.6 3.8 80.9 1.9

Source: Märdh ( 1980: 73).

The absolute figures are fairly small and therefore do not warrant far-reaching conclusions. The most important difference between the two papers, The Times and the Daily Mirror are the percentages of prepositional phrases and restrictive relative clauses as they appear in nominal headlines. 87 per cent of all postmodifiers in The Times are prepositional phrases as compared to 81 per cent in the Daily Mirror. The restrictive relative clauses, which appear to be exceedingly rare in The Times, make up ten per cent of all the postmodifiers in the nominal headlines of the Daily Mirror.

8.1.1. Non-verbal postmodifiers The non-verbal postmodifiers are numerically by far the most important subtype of all postmodifiers. They comprise the prepositional phrases, which are not just the most important type of postmodifiers in this category but among all postmodifiers but they also comprise postnominal adjectives, nouns and names. Table 8.2 shows the frequency of all non-verbal postmodifiers in relation to all postmodifiers. There is relatively little variation across the newspapers and the sections. Over the entire sample almost exactly three quarters of all postmodifiers are non-verbal. For all the subsamples that consist of more than one section, the percentage is very stable within a range from about 71 per cent to 78 per cent. Individual sections have percentages that are somewhat higher and lower than that. For the three newspaper categories a fairly unclear picture emerges. The percentage is lowest for the up-market papers and highest for the mid-market papers with the down-market papers somewhere in between. But there is such extensive overlap between the three categories that a significant variation

180

Postmodifications

Table 8.2. Non-verbal postmodifiers in per cent of postmodifiers. arts

business

foreign news

home news

sports

all sections

72.63 78.60 74.40 75.39

76.73 71.97 72.96 73.89

74.50 75.81 72.93 74.22

DMi Star Sun all down

71.71 71.71

DE DMa To all mid

75.34 74.52 74.75 74.88

75.78 69.62 71.24 72.16

72.20 79.10 85.45 79.00

80.82 76.70 76.24 78.05

76.09 74.77 76.89 75.91

73.99

DT FT G l T all up

76.42 72.20 67.22 76.35 71.54 72.86

81.34 80.14 74.61 76.54 77.54 77.98

71.59 71.73 74.09 74.56 72.86 72.97

77.04 80.48 70.03 68.70 67.82 73.02

73.68 66.53 71.90 71.41

75.86 76.29 71.90 72.81 72.45 73.77

all papers

72.86

76.45

72.70

75.03

74.02

74.37

across the newspaper categories is ruled out. The variation that does exist must be entirely due to sampling error. A 1-way analysis of variance (ANOVA) confirms this intuitive impression derived from the data. If the test is applied to the absolute figures underlying the percentage figures given in table 8.2, the three socio-economic newspaper categories as independent factor turn out to be highly significant at the 0.1 per cent level (p = 0.0001). This means that there is considerable variation in the absolute number of non-verbal postmodifiers across the three newspaper categories. However, this is not the relevant question to be asked. I have established above that the entire number of modifiers varies significantly across the newspaper categories. The varying numbers of this particular type of postmodifier is therefore very likely to be a direct result of the varying total number. It is of interest, however, to see whether the relative proportion of this type of postmodifier varies significantly across the newspaper categories. This is clearly not the case, as pointed out above. A 1-way ANOVA over the ratios of non-verbal postmodifiers per total number of postmodifiers for each section shows that the newspaper category as an independent factor is not significant. The variation across the different sections, however, turns out to be significant at the 5 per cent level (p = 0.0114) for the ratio of non-verbal postmodifiers to the total number of postmodifiers. The ratio is highest in the business section and smallest in the arts and foreign news section, even though this pat-

Types of postmodifiers

181

tern is not perfect and not followed by all individual papers. The home news sections tend to have a higher proportion of non-verbal postmodifiers but again there are newspapers that differ from this pattern. Tables 8.3 and 8.4 give an indication of the frequency of the nominal and the prepositional non-verbal postmodifiers respectively. They are given in per cent of the total number of postmodifier. Table 8.3 gives the relevant figures for the nominal postmodifiers, that is to say postnominal adjectives, nouns and names. Table 8.3. Nominal postmodifiers (adjectives, nouns and names) in per cent of postmodifiers. arts

business

foreign news

home news

sports

all sections

18.42 20.93 22.02 20.42

15.09 17.83 19.50 17.47

16.91 19.62 19.55 18.83

DM/ Star Sun all down

17.56 17.56

DE DMa To all mid

19.28 8.17 13.64 13.83

23.32 14.77 18.58 18.80

20.98 19.90 24.41 21.81

20.55 12.14 13.81 15.68

21.03 13.73 17.85 17.56

16.14

DT FT G I T all up

14.41 19.31 11.62 20.27 9.76 15.34

11.96 12.33 7.81 9.05 14.39 11.21

16.61 11.66 13.50 14.84 12.27 13.77

8.95 8.90 12.54 18.70 14.94 12.58

8.13 13.81 13.22 12.92

13.62 12.88 13.04 15.49 12.97 13.18

all papers

15.34

12.60

15.44

16.63

14.84

15.01

There seems to be a fairly clear tendency for the down-market papers to have relatively high shares of these nominal postmodifiers, and for the upmarkets to have relative low shares with the exception of the arts section in the up-markets which has a higher share than all the other up-market sections. The three subtypes, too, vary significantly for one or the other independent factor. Postnominal names vary significantly across the three newspaper categories but not across the five newspaper sections. Postnominal adjectives, on the other hand, vary significantly across the sections but not across the newspaper categories. Postnominal nouns do not vary significantly across either of these independent factors.

182

Postmodifications

These variables will be reviewed in more detail in the subsequent section, in which I provide the exemplifications for all postmodifier types within the six styles established in the cluster analysis in chapter 6 above. The most important type not just of the non-verbal postmodifiers but of all the postmodifiers is the prepositional phrase. In all samples more than half of all the postmodifiers and in some cases as many as two thirds of all postmodifiers are prepositional phrases. The details are given in table 8.4. Table 8.4. Prepositional postmodifiers in per cent of postmodifiers. arts

business

foreign news

home news

sports

all sections

54.21 57.67 52.38 54.97

61.64 54.14 53.46 56.42

57.59 56.18 53.38 55.39

DMi Star Sun all down

54.15 54.15

DE DMa To all mid

56.05 66.35 61.11 61.05

52.47 54.85 52.65 53.35

51.22 59.20 61.03 57.19

60.27 64.56 62.43 62.38

55.06 61.03 59.05 58.35

57.85

DT FT G I T all up

62.01 52.90 55.60 56.08 61.79 57.51

69.38 67.81 66.80 67.49 63.16 66.77

54.98 60.07 60.58 59.72 60.59 59.20

68.09 71.58 57.49 50.00 52.87 60.44

65.55 52.72 58.68 58.49

62.24 63.41 73.06 57.32 59.48 60.59

all papers

57.51

63.85

57.26

58.40

59.18

59.36

The ratio of prepositional phrases is highest in the up-market papers and lowest in the down-markets with the mid-market papers lying in between. However, the variation between the groups is not wide enough to justify a rejection of the null hypothesis. The newspaper category is not a significant independent factor for this variable. The newspaper section, too, turns out not to be significant for this variable. Hence prepositional phrases, in spite of their pervasiveness, appear to be of little stylistic significance in contrast to the numerically far less important nominal postmodifiers. Before I can exemplify these constructions in the six different styles, I shall turn to the more explicit form of postmodifiers, i.e. to the non-finite and to the finite verbal postmodifiers.

Types of poslmodifiers

183

8.1.2. Non-finite verbal postmodifiers This type of postmodifier is characterised by a non-finite verb possibly with complements. The verb can be an infinitive or an -ing or -ed participle. Table 8.5 gives the details for all non-finite verbal postmodifiers in per cent of the total number of postmodifiers. Table 8.5. Non-finite verbal postmodifiers (past and present participle phrases and infinitive clauses) in per cent of postmodifiers. arts

business

foreign news

home news

sports

all sections

11.58 7.91 12.50 10.47

10.69 7.64 8.18 8.84

11.17 7.80 10.34 9.82

DMi Star Sun all down

10.24 10.24

DE DMa To all mid

10.31 10.10 10.61 10.33

10.31 10.55 19.03 13.27

13.66 7.46 5.16 8.72

6.85 8.25 8.29 7.76

10.23 9.15 11.00 10.12

8.07

DT FT G I T all up

10.92 12.36 12.03 11.15 13.01 11.88

9.09 9.59 10.16 10.29 13.33 10.58

16.24 13.07 11.31 14.13 13.01 13.55

12.84 9.93 17.77 12.17 18.77 14.32

6.70 12.13 11.16 9.64

11.69 11.19 14.27 12.01 13.89 12.18

all papers

11.88

10.48

13.46

12.07

8.88

11.36

Non-finite verbal postmodifiers tend to occupy a greater share of all postmodifiers in the up-markets, and a smaller share in the down-markets. There is, however, considerable overlap between the three newspaper categories and a 1-way ANOVA reveals that the variation in this case is yet again statistically not significant for the independent factor newspaper category, but it is significant at the 5 per cent level for the independent factor newspaper section (p = 0.0114). It turns out, however, that two of the subtypes of this variable show significant variation across the five sections. These are the -ing and -ed participle clauses. The infinitive clauses, on the other hand, show no significant variation. The newspaper category as an independent factor is not statistically significant for any of these three subtypes.

184

Postmodifications

8,1.3. Finite verbal postmodißers The finite verbal postmodifiers are the most complex, not just among the postmodifiers but among the modifiers generally. They have an internal clause structure with a finite verb and all the relevant complements. The two types are appositional clauses and relative clauses. The percentages for both types of finite verbal postmodifiers are given in table 8.6. Table 8.6. Finite verbal postmodifiers (appositional and relative clauses) in per cent of postmodifiers. arts

business

foreign news

home news

sports

all sections

15.79 13.49 13.10 14.14

12.58 20.38 18.87 17.26

14.33 16.40 16.73 15.96

DMi Star Sun all down

18.05 18.05

DE DMa To all mid

14.35 15.38 14.65 14.79

13.90 19.83 9.73 14.58

14.15 13.43 9.39 12.28

12.33 15.05 15.47 14.19

13.68 16.08 12.10 13.98

17.94

DT FT G I T all up

12.66 15.44 20.75 12.50 15.45 15.26

9.57 10.27 15.23 13.17 9.12 11.44

12.18 15.19 14.60 11.31 14.13 13.48

10.12 9.59 12.20 19.13 13.41 12.66

19.62 21.34 16.94 18.95

12.45 12.52 19.38 15.18 13.66 14.05

all papers

15.26

13.07

13.84

12.90

17.10

14.27

The picture is again very similar to the other types of postmodifiers. The newspaper category is not a significant factor for the variation of the variable, but the newspaper section is significant at the 5 per cent level (p = 0.0351). The relative clauses themselves also have significant variation across the sections but not across newspaper categories, whereas the appositional clauses do not show a significant variation for either of these independent factors.

8.2. Stylistic significance of postmodifier variation The same points that were made about the value of statistics need to be made again in this context. Statistically significant variation does not necessarily entail stylistically significant variation. The statistical tests that were used above

Stylistic significance of postmodifler variation

185

have their main value in the fact that they safeguard the analyst against attributing too much weight to variations in the data that numerically appear to be important but are merely a result of random sampling. The categorisation used above - and introduced in chapter 4 - is based on syntactic criteria as for the premodifiers. However, for the premodifiers, the main criteria were distributional relations in the modified noun phrase itself, that is to say the potential or actual ordering of the premodifiers was taken as one of the main criteria. For the postmodifiers, on the other hand, it is the internal structure of the modifier that provides the basis for the categorisation. It is therefore not a straightforward matter to correlate this syntactic categorisation with semantic differences that might be used in an interpretation of the different frequencies with which individual types of postmodifiers are used. The relative shares of only two of the three major types of postmodifiers, the non-verbal and the non-finite verbal, but not the finite verbal, show statistically significant variation across the newspaper sections. The newspaper category, which was shown to be a major factor for the variation in the total number of modifiers, and also for the total number of premodifiers and postmodifiers, is not significant for the choice of specific types of pre- or postmodifiers. Some of the subtypes of the three types of postmodifiers also vary significantly across the sections, notably the -ing and -ed clauses and the relative clauses, but also some types of nominal postmodifiers, which are subtypes of the non-verbal postmodifiers. Numerically by far the most important postmodifler is the prepositional phrase, which accounts for between half and two thirds of all the postmodifiers in every individual sample. It appears, however, that it is exactly the types of postmodifier that are numerically relatively unimportant which show significant variation, whereas the important ones with high shares of postmodifiers do not vary significantly. This suggests that a great deal of caution is called for in the interpretation of the results, because it might be the case that the statistical significance for numerically small variables is a result of their skewed distribution. The numerically large variables, on the other hand, are more likely to have a normal distribution, and the result of the statistical tests is therefore more reliable for these variables. This of course casts doubt on whether the postmodifiers are stylistically significant at all. So far the evidence is not very strong.

186

Postmodifications

8.3. The individual styles In this section, I shall present an analysis of the different types of postmodifier in terms of the six styles which have been established on the basis of the cluster analysis in chapter 6 above. Table 8.7 gives the average number of postmodifiers of each type per section within one style, and it gives the percentage for each type in terms of the total number of postmodifiers. The corresponding figures for the major categories, finite and non-finite verbal, and non-verbal are not given for reasons of space. They can be deduced by adding the figures of the relevant subcategories. Table 8.7. Average number of types of postmodifiers in six styles

total 1 (III) 2 (III) 3 (IVb) 4 II 5 (IVa) 6 (I)

199.4 208.7 208.5 254.2 262.7 280.0

finite rel

app

non-finite -ed -ing

inf

non-verbal pp noun

name

adj

31.2 15.6 26.6 12.7 27.3 13.1 37.4 14.7 26.5 10.1 33.0 11.8

2.9 1.5 2.0 1.0 5.3 2.5 1.4 0.6 5.7 2.2 5.3 1.9

3.3 1.7 7.4 3.6 5.5 2.6 8.2 3.2 9.8 3.7 8.0 2.9

6.0 3.0 8.8 4.2 6.3 3.0 13.4 5.3 14.1 5.4 19.7 7.0

8.4 4.2 6.6 3.1 9.8 4.7 8.6 3.4 9.8 3.7 8.3 3.0

118.0 59.2 115.0 55.1 123.8 59.4 146.2 57.5 164.3 62.6 168.3 60.1

16.1 8.1 29.3 14.1 16.0 7.7 12.6 5.0 12.1 4.6 17.0 6.1

2.2 1.1 2.8 1.3 2.3 1.1 8.2 3.2 2.8 1.1 4.7 1.7

11.3 5.7 10.2 4.9 12.5 6.0 18.2 7.2 17.5 6.7 15.7 5.6

Average number of types of postmodifiers (first line of each row) and in per cent of all postmodifiers (second line). Roman figures in brackets indicate the cluster as given in the dendrogram of figure 6.1. Style 1: sports style 2: mid and down, home news and foreign news style 3: mid and down, business style 4: arts; style 5: up, business, home news and foreign news (excluding foreign news of FT, G and I) style 6: foreign news of the Financial Times, The Guardian and The Independent).

As can be seen in table 8.7, the picture that emerges from this grouping into six styles is not as clear as it was for the premodifiers. However, a closer analysis of the figures shows that the picture is not quite as confusing and unordered as it may appear at first glance. The main contrast lies between style 5, the business, foreign news and home news sections of the up-market papers, and style 2, the home news and foreign news sections of the mid- and downmarket papers. In some instances style 5 contrasts with style 1, the sports sections of all paper categories, rather than with style 2.

The individual styles

187

Thus it is style 5 that has the highest share of prepositional phrases, and style 2 that has the lowest. Style 5, on the other hand, has the lowest share of postnominal names, whereas style 2 has the highest. Style 5 also has the lowest share of relative clauses, but here it is style 1 that has the highest rather than style 2. In the category of non-finite verbal postmodifiers, again style 1 has the lowest share, and style 5 together with style 6, the foreign news sections of the Financial Times, The Guardian and The Independent, the highest share. The 1-way ANOVA test shows that the above grouping into six styles is statistically significant at the 5 per cent level for five of the nine variables, that is for the relative clauses (p = 0.0015), the appositional clauses (p = 0.0196), the -ing clauses (p = 0.0047), the postnominal names (p = 0.0001), and the postnominal adjectives (p = 0.0016).

8.3.1. The sports section This style is the only one which combines sections from all three newspaper categories. It has the smallest average number of postmodifiers as well as the smallest average number of modifiers. There are on average 411 modifiers per 1000 noun phrases. 199 of them are postmodifiers. These figures compare with 499 and 232 respectively for the entire corpus. As I showed in chapter 6 above, the noun phrase modifications in the sports section are shallower than in any other section, that is to say the modifiers are less frequently embedded within each other. This style has a higher percentage of relative clauses among the postmodifiers than any other style, and it has the lowest percentage of non-finite verbal postmodifiers. There are very few participle clauses. The following examples give an idea of the range of postmodifiers that are used in the sports section by all the papers. (1) (2) (3)

his manager at Monaco, Arsene Wenger, DMi, 08.10.87, 30.3 the more serious stuff ahead, I, 06.10.87, 31.2 if they can be thought to have something slightly wrong with them, T, 08.10.87,42.5

In examples (1) to (3), the headnouns are all modified by nominal, i.e. nonverbal postmodifiers. (1) is an example of a noun phrase name apposition which will be dealt with in more detail in chapter 9. This can be seen as a name which non-restrictively modifies the head noun. In (2) and (3) the head nouns are modified by adjectives in postmodifying position.

188

Postmodifications

(4)

Flamboyant trainer Rod Simpson, renowned for the outstanding placing of his runners, DE, 05.11.87, 44.1 the Board's own inquiry launched yesterday, DT, 18.12.87, 23.6 I was given a great amount of self-confidence and determination to beat the handicap, DMa, 18.12.87, 33.4 the chance to revert to his usual place on the left side of midfield, DMa, 19.12.87, 34.4f ..., although one example of forward planning is the provision of sanitary towels. To be worn back to front, inside the flannels, in case of accidents. I, 06.10.87, 31.4 with seven-and-a-half overs to spare, G, 31.10.87, 17.7

(5) (6) (7) (8)

(9)

The examples (4) to (9) show somewhat more complex postmodifiers. They are all verbal in that they include a verb that may have its own arguments. In (4) and (5), the verb appears in the form of a past participle, whereas in (6) to (9) it has the form of an infinitive. The infinitive clause in (6) can be understood both as a separate clause and as a postmodifier. As a separate clause it could be introduced by in order to without any change of meaning. Example (7) is appositional in that the head noun does not occupy a place in the argument structure of the modifying clause. Example (8) is given in a rather unorthodox punctuation. There is a full stop intervening between the head noun and its postmodifier clause, which starts with a capital letter. This is presumably an indication of a marked pause between the head noun which might raise some curiosity. The explanation is momentarily suspended, as it were, in order to heighten the humorous effect. (10)

(11) (12)

(13) (14) (15) (16)

This is always a highly-charged fixture, emotions heightened by the post-war records which show United winning most of the acclaim ..., 1,16.11.87,32.2 "There's not a lot worse can happen to you in one afternoon, except for being run down by a bus", Star, 18.12.87, 25.1 Embury, whose invaluable innings of 70 rescued England from a nightmare 85 for 6 to a respectable 222 for 7 by the close, Star, 17.12.87, 32.4 players whose fitness is in doubt 48 hours before a match, T, 08.10.87, 42.5 All it did in fact lack was a decent one-day pitch. I, 06.10.87, 31.2 DeFreitas is not currently quite the same confident cricketer he was when he left Australia last February, T, 08.10.87, 42.5 the thought that Chelsea might not survive, T, 08.10.87,42.2

The individual styles

(17)

189

[It] stems from the fact that my opponents are always bigger than me, Star, 17.12.87,29.1

The excerpts in (10) to (17) illustrate some uses of finite postmodifiers. (10) to (15) show relative clauses with relative pronouns in different syntactic functions within their clauses. In (10) and (11), the relative pronoun is in the subject position of its clause. (11) is particularly interesting because it uses the zero relative pronoun, which in standard English is only possible for nonsubject positions. In (12) and (13) the relative pronoun is in genitive case. They are non-restrictive and restrictive respectively. (14) also has a zero relative pronoun but in this case it is in direct object position of its clause and therefore perfectly standard. In (15) the relative pronoun, again in zero form, functions as subject complement in its clause. Such uses are very rare in my corpus, but as the above example shows they do occur, and they do not appear to be contrived or stilted. Examples (16) and (17) both illustrate finite appositive clauses, in which the head noun does not function as an argument of the modifying clause. The following examples are more complex. They show postmodifiers in sequences. In some instances they are concatenated and modify the same head noun, while others are embedded within each other. Square brackets indicate the internal structure of the postmodifiers. (18) (19) (20)

(21) (22) (23)

The on-off transfer [of Steve Bruce] [from Norwich City] [to Manchester United] finally went through, To, 18.12.87, 40.1 Since his arrival [at Anfield] [from Watford] [for a fee of £900,000], John Barnes has ..., G, 31.10.87, 18.3 Graham backtracked on a statement [in midweek] [that his young team were not quite good enough to overtake Liverpool], DMa, 19.12.87, 34.5 repaired the composure [of his colleagues] [which, moments beforehand, looked to have been broken decisively], T, 08.10.87, 42.3 There is every chance [that Robson will select a one-off team [based on a determination [to get a victory [against Turkey]]]], DMi, 08.10.87, 30.3 He was there yesterday, looking now older, as was the patrician figure of Jahangir Khan, [father [of Majid]], [uncle [of Imran]] and [the Cambridge bowler [who bowled the ball [which killed the sparrow [which is stuffed and in the Long Room at Lord's]]]], T, 08.10.87,42.6

In (18) and (19), there are three concatenated prepositional phrases. They all modify their respective head noun, transfer in (18) and arrival in (19). In (20)

190

Postmodifications

it is a prepositional phrase and an appositional clause, and in (21) a prepositional phrase and a restrictive relative clause that are concatenated, modifying statement and composure, respectively. (22) and (23) are more extreme examples with fairly deep embeddings of postmodifiers. In (22) the head noun chance is modified by an appositional clause which contains the noun team that is modified by a non-finite participial clause. The noun determination within this clause is further modified by an infinitive clause which in turn contains the noun victory modified by a prepositional phrase. The name Jahangir Khan is modified by three concatenated noun phrases, the first two of which are modified by internal prepositional phrases. The last one, however, is modified by a restrictive relative clause that contains two further relative clauses, one embedded within the other. The triple embedding of relative clauses creates a humorous effect which is presumably intentional. In spite of the increasing complexity of the above examples, none appears to be particularly difficult to comprehend. It is, however, not coincidental that the examples of structurally fairly simple postmodifiers regularly show premodifiers as well, whereas the more complex examples avoid premodifiers almost completely. Hence the resulting constructions are almost exclusively right-branching, that is to say that further modifiers can quite freely be added to the right of the rightmost element of its containing constituent.

8.3.2. The home-news and foreign news sections in the down- and mid-markets This style comprises all the home news sections of the down-market and midmarket papers as well as the foreign news sections of the mid-market papers. As table 8.7 above made clear, the difference between this style and the previous style of all the sports sections is not very marked. On average there are slightly more postmodifiers per 1000 noun phrases (449 as compared to 411). The share of finite postmodifiers among all the postmodifiers is smaller (14 and 17 per cent respectively), and the share of non-finite verbal postmodifiers is correspondingly greater (11 and 9 per cent respectively), while the share of the non-verbal postmodifiers is almost identical in this and in the previous style (74 and 75 per cent respectively). The ratio of modifier embedding is slightly increased in comparison with the previous style but it is still not very high. (24) (25)

McNamee, intelligent and intense, was 14 when his father went out for a drink but..., DMi, 28.10.87, 3.4, home he repaired gaming machine circuitry - similar to that used in timedelay bombs, DMi, 28.10.87, 2.3, home

The individual styles

(26)

191

Passions still simmer, but the missing link, scheduled to be opened in 1972, is now finished, and 1-95, nearly 400 miles longer than the Great Wall of China, is complete. DMa, 19.12.87, 10.5, foreign

Examples (24) to (26) show postnominal adjectives. In (25) and (26), the postmodifiers are fairly complex adjective phrases and not just simple adjectives. This is the reason why they are put postnominally. Postnominal adjectives, as I indicated in chapter 4, section 4.3.5 above, differ from their prenominal counterparts in that they can be seen as elliptical structures based on underlying predicate clauses, as shown in (24a) to (26a). (24) (25) (26)

a. McNamee [who is] intelligent and intense a. gaming machine circuitry [which is] similar to ... a. 1-95, [which is] nearly 400 miles longer than ...

This indicates that it is one of the conditions for adjectives to be used postnominally that they must be able to occur in predicate position. (27)

(28) (29)

They have been talking with interested foreign airlines, including KLM, Belgium's Sabena, Northwest Airlines and American Airlines, DE, 05.11.87, 8.2, home his refusal to appoint a Muslim teacher, Sun, 08.10.87, 6.1, home we'd even given him a fake gun to put in his holster, Sun, 08.10.87, 9.2, home

Examples (27) to (29) show non-finite verbal postmodifiers. In (27) the postmodifier is a non-restrictive -mg-participle clause, in which the head noun is understood as the active subject. The head noun in (28) does not represent an argument of its postmodifying infinitive clause but in (29) it is understood as the direct object of the infinitive clause (see again chapter 4, for more details on this difference). (30) (31) (32)

(33)

the new "get-up-and-go" generation who have built thriving businesses from a shoestring, Star, 18.12.87, 2.5, home a defiant Mr Baker, who told the conference that he would not be deterred, DMi, 08.10.87, 4.3, home leading Soviet historian Georgy Smirnov hinted that it was also time his country thought again about the 1968 invasion of Czechoslovakia, DE, 05.11.87, 4.2, foreign As an employee, the man was an utter disaster whose dereliction of duty was legendary, DE, 23.11.87, 18.2, home

192

Postmodifications

(34)

Mrs Peterson and Watson, whose neighbours included actress Wendy Craig and comedian Tim Brooke-Taylor, DMa, 22.12.87, 3.4, home

Examples (30) to (34) show a range of relative clauses. In (30) and (31) the relative pronoun occupies subject position within the relative clause in (32) a zero relative pronoun stands for a time adverbial and in (33) and (34) the relative pronoun is in the genitive case (cf. Quirk et al. 1985: 366). (30), (32) and (33) are clearly restrictive, the relative clauses are needed to pick out the intended referent, whereas (31) and (34) are non-restrictive. The latter add further descriptive details. (35)

(36) (37)

(38) (39)

(40)

the small fishing community of Stuart, where the highway came to symbolise the struggle between developers and ..., DMa, 19.12.87, 10.5, foreign it was an open secret among BBC staff that they were in love, To, 19.12.87, 1.1, home a memo written by Admiral John Poindexter, which was released by the investigating congressional committee into the Iran-Contra affair. DMa, 18.12.87, 10.1, foreign Michael Fish, who will go down in history as the forecaster who denied the hurricane was coming, To, 18.12.87, 5.1, home the Interstate system, a 42,500-mile network of super highways created by President Eisenhower that, along with telephones, television and the jet plane, have reshaped American culture. DMa, 19.12.87, 10.5, foreign Former model Barbro Peterson, [40-year-old widow [of Swedish Grand Prix star Ronnie Peterson [ - who was killed in a race - ]]] was slumped in the bath at her "600,000 Tudor-style house ..., DMa, 22.12.87, 3.1, home

The modifications in the examples (35) to (40) are increasingly complex, and accordingly less typical of this particular style. In (35) and (36) a relative clause is preceded by a prepositional phrase, and in (37) it is preceded by a non-finite -ed participle clause. In all these cases the two modifiers are concatenated rather than embedded within each other. They modify the same head noun in each case. In (38) a restrictive relative clause is embedded within a non-restrictive one. The embedded relative clause modifies a head noun which has the same referent as the subject, that is to say the relative pronoun, of this clause, and ultimately the same referent as the head noun of the superordinate relative

The individual styles

193

clause, i.e. the name Michael Fish. In (39) the head noun system is modified first by an apposition with the noun network as its head. The apposition is full, strict and non-restrictive (cf. Quirk et al. 1985: 1305), that is to say both appositives are omissible without loss of grammaticality; they are of the same syntactic class; and they contribute independent information. The first appositive picks out the correct referent, while the second adds additional information without restricting the reference set up by the first appositive. The two postmodifiers, the -ed participle clause created by President Eisenhower and the relative clause that ... have reshaped American culture, are syntactically embedded within the prepositional phrase modifying the second appositive, i.e. of super highways. The relative pronoun that has the function of the subject in its clause and therefore is responsible for the form of the verb, which is have rather than has. Hence the relative pronoun and its antecedent must be plural, which leaves highways as the only possibility. Semantically there would be little difference if both the relative clause and the -ed participle clause were understood as modifying either system or network. The excerpt in (40), finally, shows a fairly complex example of modifier embedding. It is therefore rather untypical of this style even more so in view of the fact that it is a subject noun phrase that is modified in this way. The subject, in fact, consists of a noun phrase name apposition made up by the descriptive appositive former model with a zero determiner and the name appositive Barbro Peterson. This is an instance of a partial, strict, restrictive apposition because the descriptive noun phrase could not stand alone. As I will show in detail in chapter 9, this is the typical form of noun phrase name appositions both in the down-market papers and in the mid-markets, of which (40) is an example. This noun phrase name apposition is modified by a further appositive with the head noun widow. This appositive is a non-restrictive addition. It provides additional information on a referent that has been sufficiently identified. The noun widow in turn is modified by another noun phrase name apposition, i.e. Swedish Grand Prix star Ronnie Peterson, which follows the same pattern as the superordinate one in that it also lacks a determiner. This noun phrase name apposition is finally modified by a non-restrictive relative clause. The above range of examples include very typical and fairly simple structures as well as more complex examples that are less typical of the language of the home news and foreign news sections of the down-market and midmarket papers. It is no coincidence, however, that all the more complex examples are taken from mid-market papers.

194

Postmodifications

8.3.3. The business section in the down- and mid-markets The business sections in the three mid-market papers and in the one downmarket paper that regularly features such a section, The Sun, were set up as an independent style because the cluster analysis presented in chapter 6 above suggested that they are more similar to each other than to any other section. In the previous chapter I pointed out to what extent the premodifier variables are different from other styles. The differences in the postmodifier variables are less obvious. The average number of postmodifiers is almost identical to the previous style (209), but in relational terms this means that the share of postmodifiers is smaller because this style has a higher average number of all modifiers (483 as compared to 445 in the home news and foreign news section of the mid- and down-markets). There are more finite verbal postmodifiers and fewer non-verbal ones, while the infinite verbal postmodifiers have almost identical shares in the two styles. The following examples may suffice to give an idea of the complexity of the noun phrases in this style. (41) (42) (43) (44) (45)

(46)

(47)

(48)

to pacify MPs uneasy about the great new privatisation experiment, DMa, 19.12.87, 29.1 rival suitor Atlantic Richfield took advantage, To, 19.12.87, 33.2 oil giant ARCO, the would-be rival bidder for Britoil, DMa, 22.12.87, 25.5 And by last night, it stood at $ 1.7810 against sterling, a mammoth drop of almost 4 cents in just 24 hours. DE, 06.11.87, 41.5 Torquil Norman, chairman of Bluebird Toys, the biggest UK-owned toy-maker famous for My Yellow Teapot, A la Carte Kitchen, and most recently Manta Force space vehicles, says..., DE, 23.11.87, 30.3 That other top earner, Sir Ralph Halpern of Burtons, (£ 1,004,000 a year) who has become a celebrity through his passion for fast cars and girlfriends, Sun, 8.10.87, 23.2f And British Oxygen boss Dick Giordano (£ 772,000 a year) famed for years as the biggest earner in the land - with a love of sleek yachts takes home less than a third of Heath's salary, Sun, 8.10.87, 23.3 Perhaps that is the explanation for the strange story [put about over the weekend] [that Lord Young, Trade and Industry Secretary, has swung round to favouring BA ownership of B-Cal] - a pure flight of someone's fancy, it seems. DMa, 22.12.87, 25.1

The first example again shows a postnominal adjective phrase. As I pointed out above in section 8.3.2, it can be seen as an elliptical rendering of a relative

The individual styles

195

clause with an omission of both the relative pronoun and the copula. (42) to (46) show appositions of different types and of varying complexity. The apposition in (42) is partial, strict and restrictive - according to Quirk et al.'s (1985: 1300-6) classification. This construction is similar to the noun phrase name constructions that will be discussed in chapter 9, but in this case the referent is not a person but a firm. The same type of apposition appears as the first appositive in (43). The noun phrase oil giant in a partial, strict and restrictive apposition to the proper noun ARCO forms the first appositive in a full, strict and non-restrictive apposition. Excerpt (44) gives a rather unusual example of an apposition because the two appositives, strictly speaking, do not have the same referent. The head noun of the first appositive is the currency designation $ with the cardinal number in determiner position, in spite of the orthographic convention of preceding the numeral with the currency designation. The first appositive refers to a specific level of the exchange rate between dollars and pound sterling, whereas the second appositive refers to the difference between this level and some previous level. The intended meaning is of course that this particular level is the result of a "mammoth drop" within the last 24 hours. Appositions also exist embedded within each other. In (45) a personal name appears in apposition to a noun phrase that is modified by a prepositional phrase containing another apposition. The first apposition is partial, strict and non-restrictive, whereas the embedded apposition is full, strict and nonrestrictive. The head noun of the second appositive of the embedded apposition is in turn modified by a postnominal adjective phrase of some complexity. Nevertheless in spite of the successive embeddings the resulting construction is not particularly difficult to comprehend because all the embeddings are added at the right boundary of their containing constituents. Excerpts (46) and (47) are taken from The Sun. Its business section is called "SUN MONEY" and even includes a list of some top share prices. The material that is presented in it differs very considerably from the corresponding sections in the mid-market papers and even more so from those in the upmarkets, but structurally the differences are not quite as dramatic as might be expected. Both excerpts (46) and (47) are examples of noun phrase name appositions which are further modified by two and three postmodifiers respectively. In a kind of apposition, put inside parentheses, the annual earnings of the two individuals are given. In (47) this is followed by a non-restrictive relative clause and in (48) by an -ed participle clause and a prepositional phrase. (47) is unusual in that the participle phrase precedes the prepositional phrase rather than follows it. This and the dash brackets around the prepositional phrase give it special prominence.

196

Postmodifications

Example (48) finally is again a fairly complex noun phrase construction. The noun story, which is the head of the prepositional phrase modifying explanation, is itself modified by a premodifier and two postmodifiers, an -ed participle clause and an appositional clause. The subject of the appositional clause is a noun -phrase name apposition, but this time the name precedes the descriptive noun phrase. The entire construction is followed by a further noun phrase, a pure flight of someone 's fancy, which stands in apposition either to explanation or possibly to story.

8.3.4. The arts section in the up-markets The average number of modifiers in the arts sections is considerably higher than in any of the previous sections. There are on average 552 modifiers per 1000 noun phrases, and 254 of them (46.1 per cent) are postmodifiers. In the previous chapter, I showed that the arts sections in the up-market papers have outstanding values for some of the premodifier variables. They have a higher share of descriptive and central adjectives than any other style. But none of the postmodifier variables is outstanding in this respect. They all lie well within the range of values established by the other styles. The following examples, however, give some idea in what respect this style nevertheless differs from the others in spite of the fact that numerically the differences are not very apparent. (49)

(50) (51)

(52)

the city would surely have been packed with volunteers, art historians and fund-raisers [anxious [to save a cultural heritage [they felt they shared]]]. G, 31.10.87, 14.4 The declared policy is to deal only with the art of the 20th century, British and international. FT, 26.01.88, 17.4 And then there is the building itself, [the block [of Jesse Hartley's magnificent 1840s dock] [on the prime corner site] [that looks across the Mersey to Birkenhead and a few hundred yards along to the Liver and Cunard Buildings at the Pier Head]. FT, 26.01.88, 17.2 In the first of its highlights we see Rivera as a fine second generation Cubist, [taking apart likenesses of his friends], [flattening them into playing cards] and [putting them back together again in a new order, as if the house of cards had fallen down]. G, 31.10.87, 14.4

Example (49) contains a postnominal adjective phrase consisting of an adjective which could be understood as the subject complement of an elliptical relative clause and an infinitive clause which in turn contains another relative

The individual styles

197

clause. This embedded relative clause consists of a subject, they, a verb felt and an appositional clause lacking the complementiser that. The relative pronoun of the clause corresponds to the direct object of the embedded appositional clause. The entire construction is right branching in that the embeddings are always added at the right boundaries of the containing constituents. In spite of the successive embeddings, however, this example does not contribute significantly to the statistics reflecting a high degree of complexity. Because of the chosen coding scheme (cf. appendix I), the three head nouns volunteers, art historians and fund-raisers would be marked as taking a postnominal adjective, and heritage takes both a prehead adjective and a restrictive relative clause, whereas all the other noun phrases are unmodified. Besides the subject noun phrase the city there are two personal pronouns they and a zero relative pronoun. Example (50) shows another noun phrase with a postnominal adjective phrase. In this case there are two coordinated adjectives, British and international. They are in postnominal position for reasons quite different from (49). In (50), but not in (49), the more orthodox prenominal position would have been possible, as is demonstrated by (49a) and (50a). (49) (50)

a. *the city would have been packed with anxious to save a cultural heritage volunteers, art historians and fund-raisers a. to deal only with British and international art of the 20th century

Even without the embedded relative clause, (49a) violates the right-branching principle. The modification of the adjective phrase is added within the noun phrase, between the adjective and the head noun, creating an ungrammatical structure. (50a), on the other hand, violates no such principle and would therefore be possible. In postnominal position, however, the two adjectives are heavily emphasised. Example (51) is also fairly long and complex in a way that is not sufficiently reflected in the statistics. Basically it is a full, strict and non-restrictive apposition. The second appositive is modified by two prepositional phrases and a relative clause. The relative clause is fairly long but structurally fairly simple. The subject consists of the relative pronoun, and the verb is equally short and simple, but the remainder of the clause is fairly long, combining the rightbranching condition with the end-weight principle introduced in chapter 6 above. The noun Cubist in (52) has three concatenated postmodifiers. All of them are -ing participle clauses. Again it is only the last one of these participle clauses which contains a further clause embedded within it.

198

Postmodifications

(53)

James Stirling, whose vision of the modern museum has been the architectural flavour of the decade, FT, 26.01.88, 17.1 was admirably ostentatious in the steely-fingered passagework with which he dispatched Prokofiev's First Piano Concerto, T, 08.10.87, 20.4 The present exhibition was reviewed two years ago in the Pan-Am inflight magazine where it was praised, at the very beginning of its tour, as a collection of easel paintings, portraits, landscapes, still lives, mercifully free of the "blatant propagandist" content of his murals. G, 31.10.87, 14.4 the real lack of intellectual resources [that these kids have] [with which to make their own decisions and build their own lives], T, 08.10.87, 20.2 These do not differ radically in style from the ones [they replace], [which were a collaboration between Walker (costumes) and Henry Bardon (sets)], DT, 18.12.87, 10.5

(54) (55)

(56)

(57)

Examples (53) to (57) illustrate the use of relative clauses in the arts sections of the up-market papers. They range from the fairly simple, in (53) and (54), to more complex ones, in (55) to (57). All these relative clauses are not the most typical ones in that the relative pronoun does not act as the subject of its clause. In (53) the relative pronoun is the genitive determiner of the noun vision, and in (54) it is the head of a prepositional phrase acting as an adverbial. (55) is a good example of how much information can be packed into the postmodification of a noun. Indeed it is reasonable to assume that the critic writing this sentence wanted to say something about how the exhibition had previously been received. The fact that the previous review appeared in a Pan-Am inflight magazine is only incidental. Hence the main informational load appears in the relative clause which is itself of considerable complexity. In (56) and (57), the nouns are both modified by two concatenated relative clauses. In both cases the first one has the relative pronoun as the direct object of the clause, whereas the second has it as head of a prepositional phrase and as subject respectively. (58)

It's Bruce's Blood on the Tracks, the charting of a relationship from expectation ("Ain't Got You") and hopeful singles-bar pick-up ("Tougher than the Rest") through mistrust and suspicion (the single "Brilliant Disguise", vastly improved in the context of the album) to desolation ("When You're Alone") and the eventual redemption of "Valentine's Day", a release powerful enough to bring tears to the eyes of a hardened cynic. I, 06.10.87, 16.2

The individual styles

199

The sentence in (58) contains one of the longest and most complex noun phrases in my entire corpus. The sentence structure is very simple. It consists of the subject pronoun it and the abbreviated copula 's. The remainder of the sentence is one long subject complement, which consists basically of three noun phrases in full, strict and non-restrictive apposition. Any one or two of the headnouns of the three appositives could be deleted without affecting the grammaticality of the resulting construction. The first appositive is simply the title of an album by Bruce Springsteen, Blood on the Tracks. The second appositive with the headnoun charting is the most complex one consisting basically of a head noun and three prepositional phrases. The great complexity results from the fact that the second and the third prepositional phrases contain coordinated nouns that are themselves modified, and most of the nouns in all three prepositional phrases are non-restrictively modified by the title of a particular song on this album thus forming an apposition. In one case the second appositive consists itself of an apposition, the single "Brilliant Disguise'", which is even further modified by an -ed participle clause. With the prepositional phrase of the album, which modifies a noun in adverbial position within this clause, the deepest level of modifier embedding is reached. Finally, there is a third appositive at the level of the entire construction. The album, named in the first appositive and described as the charting of a relationship in the second, is now called a release, with the added description of its likely effects on prospective listeners to this album. This last example is an extreme illustration of the end-weight principle. The subject and the verb together take up only three letters whereas the subject complement consists of 64 words and covers some 12 lines in the original column layout of The Independent. However, it seems likely that many readers might have problems with the correct interpretation of this sentence on a first and fairly fast reading because it violates the right-branching principle, in that the embeddings are not exclusively added at the right boundary of the containing constituent. There are several coordinate constructions within this subject complement in which all the coordinatives are separately modified. Thus it might be difficult for the reader to find the right superordinate level again after the -ed participle clause, vastly improved in the context of the album, in the third bracket. The reader moves down, as it were, through the consecutive embedding, but it is difficult to establish on which superordinate level the construction proceeds, once the end of a particular embedding is reached.

200

Postmodifications

8.3.5. The business, home news and foreign news sections in the up-markets The business, home news and foreign news sections of the up-market papers finally constitute two styles as outlined in the previous chapter. The foreign news sections of the Financial Times, The Guardian and The Independent were shown to be sufficiently distinct from the foreign news sections in the other two up-markets and from all the other sections in the up-market papers to warrant the postulation of an independent style. These three sections have the highest average number of modifiers. There are 611 modifiers per 1000 noun phrases, and 280 of them are postmodifiers. This compares with 499 modifiers and 232 postmodifiers for all 43 samples and with 567 modifiers and 263 postmodifiers for the main up-market style. (59) (60)

(61)

(62)

Heads to roll may include ..., DT, 27.11.87, 10.1, foreign The company is also calling for a review committee to consider a common pay structure and changes in working practices, including several Japanese techniques, G, 31.10.87, 2.3, home Could Mr Rupert Murdoch summon up the financial firepower to launch a £ 2bn takeover bid for Pearson, the British conglomerate which owns the Financial Times? FT, 22.01.88, 18.1, business the decision [on Tuesday] [by ITT, the US telecommunications giant], [to sell its 24 per cent stake in the UK electronics company to Northern Telecom of Canada for £ 447 million]. T, 08.10.87, 25.6, business

Examples (59) to (62) show the great range of complexity that is possible within one type of postmodifier. They all contain nouns that are modified by an infinitive clause. In (59) the clause consists only of an intransitive infinitive without any additional arguments. The head noun is the subject of this clause. In (60) the head noun committee is also the subject of the infinitive clause but in this case there is also a direct object with its own modifications. The direct object consists of two conjoined noun phrases, the second of which is postmodified by an -ing participle clause, which includes another premodified noun phrase. Thus the end-weight principle and the right-branching principle are adhered to. All the successive embeddings occur at the right boundary of the containing category. A further example of the end-weight principle and the right-branching principle is given in (61). It is again an infinitive clause that modifies a nonsubject noun phrase and contains further noun phrase modifications. As in the superordinate clause, it is also the right-most noun phrase which is modified

The individual styles

201

by some fairly complex postmodification. In this case, the proper noun Pearson is the first appositive of a full, strict and non-restrictive apposition. The second appositive is further postmodified by a restrictive relative clause with the relative pronoun as subject of its clause. In (62) the principle of end-weight is demonstrated in that the postmodifiers get increasingly complex, even though they are not embedded within each other. The head noun decision is modified by two prepositional phrases and an infinitive clause. The first prepositional phrase consists of only one noun besides its preposition, whereas the second consists of a full, strict and nonrestrictive apposition. The infinitive clause, finally, has a full argument structure with a direct and an indirect object as well as an adverbial. (63)

(64)

the demand [from environmental groups] [that the limestone needed in the FGD process should not be extracted from areas of outstanding beauty], and [that the slurry left over be dealt with properly], T, 08.10.87, 2.4, home He was blamed for the fiasco of Vitaly Yurchenko, the KGB officer who fooled the CIA into thinking he was defecting to the United States in 1985 but later returned to Moscow claiming that the agency had kidnapped him. DT, 27.11.87, 10.2, foreign

Examples (63) and (64) are further illustrations of the way in which postmodifications are used to cram a great deal of information into one single noun phrase. The noun demand in (63) is modified by a prepositional phrase and two conjoined appositional clauses. These postmodifiers are concatenated and modify the head noun directly. Any one or two of them could be left out without affecting the grammaticality of the remaining construction. Excerpt (64) is another example of the way in which noun phrase modifications are made to carry a far heavier load than the rest of the sentence. The entire sentence consists of a personal pronoun in subject position, a passive verb and then a prepositional phrase that contains a great deal of information in several layers of embedding. The noun fiasco is modified by a prepositional phrase which contains a noun phrase name apposition. In this case the descriptive noun phrase follows the name, and it is further modified by a fairly complex restrictive relative clause. It contains two conjoined main clauses, both of which contain a subordinate clause. (65)

At the heart of the problems facing the health districts is the cumulative underfunding of the pay awards - over which the districts have no control of any kind, DT, 27.11.87, 18.1, home

202

Postmodifications

(66)

A full prospectus is expected to be available today which should clarify the matter further, I, 06.10.87, 20.5, business

Examples (65) and (66) are not particularly complex but they are problematic in that it is not immediately clear what the antecedent of their relative clause is. In excerpt (65), is it the cumulative underfunding or the pay awards, over which the districts have no control? And in (66), is it the full prospectus which clarifies the matter or is it its imminent availability which promises to clarify the matter? In the latter case the relative clause would be understood to have a sentential antecedent, whereas in the former case the relative clause is not embedded within its noun phrase anymore. (67) (68)

In the region of Quilali contras have killed 127 people in the last week, many of them civilians. G, 31.10.87, 8.1, foreign There was loud laughter and cheers from the ladies and gentlemen of the press, many of whom felt that there but for the grace of God went they. G, 31.10.87, 9.1, foreign

Examples (67) and (68) are again fairly unusual. In (67) the postmodifier many of them civilians defies easy classification. It modifies the noun people, but it is separated from it by an intervening adverbial phrase. Moreover it is not entirely clear how this postmodifier has to be analysed. The two parts many of them and civilians stand in a complement relationship. If a copula like were is inserted, the postmodifier is promoted to main clause status and would serve as an independent continuation of the previous sentence. No really satisfactory analysis can be offered however. The situation in (68) is somewhat clearer. The conjoined nouns ladies and gentlemen are modified first by a prepositional phrase and subsequently by a non-restrictive relative clause. The unusual feature here is the quantifier and the preposition in front of the relative pronoun, which turn it into object case morphology in spite of the fact that it is the subject of the relative clause. The following examples are all taken from the foreign news sections of the Financial Times, The Guardian or The Independent. In contrast to the noticeable differences in the use of premodifiers, the postmodifiers do not differ significantly in these three sections from the foreign news sections of The Daily Telegraph and The Times, and from the home news and business sections of all up-market papers. (69) to (72) illustrate noun phrases of average complexity with two or more concatenated postmodifiers. (69)

his speech on bank secrecy, which bankers had expected him to deliver, FT, 26.01.88, 3.1, foreign

Summary

(70) (71)

(72)

203

a compromise package worth about $ 50m, comprising mainly nonlethal assistance, FT, 26.01.88, 3.1, foreign Rebel leaders have also expressed concern at an amnesty bill passed by the National Assembly this week which appears to pardon all deathsquad killers, G, 31.10.87, 8.2, foreign there are reports that the contras are failing to respect the unilateral ceasefires declared by the Government and that CIA-organised air drops have increased. G, 31.10.87, 8.1, foreign

In (69) the postmodifiers are a prepositional phrase and a relative clause in (70) an adjective phrase and an -ing participle clause in (71) a prepositional phrase, an -ed participle clause and a relative clause and in (72), finally, two appositional clauses.

8.4. Summary Postmodifiers are not to the same extent stylistically salient as premodifiers are, in spite of the fact that the absolute numbers of postmodifiers vary considerably and indeed at a level of statistical significance across the newspaper categories and across the newspaper sections. But these differences are a direct result of the varying overall numbers of modifiers. In other words, in two samples of the same size, one from a down-market paper and the other from an up-market paper, it is highly probable that there will be more prepositional phrases in the up-market sample than in the down-market sample. But in relation to all the postmodifiers or indeed to all the modifiers, the respective proportions of prepositional phrases are not likely to vary very much. But again, as in the case of the premodifiers, it turned out to be essential to analyse in detail a considerable range of actual examples. I presented them systematically according to the six styles, which I established in the previous chapter mainly on the basis of the cluster analysis of chapter 6. It turned out that I had to invoke two principles in order to account for the increasing complexity of the examples, viz. the right-branching principle and the end-weight principle. It is now time to formulate them in a more precise way on the basis of all the above examples. The two principles are of course interrelated. They are two aspects of the same structuring principle. The endweight principle was introduced in chapter 6 above, where I gave the statistical evidence for the fact that subjects are structurally "lighter" than nonsubjects. This means that high informational loads can be packed into noun phrases that appear after the verb and preferably towards the end of a clause.

204

Postmodifications

Subjects, on the other hand, which typically precede the verb, tend to be structurally simpler. The right-branching principle stipulates that embeddings of modifiers are preferred at the right boundary of the containing constituent. This allows for long sequences of concatenated and/or embedded postmodifiers if the embeddings are added at the end of the last element in a sequence of concatenated postmodifiers. This principle is schematically illustrated in figures 8.1 and 8.2. b)

Figure 8.1. Schematic representation of a noun phrase with two embedded postmodifiers

Figure 8.1 shows a noun phrase with two embedded postmodifiers. In a) they are added at the right boundary of the containing noun phrase constituent, that is to say after the head noun. In b), on the other hand, there are two embedded postmodifiers that are embedded within the containing constituent. This is theoretically possible in that the first embedded postmodifier could modify a nominal premodifier of the main head noun, but it would be a violation of the right-branching principle. A reader might find it difficult to establish the correct superordinate level again after having reached the end of the last embedded postmodifier. In a) of course there is no need to go back to the superordinate level because the end of the last embedded postmodifier will also be the end of the containing constituent. Figure 8.2 is a schematic representation of a noun phrase with three concatenated postmodifiers. The representation in a) is in accordance with the right-branching principle but b) is not. In a) the last of the three concatenated postmodifiers is further modified by two embedded postmodifiers. This creates no problems because the end of each subsequent postmodifier is also the end of the containing constituent. But in b) there are two postmodifiers embedded at the end of the first concatenated postmodifier of the main noun phrase. This is potentially problematic because it might not be clear to a reader, confronted with a string of five postmodifiers, that the fourth in the row is not embedded in the third but is actually concatenated to the first.

Summary

b)

205

NP

Figure 8.2. Schematic representation of a noun phrase with three concatenated postmodifiers, one of which is further modified

However, it must be stressed again that both the end-weight principle and the right-branching principle are merely strong tendencies. They are not constraints in the sense that a violation would produce ungrammatical results. Moreover, they are of course not restricted to newspaper language. It seems eminently plausible that they apply in a similar fashion to most varieties of English. But the examples discussed in the course of this chapter have shown how these principles help to make the noun phrase into a vehicle capable of carrying very heavy informational loads.

9. Noun phrase name appositions 9.1. Introduction One of the most clearly stratifying features in British newspaper language is the use of noun phrase name appositions. The term apposition itself - as I explained in some detail in chapter 4 above - refers to a construction in which two linguistic units, which typically but not invariably are noun phrases, are related to each other in one clause. They are called appositives and must normally be identical in reference and fulfil basically the same syntactic role within one single clause. If they are not identical in reference, the reference of one of them is included in the reference of the other. If both appositives belong to the same syntactic class, the apposition is called strict, otherwise it is weak. The apposition is full if either appositive could be left out while preserving a fully grammatical sentence, and it is called partial if only one appositive can be left out without violating the grammaticality. Finally, the apposition can be both restrictive or nonrestrictive (see section 4.3.6 above for more details. The terms are adopted from Quirk et al. 1985: 1300-1306 but see also Meyer 1987. Exemplifications will be given below). In this chapter, I am exclusively concerned with appositions that combine a noun phrase and a personal name. The media often relate news about people, and these people have to be referred to by linguistic means. Under normal circumstances, personal names have unique reference, but very often a person's name alone does not enable the reader to identify the designated individual. It is generally his or her role in public life that constitutes the newsworthiness of the individual thus designated. Appositions are of particular interest for the study of social variation in the syntax of newspaper language because there are several distinct patterns that do not differ with respect to their referential meaning. They refer to one single individual, who is equally unambiguously identified by all of them. There are four basic patterns of noun phrase name appositions: (1)

(a) (b) (c) (d)

[DN']Na [0N']Na Na[DN'] Na[0N']

where D stands for a non-zero determinative element, N' for a noun that may or may not be modified itself, Na for a personal name, and 0 for the zero

208

Noun phrase name appositions

article. The symbol N' is of course borrowed from X-bar syntax. Its usage here merely serves as a mnemonic device to indicate that it is not a noun phrase in the sense of the maximal projection of a noun because it does not include the determiner. It may, however, include pre- and postmodifiers. The name itself may consist of an honorific or title, such as Mr, Mrs, Dr, Sir, President or Archbishop, a given or first name and a surname. In patterns (a) and (b), the descriptive appositive is preposed whereas in (c) and (d) it is postposed. In (a) and (c), the determinative element is realised by a central determiner such as a definite or indefinite article, or a possessive pronoun or by a noun phrase in genitive case in (b) and (d) it is realised by the zero article. The three possibilities of realising the determinative element give rise to the following subtypes (a stands for the indefinite (indef) or definite (def) article and gen for a noun phrase in genitive case or a possessive pronoun): (2)

(al) (a2) (a3) (cl) (c2) (c3)

[a(indef) N'] Na [a(def)N']Na [genN']Na Na [a(indef) N'] Na[a(def)N'] Na[genN']

The following hypothetical examples illustrate all the possibilities. They are all identical in reference in that they refer to one single individual. However, for reasons to be expounded below, not all these locutions are equally likely to be used for this particular individual. (3)

(a 1) (a2) (a3) (b) (cl) (c2) (c3) (d)

a Labour MP, Neil Kinnock the Labour Leader, Neil Kinnock Labour's Leader, Neil Kinnock Labour Leader Neil Kinnock Neil Kinnock, a Labour MP Neil Kinnock, the Labour Leader Neil Kinnock, Labour's Leader Neil Kinnock, Labour Leader

If the determinative is realised by a definite article or a noun phrase in the genitive case, as in (a2), (a3), (c2) and (c3), both appositives are equally specific and would suffice on their own to refer unambiguously to one single individual only. This is particularly obvious in cases of well-known personalities and designations that in any one particular historical and geographical context are

Introduction

209

held by one single person only, as for instance "Prime Minister" or "Labour Leader". In a British context and at one particular time in history, both the Labour Leader and Neil Kinnock refer to the same individual. The lower the informative value of N is, the higher is the degree of Nredundancy. In cases [of well-known personalities like President Nixon or Prime Minister Edward Heath] the name itself is sufficient for the reader to identify the person referred to and hence is the primary element, whereas in, say, driver Jim Tweedie the noun, being designatively selective, is the focal item. (Ryden 1975: 36) In the period under investigation, Margaret Thatcher had been the British Prime Minister for more than nine years. This presumably explains why in all the newspapers of my corpus she is almost invariably referred to by either her role or by her name alone but rarely by an apposition including both of them. Other personalities that appear in newspapers, including the Opposition Leader and even long-standing members of the government, are regularly referred to by means of appositions. The definite article - as I pointed out in some detail in chapter 4 above - indicates specific reference for the descriptive appositive alone, even in cases of comparatively little-known figures in public life as in the following example (4)

Mr Simpson, the vicar of Barmston

In the case of an indefinite article the reference of the nominal appositive is less specific than the name, but it always includes the person referred to by the name. (5)

A self-made multi-millionaire, Harry Goodman

The preposed noun is less specific than the personal name because there are probably several individuals who might fit that description, and the one referred to by means of the name is claimed to be one among them. The term specific needs some further elucidation because Quirk et al. (1985: 1308-311) seem to be using it in a slightly different way. They distinguish four types of equivalence appositions: appellation, identification, designation and reformulation, the first three of which are applicable to noun phrase name appositions. An appellation is exemplified by (a2) or (a3), in which a preposed nominal appositive is definite and a phrase like that is (to say) or in other words could be inserted between the two appositives. In an identification, the preposed descriptive appositive is typically an indefinite noun phrase, as in (al). In this case namely may be inserted but not, for instance, in other words.

210

Noun phrase name appositions

A designation is the opposite of both the appellation and identification. Here the descriptive appositive is postposed rather than preposed, and it is usually definite, that is cases (c2) and (c3). In all these cases, Quirk et al. (1985: 1309, 1310) claim that the descriptive appositive is less specific than the name appositive. A distinction has to be made between nouns with a definite and those with an indefinite article. The definite article can only be used if the noun phrase refers to "something which can be identified uniquely in the contextual or general knowledge shared by speaker and hearer." (Quirk et al. 1985: 265). As I have argued in chapter 4, it does not matter whether the referent of the noun can be uniquely identified on the basis of world knowledge or on the basis of the linguistic context, but it is crucial that unique reference is claimed for such nouns. Thus, I want to say that the two appositives are equally specific both in the case of appellations, in which a definite noun phrase preposes a name ((a2) and (a3)) and in the case of identifications, in which a definite noun phrase is postposed to the name ((c2) and (c3)). There is, however, a clear difference between (6) and (7). (6) (7)

Harry Goodman, the self-made multi-millionaire Harry Goodman, a self-made multi-millionaire

In (6), an example of a designation, both appositives have unique reference, the only difference being that the first appositive has direct reference to its real world referent, whereas the second appositive, in this case, may have anaphoric reference to information given earlier in the text, which enables the reader to uniquely identify the referent of this phrase at this particular point in the text or it may imply that we are all supposed to know who he is. In (7), an attribution, which is not strictly speaking an equivalence apposition (Quirk et al. 1985:1313), the postposed noun phrase is indefinite and therefore does not claim to have unique reference. In the present context, however, it adds additional, nonrestrictive information about the person referred to by name in the first appositive. Table 9.1 summarises the various patterns that exist and the categories assigned to them by Quirk et al. (1985) (cf. also Meyer 1987: 113). In the locution a Labour MP, Neil Kinnock (al), the name adds nonrestrictive information to the nominal appositive. The indefinite article implies that the referent of this noun phrase has not been mentioned before and that it is unfamiliar to the reader. It further implies that the noun phrase does not have unique reference. There are several individuals who could fit the description given in the nominal appositive. If the noun phrase has unique reference, the

Introduction

211

Table 9.1. Patterns of noun phrase name appositions type

pattern

full or partial

restrictive or nonrestrictive

terminology

identification appellation sra appellation sra sra attribution designation designation attribution

preposed al a2

[a(indef) N1] Na

[a(def) N ] Na

full full

a3

[gen N1] Na

full

b

[0 N1] Na

partial

nonrestrictive nonrestrictive restrictive nonrestrictive restrictive restrictive

full full full partial

nonrestrictive nonrestrictive nonrestrictive nonrestrictive

cl c2 c3 d

1

postposed Na [a(indef) N'] 1

Na [a(def) N ]

Na [gen N1] Na [0 N']

Terminology after Quirk et al. (1985: 1309-1313). Strict restrictive appositions (sra) are not labelled independently by them.

indefinite article cannot be used. *Λ Prime Minister, Margaret Thatcher is ruled out as well as *a Labour Leader, Neil Kinnock. These locutions are only possible if they are seen in a wider historical context including their respective predecessors and successors, but even then these forms are only likely to be used in connection with further qualifications such as a prime minister of the seventies, Harold Wilson. It is therefore quite likely that public figures, such as politicians, who depend on being known by a large segment of the newspaper reading public - and electorate - may take offence at being referred to by an indefinite noun phrase, because of its implications that the author does not think they are familiar enough to warrant the use of a definite article. The format may indeed be used deliberately to belittle the person concerned. In (a2) a definite article is used as in the Labour Leader, Neil Kinnock, and in (a3) the determinative element is realised by a noun phrase in the genitive case or by a possessive pronoun, as in Labour's Leader, Neil Kinnock or their leader, Neil Kinnock. In both these examples, the name is again a nonrestrictive modification of the descriptive appositive. The nonrestrictiveness is indicated by the presence of a comma between the two appositives. Patterns (a2) and (a3), however, can also be used for restrictive modifications. The differences can be seen most clearly in an opposition like his brother, John and his brother John, in which the presence of the comma suggests a nonrestrictive relationship between the two appositives indicating that there is only one brother, who happens to be called John. The absence of the comma, on the other

212

Noun phrase name appositions

hand, suggests that the relationship is restrictive in that the name specifies the reference by picking out one particular brother out of two or more brothers. In Labour Leader Neil Kinnock, pattern (b), the determiner-slot is empty, which makes the apposition partial because only one appositive (i.e. the descriptive appositive but not the name) could be omitted without violating the grammaticality. This format is always restrictive, whether the descriptive appositive has unique reference or not, and therefore there is no comma between the two appositives. This construction has got the advantage of not having any implications as to whether the referent is familiar or unfamiliar to the audience. Moreover, it is often difficult to distinguish this construction from names with a preceding honorific or title, and this is presumably its major attraction for specific segments of the media. It confers a title-like quality to all kinds of everyday nouns used to refer to people's role in society. There is not just Prime Minister Margaret Thatcher and Opposition Leader Neil Kinnock but also goalkeeper Peter Guthrie, fashion designer Norman Blackwell or pensioner Gerry Doherty . The criteria that are involved in the distinction between honorifics, titles and appositives will be expounded below. In patterns (c) and (d), the descriptive appositive follows the name. The descriptive appositive modifies the name rather than the other way round. As the name in practically all cases has unique reference, the modifications can only be nonrestrictive. The only exceptions are those cases where two individuals share the same name. (8)

I am talking about Steve Levinson the linguist and not about Steve Levinson the economics correspondent of The Independent.

Pattern (cl) with an indefinite article in the descriptive appositive, as in Neil Kinnock, a Labour MP, again implies that the referent is one among many and not familiar to the reader, whereas Neil Kinnock, the Labour Leader or Neil Kinnock, Labour's Leader implies familiarity and/or unique reference. In the case of Mr Simpson, the vicar of Barmston, it is unlikely that familiarity is claimed and therefore it must be a case of unique reference. In Harry Goodman, the self-made multi-millionaire, on the other hand, it is equally unlikely that unique reference is claimed. In this instance it must be a claim for familiarity, either on the basis of contextual knowledge or on the basis of world knowledge. The journalist is thus faced with a potentially awkward decision whether to use the definite or indefinite article. If the descriptive appositive refers to a unique role, there is no problem because the definite article can be used, but if there is no unique reference, the definite article may claim familiarity where there is none for the majority of readers, while the indefinite article may be unpleasant to the individual thus referred to.

The honorific-title-appositive gradients

2 \3

Pattern (d) with the zero article again makes no implications about the familiarity or unfamiliarity of the referent. Harry Goodman, self-made multimillionaire is a possible locution just as Mr Simpson, vicar ofBarmston, even though in the former case the descriptive appositive presumably does not have unique reference but it does in the latter case. This format, as format (b), is an instance of a partial apposition because only one appositive can be omitted without affecting the grammaticality of the remaining construction.

9.2. The honorific-title-appositive gradients In millionaire Harry Goodman, the first appositive is used as if it were a title such as professor in Professor Henry Higgins. However, there are a number of grammatical features that distinguish nouns (millionaire, pensioner, driver, etc.) from titles (Professor, President, etc.) and honorifics (Mr, Ms, Sir etc.). These features can be used to categorise any doubtful elements. Few of them actually provide hard and fast decisions in all cases, but used in combination, they pattern into three more or less distinct groups. The first of these tests ascertains whether a particular item also exists as a vocative. In normal usage honorifics cannot be used in such a way. The vocative mister is an exception and restricted to nonstandard usage (Quirk et al. 1985: 1319). Most titles and some nouns, on the other hand, are regularly used as vocatives such as Professor, Doctor, Prime Minister, Chancellor, Nurse. Some kinship nouns may also be used as vocatives, albeit less easily. Examples: brother, husband, grandmother. The second test ascertains whether an item can itself be modified. Honorifics do not allow this. *The mister who lives next door John Brown is ungrammatical whereas the present Queen Elizabeth, the president of the United States, Ronald Reagan, and the professor for English and Media Studies, Michael Stubbs are all grammatical. Nouns also allow modification whether they are preceded by an article or not, as in financial expert Peter Smith or the pilot of the other aircraft, George Brown. If titles are modified they behave as ordinary nouns, which is usually marked by the fact that they are not capitalised. Thus we have the president of the United States, Ronald Reagan, but President Ronald Reagan. Thirdly, all nouns and most titles can be used independently without a name (examples: the expert, the pilot, the Prime Minister, the Queen, the professor) but honorifics and some titles cannot be used in this way (examples: *the Miss, *the Lady, *the Dr). There is a difference between doctor and its

214

Noun phrase name appositions

abbreviation Dr only the former of which can be used independently or as a vocative. The fourth test marks whether a particular item can occur in postposition as well as preceding the name. Honorifics do not allow this, and titles only allow it if they are modified. *Linda Brown, Mrs ^Randolph, Sir *John Clark, Professor and *Fran£ois Mitterrand, President, are all ungrammatical in the context of a newspaper article, even though they would - with the exception of the second one - be normal at the end of a letter. Ronald Reagan, President of the United States, or Michael Stubbs, Professor for English and Media Studies are both acceptable. Most nouns are acceptable in postposition. Kinship terms again seem to be the exception, unless they are preceded by a possessive pronoun. George, my brother, is acceptable but *George, brother is not in spite of the possibility of brother George. The fifth test concerns the possibility of a preceding determiner, either definite or indefinite. Titles and honorifics do not allow this, whereas nouns normally require a determiner (examples: *the Mr Brown, *the President Reagan, but the financial expert, George Brown, the Prime Minister, Margaret Thatcher). It is exactly the possibility to drop the determiner in front of the preceding descriptive appositives that creates the difficulty in distinguishing nouns from titles and honorifics. Kinship terms are rarely - if at all - preceded by an article but they are regularly preceded by a possessive pronoun. Bell (1977: 377) points out that kinship terms are "semicategorically preceded by a possessive pronoun, which is semantically redundant since the relationship to the pronoun referent is already expressed in the noun itself." The sixth test shows whether an item can cooccur with an honorific. *Dr Mr Smith, ^Professor Mr Clark and ^President Mr Reagan are all ungrammatical, but (the) financial expert, Mr Jim Smith and her father, Sir John are both fully acceptable. An honorific is again acceptable after a title if the title is modified as in the president of the United States, Mr Ronald Reagan. A small number of items may actually shift category altogether. Thus we can have the nurse, Miss Susan Bloggs but also Nurse Bloggs. In the former case there is no capitalisation and an honorific is possible. It is therefore an ordinary appositive. In the latter case the noun is capitalised and the honorific would be ungrammatical. In ambiguous cases like this one, the capitalisation of the noun is taken to be the decisive feature. Capitalisation, then, is the last test. All titles and honorifics are capitalised, and a lack of capitalisation often indicates a shift from a title to an appositive, as in the examples of modified titles quoted above (such as the president of the United States, Ronald Reagan). However, there are also some appositives that

The honorific-title-appositive gradients

215

are capitalised, as for instance political terms of office with unique reference (e.g. Foreign Secretary, Chancellor, Prime Minister, Opposition Leader). Table 9.2 tabulates the relevant criteria of the appositive-title-honorific gradients. The letters a) to g) refer to the criteria outlined above. In summary, they mark a particular item as to whether it (a) (b) (c) (d) (e) (f) (g)

exists as a vocative can be modified itself can be used independently can occur in postposition can be preceded by a determiner can cooccur with honorifics has no capitalisation

Table 9.2. Appositive-title-honorific gradients

appositives expert, pilot, architect nurse, doctor husband, sister Prime Minister, Chancellor

-f + +

+ + + +

+ + + +

+ + +/+

+ + + +

+ + + +

+ + + -

titles Professor, Nurse President, Bishop Queen

+ + -

+ + +

+ + +

-

-

-

-

_ _

_

_

_

_

_ _

_ _

honorifics D r Sir, Lord, Lady M r , M s , Mrs, Miss

_ _

_ _

_ _

_

_ _

_

_

No single test provides conclusive evidence for the categorisation of a particular item. The possibility of a preceding determiner (test e) and the possibility to cooccur with an honorific (test f) are taken to be the most conclusive ones. But it is of course exactly the somewhat title-like quality afforded to the noun by the absence of a determiner and honorific which gives the papers that employ this format their particular flavour. The locutions goalkeeper Peter Guthrie, fashion designer Norman Blackwell or pensioner Gerry Doherty are unlikely to be used with an honorific. The distinction between title and appositive is deliberately blurred. It heightens the importance of a person thus designated.

_

216

Noun phrase name appositions

It implies that this person belongs to a class of human beings as exclusive as the clergy, military or nobility. It implies a uniqueness, even when we know a person's label is very common. (Bell 1985: 98)

9.3. Previous studies Quirk et al. (1985: 1314, fn b 1319) point out that the apposition format with preposed noun and zero article is characteristic of American English journalistic style. Thus they restrict this pattern geographically and stylistically. Elsewhere they say that it is typical of journalistic style without restricting it to one geographical variety (1985: 1305). Matthews (1981: 228) likewise calls this format "journalese". However, this is too general because geographically the usage is not restricted to American English, but neither is it typical of journalistic style in general because it only appears in some of the British newspapers. The most thorough studies of noun phrase name appositions in media language are Ryden (1975) and Bell (1985, 1988, 1991). Bell investigates this construction in terms of a variable rule that deletes the determiner in preposed descriptive appositives according to linguistic and extralinguistic constraints. The three extralinguistic factors that influence the application or nonapplication of the rule are geographical, social and historical (Bell 1985: 106). There is a near absolute distinction between the media language in the United States and in Great Britain (see figure 9.1). 86

100

20

121

126

80

73

203

2A8

75

50

25

c ο

a> αϊ Ο

(Οι

CD

c ο VI

m

c g t_ σ Ο αϊ, Τ3

CD C

m u

Figure 9.1. Percentage of determiner deletion on 4 British and 4 United States prestige media, 1980/82. (Number of actual rule applications in percentage of potential cases). Source: Bell 1985: 105, Fig. 1. Exact figures in Bell 1988: 336, table 1.

Previous studies

217

In the United States the rule applies in 90 and more per cent of all possible cases with the exception of the New York Times, in which deletion applies in 74 per cent of the possible cases, whereas in the comparable media channels in Great Britain, it is less than ten per cent in all cases. The clear social distinction between the tabloids and the broadsheet papers in Great Britain has been pointed out both by Ryden (1975) and by Bell (1985, 1988, 1991). Ryden investigated six British newspapers; The Observer (O), The Times (T), The Guardian (G), The Daily Telegraph (DT), the Daily Mail (DMa), and the Daily Express (DE). The results are summarised in table 9.3. Table 9.3. Overall distribution of noun-name collocations, in per cent (1971)

Collocation

Text 0 (415)

T (590)

G (397)

DT (560)

DMa (342)

DE (426)

A + N + Na 0 + N + Na Na + A + N Na + 0 + N

25.3 18.3 24.6 31.8

13.2 3.1 36.1 47.6

34.4 5.8 27.5 32.3

13.5 8.9 30.3 47.3

5 65.8 7.9 21.3

4.5 70.6 6.1 18.8

Source: Ryden 1975: 32, table 4:1. The four collocational types correspond to the patterns (a), (b), (c) and (d).

In the data that was collected in 1971 the two tabloid papers, Daily Mail and Daily Express, both of which are mid-market papers, use the pattern with preposed descriptive appositive and zero article with great frequency. Roughly two thirds of all noun phrase name appositions in these two papers appear in this form. In the three up-market dailies, on the other hand, this pattern is only of marginal importance. The Observer, the only weekly analysed by Ryden, takes an intermediate position. In terms of Bell's variable rule, the Daily Mail and the Daily Express delete the determiner in 92.3 and in 94.1 per cent of all the preposed appositives. The up-markets, The Times, The Guardian and The Daily Telegraph apply the rule in 18.8, 14.4 and 40 per cent of the possible cases. The Sunday paper The Observer applies it in 42 per cent of the cases. Bell (1985, 1988) gives the percentages of rule application for seven British papers of 1980. He gives two figures for each paper. One figure gives the number of rule applications in per cent of all potential places. But he notes that genitives are more resistant to rule application, that is to say rule application is more likely in the pensioner Gerry Doherty (to yield pensioner Gerry Doherty) than in her husband Tom Stoppard (to yield husband Tom Stoppard). Bell's and Ryden's figures can only be compared with considerable caution because there is some doubt as to whether they counted exactly the same pat-

218

Noun phrase name appositions 20

126

26

91

26

35

191

100

Articles only All determiners

75

50

25

ω οι

.g Ό ία Ο ω·

0

ω ω ω u α. χ L

C

α σι

Figure 9.2. Percentage of determiner deletion in 7 British daily papers, 1980. Source: Bell 1 985: 107, Fig. 2, actual figures in Bell 1988: 337, table 2.

terns. It is not clear, for instance, how Ryden handled the appositives that have a genitive in determiner position. If he included them, the percentage figures are bound to be lower than if he did not include them because of the reluctance of this pattern to delete the determiner. Ryden (1975: 16) points out that the pattern (b) with a preposed appositive with a zero article is on the increase, and in some instances dramatically so. He reports no more than sporadic occurrences of pattern (b) in 1920 in the Daily Mail and the Daily Express. It came into prominence in the late 1940's, and ten years later it was already established as a distinctive feature of both these papers. Bell (1985, 1988, 1991) quotes figures for the Daily Mirror. In 1920, the Daily Mirror did not delete any determiners at all, in 1940 15 per cent were deleted, in 1960 64 per cent and in 1980 94 per cent. The most dramatic increase came in the decade from 1940 to 1950, when the rule application increased by 28 percentage points (Bell 1985: 112, 1988: 339). In 1990, only possessive determiners (his, her, etc) resist deletion. All other determiners are categorically deleted in noun phrase name appositions (Bell 1991: 134). An even more rapid language change took place in the New Zealand media language over a very short period of time. Figure 9.3 shows how New Zealand radio stations increased the amount of determiner deletion in the six years from 1974 to 1980. Bell (1985: 115, 1988: 342) concludes: The prestige media in New Zealand and Britain will be inevitably drawn into deleting more and more of the determiners. ... In Britain, even the Times and BBC Overseas Service will lose their rearguard action and be

Previous studies

219

100

75

1980 1974

50

25

Ο

YA

Figure 9.3. 1984:195)

ZB

ZBR

ZM

XA

XI

Determiner deletion on six New Zealand radio stations, 1974—1980 (Bell

drawn into increasing determiner deletion. By the end of the century, they may well be where the popular press was in 1950. The elite media are linguistically conservative - often deliberately so. Yet they are unlikely to be able to resist the force of the more popular media dragging them into the news language of the twentieth century. It is unfortunate that Bell investigates this process solely in terms of his "determiner deletion rule" because, as I will show in detail below, this approach misses the point that the crucial decision facing newspaper journalists, or their editors, is not whether to use the Labour Leader, Neil Kinnock or Labour Leader Neil Kinnock but whether to use Labour Leader Neil Kinnock or Neil Kinnock, (the) Labour Leader. Ryden (1975: 37f), writing ten years previously, reached a rather more cautious conclusion as regards the further development of this format. Although on the increase in radio and TV news-reports, 0 + N + Na still has limited currency outside newspapers and magazines (at least in British English). And although it has been a regular feature of informal newspaper language for more than 20 years, ..., quality papers like The Times and the Guardian have been remarkably slow in adopting the type. Hence there is reason to believe that, for a considerable time, the 0 + N + Na collocation will remain a differentiating stylistic criterion, distinguishing popular journalese from more conventional and sober forms of British English journalistic prose. He noticed that it is not only the amount of determiner deletion in noun phrase name appositions that distinguish the different types of newspapers. They also differ in their respective preferences for preposing or postposing the descrip-

220

Noun phrase name appositions

tive appositive. The up-market papers strongly prefer postposing, whereas the mid-market papers strongly prefer preposing. This means, of course, that there are far more occasions for potential determiner deletion in the mid-market papers than in the up-market papers, in which preposing is relatively rare. 100

%

90 ΘΟ 70 60 50 ^0 30 20 10 0

DE

υΖΔ post N

DMa

DT

0

ESpre N

Figure 9.4. Preposed versus postposed descriptive appositives. Based on Ryden's (1975: 34, table 5) data.

Post-N prevails as against pre-N in all the papers except DE and DM [i.e. DMa], most markedly in T and DT .... Again, Ο [i.e. the Observer] has the most balanced distribution. (Ryden 1975: 33) Ryden further analyses in detail the structure of the N-part of the noun phrase name appositions, and he finds that simple nouns are by far the most common element of the descriptive appositive in appositions of pattern (b). They account for 42.8 per cent of all such cases. The complex modifications, accounting for 57.2 per cent, split up into many subgroups, none of which accounts individually for more than 26 per cent. However, intuitively these subgroupings do not seem to reveal any statistically significant results (cf. the exact table: Ryden 1975: 20). Ryden also distinguishes between different types of descriptive appositives on semantic grounds. He distinguishes nouns referring to the world of politics, nouns referring to sports and a rag-bag group for all the rest. Such a distinction is largely but not entirely coextensive with my own distinction according to the newspaper section in which a particular noun phrase name apposition occurs. The distinction between the various ways of using appositions to refer to individuals in the context of newspaper reporting has not escaped the attention

Previous studies

221

of at least some newspaper readers. Some twenty years ago a revealing controversy was fought out in the letter section of The Times. This English Language Sir,-The Times Diary refers (March 25) to Mr. Brno Goldfinger, the architect, as "architect Brno Goldfinger" as though his profession were a rank or a title. I know this solecism is becoming common, but I am dismayed to see its adoption by The Times it is common not only in the sense of being widespread but as a vulgarism that has nothing to do with the evolutionary process by which any language must change if it is to survive and remain comprehensible. It is as much a perversion of English as the juvenile habit of substituting "this" for the indefinite article, a practice that is equally common. If The Times is going to fall for one, why not the other - and by degrees all the rest of the grammatical gimmicks that make some other papers so boring to read? Yours faithfully, NICOLAS BENTLEY

(The Times, March 27, 1968, 11)

The next day, the author thus criticised answered in the same column, The Times Diary, and defended the usage of this particular format of noun phrase name apposition. A publisher called Bentley... A PUBLISHER called Nicolas Bentley, who also draws, took me to task in a letter printed in yesterday's Times for referring to Mr. Brno Goldfinger as "Architect Brno Goldfinger". He reported himself dismayed to see this solecism in The Times, and described it as a vulgarism "that has nothing to do with the evolutionary process by which any language must change if it is to survive and remain comprehensible". Now Mr. Bentley is clearly at liberty to regard himself as a more zealous custodian of the English language than me (though I shall read books he has edited for Andre Deutsch or Nelson with a true grammarian's eye). But has he paused for a moment to reflect why we periodically - though not slavishly - employ this formula? A sterile activity It is because it does away with the necessity of making an invidious choice between "the architect Erno Goldfinger", or "Erno Goldfinger, the architect", and "an architect called Erno Goldfinger", or "Erno Goldfinger, an architect". When does an author, painter, politician, actor, and so on, qualify to be called "the" rather than "a"? Arbitrating on such matters is a sterile activity.

222

Noun phrase name appositions

Nor are there many things more irritating in modern journalism than the undiscriminating use of "the" before the occupations of people of limited or non-existent achievement. Even though there is no doubt that Mr. Goldfinger is an architect of exceptional attainments, he can, I am sure, dispense with the pompous bolstering of the definite article. PHS (The Times, March 28 1968, 10) Still a day later, The Times published a letter from a reader who reports on a usage that rejects what most people would regard as a title in this particular position. This English Language Sir,-Mr. Nicolas Bentley's letter (March 27) recalls to me the occasion long ago when, as a probationer fellow of my college, I introduced my formidable senior, the then Professor of Poetry, to a lady as "Professor Ker". "Don't call me 'Professor'", he said: "you would not call me 'Carpenter Ker' or 'Pawnbroker Ker', had such been my trade." I suppose the rebuke would be unintelligible to an addict of the popular press today. I am, Sir, your obedient servant, DERMONT MORRAH

(The Times, March 29, 1968, 11)

These letters make clear that at least some members of the newspaper reading public are aware of the significance of the pattern architect Erno Goldfinger, which indicates that we are dealing with a marker rather than an indicator. The author, at least in some instances, makes a conscious choice, and the reader is aware of that choice. However, there are important differences in the perception of this choice. For Nicolas Bentley, the initiator of the exchange, the pattern is a solecism and a vulgarism. He deplores its adoption by The Times because he deems it to be substandard and therefore not suited to the level of language he expects from The Times. Thus he must clearly see it as a stylistic possibly even social marker. PHS, the author of the article that gave rise to Bentley's letter, on the other hand, defends its usage on semantic grounds. It is a useful pattern because it does not force the journalist to take a stand on the likely familiarity or unfamiliarity of a particular personality to a particular audience. For this author, then, the pattern architect Erno Goldfinger does not have strong stylistic or even social implications. The last letter quoted above is also interesting because it shows that professor was understood, at least by Dermont Morrah's senior colleague, as an appositive and therefore inadequate in the particular social context of being introduced.

Linguistic constraints

223

9.4. Linguistic constraints In the following, I shall present the results as they are attested in my data. Before I can show the social stratification of the individual patterns, however, it is necessary to ascertain what the linguistic constraints are that influence the choice between the existing patterns of noun phrase name appositions. This will make it possible to distinguish those effects that are common to all newspapers, and possibly valid beyond the language of the media, and those effects that vary between the three categories of newspapers or even between individual newspapers.

9.4.1. The semantic constraints So far I have assumed that the various patterns of noun phrase name appositions have got the same referential meaning in that they may refer to the same individual. This is, of course, a slight simplification which needs to be qualified. If the article is not zero in the descriptive appositive, it carries some crucial information, as has been outlined above. Thus the two locutions a selfmade multi-millionaire, Harry Goodman and the self-made multi-millionaire, Harry Goodman are not synonymous even though they refer to the same individual. The former suggests that the referent is not familiar to the audience or that the descriptive appositive does not have unique reference, whereas the latter suggests exactly the opposite. The referent is familiar to the audience and/ or the appositive has unique reference (cf. section 4.3 above). A noun phrase in genitive case or a possessive pronoun as the determinative element has the same effect as the definite article. It makes the descriptive appositive definite and claims unique reference for it. The noun phrase in genitive case in the determiner position of a superordinate noun phrase must itself be definite. Otherwise it produces an ungrammatical phrase as in (9), in which the determinative element is highlighted. (9)

*a company 's executive, Roger Gethin

This contrasts with the acceptability of (10) to (12): (10) (11) (12)

the executive of a company, Roger Gethin an executive of the company, Roger Gethin the company's executive, Roger Gethin

These three examples differ in the claims they make about the familiarity or uniqueness of the two referents, i.e. the "executive" and the "company", re-

224

Noun phrase name appositions

spectively. In (10) the "executive" is taken to be familiar, on the basis of realworld knowledge or on the basis of the preceding context, and the "company" is taken not to be familiar. The reverse is the case in (11). In example (12), both the "executive" and the "company" are taken to be familiar and/or unique. There are thus clear semantic differences between the three (a)patterns. In each case the given situation of familiarity or unfamiliarity of the individual referents influences the choice of the pattern. These distinctions must be borne in mind if the frequencies of the three (a)-patterns are to be compared. The same differences as for the three (a)-patterns holds true for the three (c)-patterns. The choice among them again depends on the assumed familiarity or unfamiliarity (or uniqueness) of the individual referents. The distinction between the three (a)-patterns on the one hand and the (b)pattern on the other, however, is of a different nature. Pattern (b) can replace any of the three (a)-patterns because it gives less information. Example (13) could stand for any of the examples (10) to (12) above. It does not make any claims about the familiarity and/or uniqueness of the referents "company" and "executive". (13)

company executive Roger Gethin

The journalist has to make a decision between the three (a)-patterns on the basis of semantic considerations. Given a certain constellation of facts concerning the familiarity and/or uniqueness of the referents, there is no choice between the (a)-patterns. But there is always the alternative to use pattern (b), which means that for any given set of facts there is a real choice to be made by the journalist. There is very little difference between the individual (a)-pattems and their corresponding (c)-patterns in semantic terms. (14) (15)

the company executive, Roger Gethin Roger Gethin, the company executive

Examples (14) and (15) are identical in referential meaning in that they refer to the same individual, "Roger Gethin". They also make the same claims about the familiarity or uniqueness of the referent. The company has presumably been mentioned in the preceding context. It is therefore assumed to be familiar to the reader. The executive, on the other hand, is taken to have unique reference. The name, Roger Gethin, is therefore a nonrestrictive modification of the descriptive appositive as it is indicated by the presence of the comma.

Linguistic constraints

225

In (15), the descriptive appositive is the nonrestrictive modification of the name. (16) (17)

Roger Gethin, an executive of the company Roger Gethin, a company executive

In (16) and (17), which illustrate pattern (cl), it is assumed that the "executive" is only one among others. The difference here is that (16) specifies the "company" as familiar to the reader whereas in (17) it can be familiar or not. Moreover (16) seems to imply that this company has more than one executive, whereas (17) is vaguer in this respect and does not carry any such implication. Pattern (c3) is practically synonymous with pattern (a3) as in (18) and (19): (18) (19)

the company's executive, Roger Gethin Roger Gethin, the company's executive

Both examples make the same claims about the familiarity. The "company" is taken to be familiar, and the "executive" unique. Format (d) with the zero article is less explicit in that it does not make any claims about the uniqueness and familiarity of the referent of the head noun. If exactly the same descriptive appositives are used, as for instance in Labour Leader Neil Kinnock and Neil Kinnock, Labour Leader, the two formats are virtually synonymous. However, as I will show in detail below, the descriptive appositive in postposed position tends to be modified by a prepositional phrase where its equivalent in preposed position would be modified by a simple noun. (20) and (21) are typical examples of the patterns (b) and (d): (20) (21)

company executive Roger Gethin Roger Gethin, executive of the company

Example (21) is more explicit than (20) by stating that the "company" is taken to be familiar to the reader. In (20) no such claim is made. Thus the choice between the three (a)-patterns, as pointed out above, is largely semantic, whereas the choice between either one of the (a)-patterns or pattern (b) is stylistic. Likewise the choice between the three (c)-patterns is largely semantic and the choice between either one of the (c)-patterns and pattern (d) is stylistic. The patterns (al), (a2) and (a3) correspond more or less closely to the patterns (cl), (c2) and (c3), respectively. Even in the cases of near semantic equivalence, there still are several factors affecting the choice between individual patterns. As implied above, it can

226

Noun phrase name appositions

be a stylistic choice a choice made by the journalist of how explicit a descriptive appositive should be and how much information it should contain , but the choice can also be more formal depending on the presence or absence of specific elements and on the number and complexity of the elements. Before I can turn to the stylistic choices made by the journalists, it first has to be established how much of the variability can be attributed to these formal or syntactic factors.

9.4.2. The syntactic constraints The following linguistic factors may be expected to influence the choice of particular noun phrase name apposition structures: Firstly, the determinative element is more resistant to deletion if it is realised by a noun phrase in the genitive case. Secondly the fact whether it is a definite or an indefinite article will affect the likelihood with which appositives appear in preposed or postposed position. And thirdly, the complexity of the descriptive appositive will also influence the possibility of determiner deletion and its position in relation to the name. The following table gives the frequency of the individual patterns as I have collected them in the eleven British dailies in samples of roughly equal size. The figures given here comprise in each case more or less equivalent samples from the home news section of each paper. Table 9.4. Types of noun phrase name appositions in percentage

T

DMi

Star

Sun

DE

DMa

To

DT

FT

G

I

al a2 a3 alia

0.7 0.0 17.5 18.2

0.0 0.0 9.8 9.8

0.0 0.0 15.7 15.7

0.8 3.9 19.4 24.0

1.5 6.5 18.8 26.8

0.8 0.0 22.1 22.9

1.6 7.9 11.0 20.5

0.8 3.2 0.8 4.8

6.1 32.7 11.6 50.3

0.9 0.0 6.4 6.8 6.4 11.3 13.8 18.1

b

67.1

70.5

69.6

58.1

54.4

55.7

4.7

1.6

0.7

2.8

cl c2 c3 alle

4.2 2.8 2.1 9.1

5.7 4.9 0.8 1.5

6.1 0.0 0.0 6.1

6.2 2.3 0.8 9.3

5.1 4.4 0.7 10.1

7.6 1.5 1.5 10.7

18.1 12.6 5.5 36.2

10.3 31.0 12.7 54.0

8.2 18.4 4.1 30.6

12.8 17.3 30.3 19.6 8.3 6.0 51.4 42.9

d

5.6

8.2

8.7

8.5

8.7

10.7

38.6

39.7

18.4

32.1 38.4

total N

143

122

115

129

138

131

127

126

147

109

0.8

133

Actual number of occurrences given as totals. For the individual figures cf. Appendix III, table A9.1.

Linguistic constraints

227

The letters (al) to (d) in table 9.4 stand for the eight patterns of noun phrase name appositions introduced above. For ease of reference, they are repeated here: (22)

(al) (a2) (a3) (b) (cl) (c2) (c3) (d)

[a(indef) N'] Na [a(def)N']Na [genN']Na [0 N'] Na Na [a(indef) N1] Na[a(def)N'] Na[genN'] Na [0 N1]

The first hypothesis mentioned above claims that genitives and possessives are more resistant to being omitted than either the definite or indefinite article. This is a claim which only applies to the tabloid papers, i.e. to the downmarket and the mid-market papers, because only these papers omit the determiner with great frequency if the descriptive appositive is in preposed position. The patterns (al) and (a2), with an indefinite and a definite article respectively, are virtually non-existent in the down-market papers and very rare in the mid-markets. However, all these papers have sizeable proportions of the pattern (a3) with a noun in the genitive case or a possessive pronoun in the determiner position. All of them have between 10 and 20 per cent in this form, and it is the second most important pattern after pattern (b) with a zero determiner for all of them. No difference has been made between genitive noun phrases in determiner position that are preceded by an article themselves and those that are not, because there are several other factors involved as well. (23) (24) (25)

Labour's Leader, Neil Kinnock (the) doctor's daughter, Carole Burton her father, Dr Geoffrey Burton

The article can only be present in example (24), and only in this case would its absence have any potential significance. Example (23), with a proper noun in genitive case in determiner position, would be ungrammatical with a preceding article. The same is true for example (25), where the determiner position is filled by a possessive pronoun, which of course can never be preceded by an article. However, as it is often difficult to decide whether in a given case the article could possibly have been present and as these cases in which such a

228

Noun phrase name appositions

decision might have mattered are not very numerous for any of the newspapers, they have all been lumped together under category (a3), whether the noun in determiner position and in the genitive case is preceded by its own article or not. The same policy was adopted for pattern (c3), in which the descriptive appositive appears in postposed position. The second hypothesis says that the choice between the definite and the indefinite article may have a bearing on whether the descriptive appositive appears in preposed or in postposed position. As outlined in the previous section, this is a fairly free choice for any given situation. The two locutions a selfmade multi-millionaire, Harry Goodman and Harry Goodman, a self-made multi-millionaire, are fairly synonymous in that they are not only identical in their referential meaning but also in their explicitness as to the familiarity and/ or uniqueness of the referent. All the papers have larger proportions of (cl)-patterns than (al). Pattern (al) is used in less than two per cent of all noun phrase name appositions in all papers with the single exception of The Guardian, where it is 6.1 per cent, whereas pattern (cl) is used in between 4 and 6 per cent in the tabloids, with the exception of Today, which has 7.6 per cent, and between 8 and 18 in the up-market papers. The third hypothesis seems particularly obvious even from a short look at the data. Preposed descriptive appositives tend to be shorter and less complex than postposed descriptive appositives. A simple count of the elements in the descriptive appositive reveals to what extent this is true. For the purpose of this count I have excluded the determiner of the appositive, prepositions, further articles and auxiliary verbs. On an average there are roughly two elements in preposed descriptive appositives and roughly four elements in postposed ones. There is little difference between the three types of papers in this respect. Table 9.5. Average number of main elements in preposed and postposed descriptive appositives

preposed postposed

DMi

Star

Sun

DE

DMa

To

DT

FT

G

I

T

1.9 3.9

1.9 5.1

1.9 4.5

1.8 4.6

1.9 4.3

2.0 4.1

1.7 3.8

2.3 3.3

1.9 4.0

1.8 4.0

1.8 4.1

Figures exclude determiners, prepositions, articles and auxiliary verbs.

There appears to be some greater variation in the up-market papers than in the tabloids, but the up-markets have only very few instances of preposed descriptive appositives, and therefore these figures are probably less than reliable in this respect.

Linguistic constraints

229

Postposed descriptive appositives are, if anything, rather shorter in the upmarkets than in the down-markets, contrary to what one might expect. The obvious explanation is that the tabloids very much prefer the preposed descriptive appositives and only resort to postposition if the descriptive appositive is too heavy in terms of the number of its elements and in terms of its complexity. The up-markets, on the other hand, routinely use the postposition even for comparatively light descriptive appositives. With only one exception, all the papers have got on an average at least twice as many elements in the postposed appositives than in the preposed appositives. The only exception is the Financial Times. This paper, however, is exceptional anyway in that it has very few instances of preposed descriptive appositives (6.4 per cent). Postposition for this paper is therefore nearly categorical whatever the weight of the appositive. These results are further corroborated if the complexity of the preposed and the postposed descriptive appositives is compared. The density of complex appositions is the percentage of appositions that have more than one modification (adjective, prepositional phrase, relative clause, etc.) of their descriptive appositive in relation to the total number of appositions. Predictably the density is considerably higher for postposed descriptive appositives than for preposed descriptive appositives. Table 9.6. Density of complex appositions

preposed postposed

DMi

Star

Sun

DE

DMa

To

DT

FT

G

I

T

16 50

16 53

19 60

13 46

19 39

19 49

16 40

25 51

23 43

13 48

18 51

Figures are percentages of appositives that have more than one modification.

The postposed appositives have a density of between 40 and 60 per cent, that is to say that roughly half of all the descriptive appositives in postposed position have got multiple modification. Again there is more variation for the upmarkets than there is either for the mid-markets or the down-markets, but no consistent pattern emerges from the above figures. If we compare the average number of elements for the individual patterns, one interesting result emerges. The pattern (a3) in all cases has got an average number of elements in the descriptive appositive that is well below the average for all preposed descriptive appositives. It ranges from 1.3 to 1.7 elements. In all the papers, the weight of the descriptive appositive is smaller for this pattern than for any other pattern. This means that (a3) appositives are very often unmodified and if they are, they are never very heavily modified. One obvious explanation for this phenomenon lies in the fact that a noun

230

Noun phrase name appositions

phrase in the determiner position very often fulfils a similar role as a modifier of the head might have done, as in the examples Labour's Leader, Neil Kinnock and the Labour Leader Neil Kinnock, the former of which has an unmodified noun in the descriptive appositive whereas the latter has a noun modified by a preceding noun.

9.5. Social constraints Ryden's (1975) paper indicates the major variables that can be expected to correlate with the social distinction between the three categories of papers. The first variable, which is also the major focus of Bell's (1985) paper, is the relation between the apposition patterns with a preposed noun and a definite or indefinite article and those with a zero article or in Bell's terminology the amount of determiner deletion. The second variable is the relation between preposed and postposed nouns in noun phrase name appositions. This relationship, which is just as significant as the determiner deletion, has been largely overlooked by Bell, presumably because he was interested in the workings of one particular hypothetical variable rule rather than in the construction of noun phrase name appositions itself. The third variable which stratifies the newspapers very clearly is the format of the name-appositive itself. The individual newspapers have very clear-cut preferences as to whether they use honorifics and titles or not. Hence, the crucial questions are firstly whether the determiner deletion has made further inroads into the up-market papers in comparison to the results reported above, secondly how coherent the usage is within the individual newspaper categories, and thirdly, beside the well-attested difference between the broadsheet papers and the tabloids taken together, whether there is also a difference between the down-market and the mid-market papers. Figure 9.5 shows the percentage of determiner deletion in the home news sections of the eleven papers. It reflects the figures given in table 9.4 above and plots the percentage of patterns (b), with zero article against the total of all preposed appositives excluding those with a genitive in determiner position. The down-market papers have categorical determiner deletion. All three of them delete the determiner in between 99 or 100 per cent of all preposed descriptive appositives, excluding pattern (a3). The mid-market papers are not quite as categorical as the down-markets but they still delete between 87 and 99 per cent. For the up-market papers the picture is less coherent. The Daily Telegraph deletes most (33 per cent) and The Guardian least (2 per cent).

Social constraints

231

ιυυ 90 80 70

60 50 40 20 10 0

I

DM i

I

Star

I

Sun

I

DE

I

I

DMa

To

I

DT

I ,

FT

.I

G

I

I

T

Figure 9.5. Determiner deletion in eleven British dailies (excluding pattern (a3) with a genitive determiner in the descriptive appositive). Based on the figures given in table 9.4

On this basis, Bell's (1985: 115) claim that "even the Times ... will lose [its] rearguard action and be drawn into increasing determiner deletion," seems to be justified. However, this claim, as well as figure 9.5, is based on the noun phrase name appositions with a preposed descriptive appositive only. Figure 9.5 uses the figures from table 9.4 above, which makes clear that for most upmarket papers preposed descriptive appositives are on the whole of very limited currency. In all the up-market papers, pattern (b) is used in less than 5 per cent of all noun phrase name appositions. It is only the fact that the other formats with a preposed descriptive appositive are rare too in these papers that produces sizeable amounts of determiner deletion. Allan Bell (private communication) makes the point that the absolute rarity of a form or a rule is irrelevant for the relative frequency of its application. For him it is therefore highly relevant that three of the five up-market papers have determiner deletion values of about 30 per cent (The Daily Telegraph, 33 per cent the Financial Times, 29 per cent and The Independent, 21 per cent). However, this point holds true only if we restrict our view to noun phrase name appositions with a preposed descriptive appositive. To take the Financial Times as an example, in absolute figures there are in my data one instance of pattern (al), four of pattern (a2) and two of pattern (b). One instance of pattern (a3) is ignored because of the different behaviour of this pattern as outlined above. This gives a determiner deletion value of 29 per cent, but it fails to appreciate that there are 118 instances of noun phrase name appositions with a postposed descriptive appositive in the same corpus. Seen in this light, it appears more likely that the two instances of pattern (b) are editorial lapses rather than an indication that the Financial

232

Noun phrase name appositions

Times is about to lose its "rearguard action" or that it gives in to "the force of the more popular media dragging [it] into the news language of the twentieth century "(Bell 1988: 342). A more adequate picture can be deduced from a comparison of the relative importance of all four patterns. 100 90

80 70 60 50 40 30 20 10 ° DMi

Ϊ77ΆΛ

Star

Sun

ES3 c

DE

DMa

To

E5553 h

DT

FT

OTTOln

Figure 9.6. Frequency of the four noun phrase name apposition patterns in per cent. The exact figures are given in table 9.4. /

Figure 9.6 reveals quite clearly that it is not so much the amount of determiner deletion that is crucial, even though this value too gives some important indication. It is rather the respective frequencies of the individual patterns that stratifies the three categories of newspapers. The tabloids have got a remarkably coherent style in this respect. Pattern (b) is by far the most important single pattern, and there is not very much difference between the patterns (c) and (d), the former being slightly more important in all cases. There is a small but discernible difference between the two tabloid categories. The down-market papers use pattern (b) in between 67 and 71 per cent of their noun phrase name appositions, whereas the percentages for the mid-market papers vary from 54 to 58 per cent. For the up-market papers the picture is altogether less coherent. The importance of pattern (b) varies considerably, as pointed out above. The Guardian is exceptional among the up-market papers because more than half of its noun phrase name appositions have the descriptive appositive in preposed position, but only very few of them are instances of pattern (b). As a result of this, the amount of determiner deletion is very low (2 per cent). The Times, on the other hand, which has a determiner deletion percentage of almost 10 per cent uses the pattern (b) just as rarely as The Guardian. It appears in 0.8 per cent of all

Social constraints

233

the noun phrase name appositions of The Times, and in 0.7 per cent of The Guardian. The Financial Times is exceptional even among the up-market papers in that only six per cent of all its noun phrase name appositions have the descriptive appositive in preposed position. The pattern (b) is used in 1.6 per cent of all noun phrase name appositions, but - as pointed out above - because of the very small total number of preposed descriptive appositives, the amount of determiner deletion turns out to be 29 per cent. In all these cases, the amount of determiner deletion is very misleading and conceals the fact that these papers are very reluctant indeed to use pattern (b). The distinction between the up-market papers and both types of tabloids is thus near categorical. In the former the postposed descriptive appositives are clearly preferred over the preposed descriptive appositives, and the pattern (b) with deleted determiner is only very rarely used, whereas the latter predominantly use pattern (b). In all of them this pattern is used in more than 54 per cent of their noun phrase name appositions. On the basis of the stratifying power of pattern (b) with deleted determiner in preposed position, it could be expected that the determiner deletion in descriptive appositives in postposed position may pattern in a similar fashion. This expectation, however, is not borne out by the data. There are, of course, individual differences, but no clear pattern emerges. 60 50 40 30 20 10 0

DMi

Star

Sun

DE

DMa

To

DT

FT

G

I

T

Figure 9.7. Determiner deletion in postposed position. Based on the figures given in table 9.4.

The Sun has got the highest percentage of determiner deletion in postposed position (54 per cent) and The Guardian the lowest (32 per cent). The following examples illustrate pattern (b) as it is used in the three types of newspapers. All these examples were taken out of the home news section of

234

Noun phrase name appositions

the individual papers. Examples are given for all three types of papers, but it must be remembered that this pattern is typical of the down-markets and the mid-markets but rather exceptional for the up-markets. down-market papers: (26) Tory health fanatic Edwina Currie (DMi hn 8.10.87, 1.1) (27) top nutritionist Dr Sally Pasonage (DMi hn 8.10.87, 1.1) (28) Royal relative Katie Baring (S hn 8.10.87, 1.2) (29) Left-wing firebrand Derek Hatton (S hn 8.10.87, 1.2) (30) Home Secretary Douglas Hurd (S hn 8.10.87, 2.2) (31) Bride Leigh Morris (S hn 8.10.87, 7.1) (32) new husband Caryl Patterson (S hn 8.10.87, 7.1) (33) owner Dr. Roy Sawyer (St hn 17.12.87, 1.1) (34) shop manageress Wendy Knell (St hn 17.12.87, 1.2) mid-market papers: (35) Spanish tourist Josefa Morelli (DE hn 5.11.87, 7.1) (36) Husband Harry (DE hn 5.11.87, 9.2) (37) Film star Jack Nicholson (DE hn 5.11.87, 13.1) (38) former Prime Minister Edward Heath (DMa hn 18.12.87, 1.1) (39) Cheeky office girl Susan Wilson (DMa hn 18.12.87, 1.2) (40) baby Alexander Davies (DMa hn 18.12.87, 2.1) (41) Delighted Employment Secretary Norman Fowler (DMa hn 18.12.87, 2.6) (42) baby Tyra Henry (To hn 18.12.87, 2.1) (43) Health Minister Tony Newton (To hn 18.12.87, 2.5) up-market papers: (44) baby David Barber (DT hn 27.11.87, 1.8) (45) missing Bristol bride Shirley Banks, (DT hn 27.11.87, 3.7) (46) Home Secretary Douglas Hurd (FT hn 22.1.88, 9.3) (47) Former Austin Rover UK sales director Mr Peter King (FT hn 22.1.88, 8.4) (48) TV presenters Mike Smith and Vicky Licorish, (G hn 17.12.87, 4.3) (49) 82-year-old publisher, Joseph Cleaver (I hn 6.10.87, 1.2) (50) businessman Harry Djanobally (T hn 8.10.87, 5.2) This more or less random selection of examples shows that the same wide range of nouns are used in all three types of papers. Designations of ministerial offices such as Home Secretary, which can also be used as vocatives, are more title-like than nouns like baby or bride, which cannot be used as voca-

Social constraints

235

lives, but examples of both types of nouns are attested in the up-market papers. Thus it does not appear to be the case that this construction makes its way into the prestige media via nouns that are to some extent ambivalent between being a title and a real appositive. Example (49) is slightly odd because it uses a comma between the two appositives in spite of the fact that the construction must be understood restrictively. It appears in the following context, which forms the first two sentences of a front page article. A sacked handyman and two accomplices went back to his former employer's New Forest mansion and murdered five of the occupants, Winchester Crown Court was told yesterday. The prosecution alleged that 82-year-old publisher, Joseph Cleaver, his wife, son, and nurse were tied up, dowsed with petrol and burnt alive. (I hn 6.10.87, 1.2)

In the second sentence the referent of the phrase 82-year-old publisher can be taken to be familiar. It is the same individual which the phrase his former employer in the first sentence refers to, but the phrase does not uniquely refer to its referent, and without a definite article it cannot be assumed to do so. The only solution to this puzzle seems to be to assume that the comma is a misprint. The following examples illustrate the remaining patterns. Pattern (a) down-market papers (51) the boss, Mr Lawless (DMi sp 8.10.87, 28,1) (52) the president of the board, General Safdar Butt (St sp 18.12.87, 28.2) (53) that vindictive crackpot Haseeb Ahsan (St sp 18.12.87, 25.6) (54) Her father, Noel Porter-Blake (DMi hn 28.10.87, 11.2) (55) his former secretary Sara Keays (S hn 8.10.87, 2.2) mid-market papers (56) A confident Chancellor Nigel Lawson (To hn 19.12.87, 2.2) (57) The obstetrician in charge of the Davies case, Mr Edmund Holt (DMa hn 18.12.87, 2.5) (58) the two youngsters Benn, 9, and Katy, 6, (DE hn 5.11.87, 9.2) (59) Their leader, Mrs Jenny Keyte (DMa hn 18.12.87, 9.2) (60) her two-year-old son Bryn (DE hn 5.11.87, 3.1) (61) Janet's eldest son by a previous marriage Mark (DE hn 5.11.87, 9.2) (62) her Environment Secretary Nicholas Ridley (DMa hn 18.12.87, 1.1) (63) his wife Bernice (DMa hn 18.12.87, 15.3) up-market papers (64) a deputy chief executive, Mr Michael Dobson (FT hn 26.1.88, 8.5)

236

(65) (66) (67) (68) (69) (70) (71) (72) (73) (74)

Noun phrase name appositions

A former Tory Health Minister, Sir Barney Hayhoe (DT hn 27.11.87, 1.8) the Chancellor, Mr Nigel Lawson (G hn 31.10.87, 1.8) The Home Secretary, Douglas Hurd (I hn 16.11.87, 5.3) the gun fanatic Michael Ryan (FT hn 22.1.88, 9.8) the fugitive property developer, Gordon Parry (G hn 31.10.87, 2.1) the administrator Mr John Scott (DT hn 27.11.87, 2.7) Last season's World Cup slalom champion, Corinne Schmidhauser (DTsp27.11.87,38.7) his fellow left-winger Bob Cryer (FT hn 22.1.88, 9.3) The group's chairman, Mr Douglas Robertson (T hn 8.10.87, 7.1) Thomas Cleaver's wife, Wendy, 46 (I hn 6.10.87, 3.4)

The patterns (al) and (a2) with an indefinite or a definite article as the determinative of the descriptive appositive are not typical of any of the papers with the one exception of The Guardian, which has got just over 50 per cent of its noun phrase name appositions in this format. The above examples illustrate that they nevertheless do appear in all three types of papers. There is no consistency in the use of a comma between the descriptive appositive and the name. The restrictiveness of the name in relation to the descriptive appositive is not a reliable guide either. In (60) to (63), for instance, the descriptive appositives can only be understood as having unique reference, and therefore only a nonrestrictive modification should be possible, which, according to common practice, requires a comma. Pattern (c) down-market papers: (75) Mrs Currie, the junior health minister who has lectured the nation on sensible eating (DMi hn 8.10.87, 1.1) (76) Justine, a computer expert, of St Johns, Worcester (S hn 20.10.87, 13.5) (77) Frank Langella, the former screen Dracula who plays arch-enemy Skeletor(Sthn 18.12.87, 15.1) (78) Gordon Wilson, the father of the nurse who died in the carnage (St hn 19.12.87, 1.5) (79) Susan, a student nurse on £ 75 a week (St hn 19.12.87, 4.3) (80) Polly Bloomfield, the beautiful mistress of smoothie actor Nigel Havers (Sthn 19.12.87, 5.1) mid-market papers: (81) Glenn Close, the lovely star of smash-hit movie Fatal Attraction (DE hn 5.11.87, 3.1)

Social constraints

(82) (83) (84)

237

Mr Davies, a management consultant (DMa hn 18.12.87, 2.1) John Biffen, the former Commons leader sacked from the Thatcher Cabinet after the election (DMa hn 18.12.87, 2.1) Michael Frost, the man she has lived with for six years (To hn 18.12.87,5.5)

up-market papers: (85) Mr Gorbachev, the Soviet leader, (DT hn 27.11.87, 1.1) (86) Mr Roger Gethin, an executive with the company (DT hn 27.11.87, 3.1) (87) Mr Jim Thomas, a national officer of the manufacturing union Tass, (G hn 31.10.87, 2.2) (88) Mr Gavin Laird, the AEU general secretary (G hn 31.10.87, 2.3) (89) Malcolm Rifkind, the Secretary of State for Scotland (I hn 6.10.87, 1.4) (90) Norman Tebbit, the Tory party chairman (I hn 6.10.87, 1.4) (91) Mr Douglas Hurd, the Home Secretary (T hn 8.10.87, 1.1) Pattern (d) down-market papers: (92) Commander George Churchill-Coleman, head of Scotland Yard's antiterrorist squad (DMi hn 28.10.87, 2.3) (93) David Storey, chairman of the right-wing Monday Club (S hn 8.10.87, 2.2) (94) Dr. Peter Phillips, chairman of the West Berks Area Health Authority (Sthn 18.12.87, 5.1) mid-market papers: (95) Robin Lees, chairman of the British Hotels, Restaurants and Catering Association (DE hn 5.11.87, 10.2) (96) Fred Carr, marketing director of Capel-Cure Mayers (To hn 18.12.87, 2.5) up-market papers: (97) Mrs Anita Roddick, managing director of Body Shop (DT hn 27.11.87, 1.4) (98) Mr John Moore, Social Services Secretary (FT hn 22.1.88, 6.1) (99) Frank Evans, chairman of the council committee (I hn 6.10.87, 2.2) (100) Mr Parkinson, Secretary of State for Energy (T hn 8.10.87, 1.4) Both the patterns (c) and (d) are typical of the up-market papers. They are clearly dispreferred by the tabloids. Again the examples were chosen to illustrate their typical usage as well as their less characteristic usage. For these two

238

Noun phrase name appositions

patterns, the comma between the appositives is obligatory, and it appears in all the examples. These examples moreover illustrate the point made above that postposition of the descriptive appositive is often chosen because of its length and/or complexity. Examples (75), (77), (78) or (84), which include a relative clause in the descriptive appositive would be impossible in a preposed format. A small number of noun phrase name appositions have more than one descriptive appositive. The possibilities appear to be almost without limits. Most patterns combine quite freely with all the other patterns. In the case of two descriptive appositives, they can both be preposed or postposed, in which case they may either be in apposition to each other or they may be conjoined by and, or one of them may be preposed and the other postposed. In most cases, the same choice for the determiner realisation applies as in the case of single descriptive appositives. The following examples give an indication of the range of possibilities as they appear in the three types of papers. (101) the father, film producer John Starke (DE hn 5.11.87, 3.1) (102) her husband, playwright Tom Stoppard (S hn 8.10.87, 3.1) (103) Mr Simpson, the vicar of Barmston, North Humberside and a local Conservative councillor (To hn 18.12.87, 4.2) (104) Volvo driver Dov Lisovsky, a 41-year-old car mechanic(S hn 8.10.87, 3.1) (105) wildlife ranger Pat O'Connell, an expert falconer who is now caring for the feathered celebrity (DMa hn 18.12.87, 5.3) (106) billionaire Lord Vetey, Britain's most notorious tax dodger (DMi hn 28.10.87,7.4) (107) his friend Timothy Hutton, a co-star from the movie The Falcon and the Snowman (To hn 18.12.87,5.1) (108) the country's biggest earner Christopher Heath, managing director of Baring Securities (To hn 18.12.87, 7.2) (109) the minister responsible for this aspect, Mr Tom King, the Northern Ireland Secretary (FT hn 26.1.88, 13.3) (110) a 39-year-old prison officer, George Jolly, a married man with two children who was taken captive on Sunday evening (I hn 6.10.87, 1.2) (111) Her friend, Mr Desmond McClean, a music teacher, aged 20 (T hn 20.11.87, 1.2) In (101) and (102) both descriptive appositives are preposed, and in both cases the first one has a non-zero determiner and the second a zero determiner. In example (103), both descriptive appositives are postposed and conjoined by and. The first one has got a definite article because the reference of the phrase

Noun phrase name appositions in the sports section

239

vicar ofBarmston is apparently unique. There is only one individual it can refer to. The second descriptive appositive, on the other hand, has got an indefinite determiner because this phrase does not have unique reference. There are quite clearly several individuals to whom the phrase local Conservative councillor could refer. The remaining examples illustrate a range of combinations of preposed and postposed descriptive appositives as they are used in the different types of papers. The comma is obligatory in front of postposed descriptive appositives, and it never occurs after preposed ones that have a zero article. The usage of preposed descriptive appositives with a non-zero article is again inconsistent. In most cases there is a comma, but in example (108), in which a comma would be expected on the assumption that the phrase the country's biggest earner must have unique reference and that therefore the name can only be a nonrestrictive modification.

9.6. Noun phrase name appositions in the sports section The above results were all based on data taken from the home news sections alone because the considerable differences between this and other sections would have skewed any aggregate samples to an unacceptable degree. In this section, I would like to show how much difference there actually is be80 % 70

60 50 40 30 20 10 0 Υ//Λ hnmp news

KXVispnrt

Figure 9.8. Noun phrase name appositions in the home news and sports sections of the down-market papers in per cent. For the exact figures cf. Appendix III, tables A9.1 and A9.2.

240

Noun phrase name appositions

tween just two sections, that is between the home news section and the sports section. As I have shown above, it is not so much the amount of determiner deletion that stratifies the three types of newspapers, but rather the distribution of the four main types of apposition patterns. The following figures show the results for the combined sets of papers for each category. Figure 9.8 shows the combined set of down-market papers (Daily Mirror, Sun and Star). The individual columns indicate the percentage of one particular pattern in all the noun phrase name appositions of either the home news section or the sports section. There is hardly any difference at all between the two sections. Pattern (b) is by far the most frequent, followed by pattern (a), most of which are instances of pattern (a3). The patterns (c) and (d) are fairly insignificant. If a chi-square test is applied to the actual figures underlying this figure, the difference between the combined sets of the home news sections and the sports sections turn out to be not statistically significant, (χ2 = 3.71, df = 3, p: not significant). For the combined set of mid-market papers (Daily Express, Daily Mail and Today) there is more variety, as can be seen in figure 9.9.

80 r 70 60 50

30 20 10 0 \//Λ hnmp news

KNNM

Figure 9.9. Noun phrase name appositions in the home news and sports sections of the mid-market papers in per cent. For the exact figures cf. Appendix III, tables A9.1 and A9.2.

In the sports sections, the pattern (b) is more frequent than in the home news section, mainly at the expense of pattern (a), which has got a greater frequency in the home news sections. Even though the difference does not look very decisive, a chi-square test applied to the actual figures underlying figure 9.9 indi-

Noun phrase name appositions in the sports section

241

cates that the difference is statistically significant at the 0.1 per cent level (% 2 =19.49,df=3,p